Search Results

Search found 18028 results on 722 pages for 'atomic values'.

Page 374/722 | < Previous Page | 370 371 372 373 374 375 376 377 378 379 380 381  | Next Page >

  • Sliding collision response

    - by dbostream
    I have been reading plenty of tutorials about sliding collision responses yet I am not able to implement it properly in my project. What I want to do is make a puck slide along the rounded corner boards of a hockey rink. In my latest attempt the puck does slide along the boards but there are some strange velocity behaviors. First of all the puck slows down a lot pretty much right away and then it slides for awhile and stops before exiting the corner. Even if I double the speed I get a similar behavior and the puck does not make it out of the corner. I used some ideas from this document http://www.peroxide.dk/papers/collision/collision.pdf. This is what I have: Update method called from the game loop when it is time to update the puck (I removed some irrelevant parts). I use two states (current, previous) which are used to interpolate the position during rendering. public override void Update(double fixedTimeStep) { /* Acceleration is set to 0 for now. */ Acceleration.Zero(); PreviousState = CurrentState; _collisionRecursionDepth = 0; CurrentState.Position = SlidingCollision(CurrentState.Position, CurrentState.Velocity * fixedTimeStep + 0.5 * Acceleration * fixedTimeStep * fixedTimeStep); /* Should not this be affected by a sliding collision? and not only the position. */ CurrentState.Velocity = CurrentState.Velocity + Acceleration * fixedTimeStep; Heading = Vector2.NormalizeRet(CurrentState.Velocity); } private Vector2 SlidingCollision(Vector2 position, Vector2 velocity) { if(_collisionRecursionDepth > 5) return position; bool collisionFound = false; Vector2 futurePosition = position + velocity; Vector2 intersectionPoint = new Vector2(); Vector2 intersectionPointNormal = new Vector2(); /* I did not include the collision detection code, if a collision is detected the intersection point and normal in that point is returned. */ if(!collisionFound) return futurePosition; /* If no collision was detected it is safe to move to the future position. */ /* It is not exactly the intersection point, but slightly before. */ Vector2 newPosition = intersectionPoint; /* oldVelocity is set to the distance from the newPosition(intersection point) to the position it had moved to had it not collided. */ Vector2 oldVelocity = futurePosition - newPosition; /* Project the distance left to move along the intersection normal. */ Vector2 newVelocity = oldVelocity - intersectionPointNormal * oldVelocity.DotProduct(intersectionPointNormal); if(newVelocity.LengthSq() < 0.001) return newPosition; /* If almost no speed, no need to continue. */ _collisionRecursionDepth++; return SlidingCollision(newPosition, newVelocity); } What am I doing wrong with the velocity? I have been staring at this for very long so I have gone blind. I have tried different values of recursion depth but it does not seem to make it better. Let me know if you need more information. I appreciate any help. EDIT: A combination of Patrick Hughes' and teodron's answers solved the velocity problem (I think), thanks a lot! This is the new code: I decided to use a separate recursion method now too since I don't want to recalculate the acceleration in each recursion. public override void Update(double fixedTimeStep) { Acceleration.Zero();// = CalculateAcceleration(fixedTimeStep); PreviousState = new MovingEntityState(CurrentState.Position, CurrentState.Velocity); CurrentState = SlidingCollision(CurrentState, fixedTimeStep); Heading = Vector2.NormalizeRet(CurrentState.Velocity); } private MovingEntityState SlidingCollision(MovingEntityState state, double timeStep) { bool collisionFound = false; /* Calculate the next position given no detected collision. */ Vector2 futurePosition = state.Position + state.Velocity * timeStep; Vector2 intersectionPoint = new Vector2(); Vector2 intersectionPointNormal = new Vector2(); /* I did not include the collision detection code, if a collision is detected the intersection point and normal in that point is returned. */ /* If no collision was detected it is safe to move to the future position. */ if (!collisionFound) return new MovingEntityState(futurePosition, state.Velocity); /* Set new position to the intersection point (slightly before). */ Vector2 newPosition = intersectionPoint; /* Project the new velocity along the intersection normal. */ Vector2 newVelocity = state.Velocity - 1.90 * intersectionPointNormal * state.Velocity.DotProduct(intersectionPointNormal); /* Calculate the time of collision. */ double timeOfCollision = Math.Sqrt((newPosition - state.Position).LengthSq() / (futurePosition - state.Position).LengthSq()); /* Calculate new time step, remaining time of full step after the collision * current time step. */ double newTimeStep = timeStep * (1 - timeOfCollision); return SlidingCollision(new MovingEntityState(newPosition, newVelocity), newTimeStep); } Even though the code above seems to slide the puck correctly please have a look at it. I have a few questions, if I don't multiply by 1.90 in the newVelocity calculation it doesn't work (I get a stack overflow when the puck enters the corner because the timeStep decreases very slowly - a collision is found early in every recursion), why is that? what does 1.90 really do and why 1.90? Also I have a new problem, the puck does not move parallell to the short side after exiting the curve; to be more exact it moves outside the rink (I am not checking for any collisions with the short side at the moment). When I perform the collision detection I first check that the puck is in the correct quadrant. For example bottom-right corner is quadrant four i.e. circleCenter.X < puck.X && circleCenter.Y puck.Y is this a problem? or should the short side of the rink be the one to make the puck go parallell to it and not the last collision in the corner? EDIT2: This is the code I use for collision detection, maybe it has something to do with the fact that I can't make the puck slide (-1.0) but only reflect (-2.0): /* Point is the current position (not the predicted one) and quadrant is 4 for the bottom-right corner for example. */ if (GeometryHelper.PointInCircleQuadrant(circleCenter, circleRadius, state.Position, quadrant)) { /* The line is: from = state.Position, to = futurePosition. So a collision is detected when from is inside the circle and to is outside. */ if (GeometryHelper.LineCircleIntersection2d(state.Position, futurePosition, circleCenter, circleRadius, intersectionPoint, quadrant)) { collisionFound = true; /* Set the intersection point to slightly before the real intersection point (I read somewhere this was good to do because of floting point precision, not sure exactly how much though). */ intersectionPoint = intersectionPoint - Vector2.NormalizeRet(state.Velocity) * 0.001; /* Normal at the intersection point. */ intersectionPointNormal = Vector2.NormalizeRet(circleCenter - intersectionPoint) } } When I set the intersection point, if I for example use 0.1 instead of 0.001 the puck travels further before it gets stuck, but for all values I have tried (including 0 - the real intersection point) it gets stuck somewhere (but I necessarily not get a stack overflow). Can something in this part be the cause of my problem? I can see why I get the stack overflow when using -1.0 when calculating the new velocity vector; but not how to solve it. I traced the time steps used in the recursion (initial time step is always 1/60 ~ 0.01666): Recursion depth Time step next recursive call [Start recursion, time step ~ 0.016666] 0 0,000985806527246773 [No collision, stop recursion] [Start recursion, time step ~ 0.016666] 0 0,0149596704364629 1 0,0144883449376379 2 0,0143155612984837 3 0,014224925727213 4 0,0141673917461608 5 0,0141265435314026 6 0,0140953966184117 7 0,0140704653746625 ...and so on. As you can see the collision is detected early in every recursive call which means the next time step decreases very slowly thus the recursion depth gets very big - stack overflow.

    Read the article

  • More Fun with C# Iterators and Generators

    - by James Michael Hare
    In my last post, I talked quite a bit about iterators and how they can be really powerful tools for filtering a list of items down to a subset of items.  This had both pros and cons over returning a full collection, which, in summary, were:   Pros: If traversal is only partial, does not have to visit rest of collection. If evaluation method is costly, only incurs that cost on elements visited. Adds little to no garbage collection pressure.    Cons: Very slight performance impact if you know caller will always consume all items in collection. And as we saw in the last post, that con for the cost was very, very small and only really became evident on very tight loops consuming very large lists completely.    One of the key items to note, though, is the garbage!  In the traditional (return a new collection) method, if you have a 1,000,000 element collection, and wish to transform or filter it in some way, you have to allocate space for that copy of the collection.  That is, say you have a collection of 1,000,000 items and you want to double every item in the collection.  Well, that means you have to allocate a collection to hold those 1,000,000 items to return, which is a lot especially if you are just going to use it once and toss it.   Iterators, though, don't have this problem.  Each time you visit the node, it would return the doubled value of the node (in this example) and not allocate a second collection of 1,000,000 doubled items.  Do you see the distinction?  In both cases, we're consuming 1,000,000 items.  But in one case we pass back each doubled item which is just an int (for example's sake) on the stack and in the other case, we allocate a list containing 1,000,000 items which then must be garbage collected.   So iterators in C# are pretty cool, eh?  Well, here's one more thing a C# iterator can do that a traditional "return a new collection" transformation can't!   It can return **unbounded** collections!   I know, I know, that smells a lot like an infinite loop, eh?  Yes and no.  Basically, you're relying on the caller to put the bounds on the list, and as long as the caller doesn't you keep going.  Consider this example:   public static class Fibonacci {     // returns the infinite fibonacci sequence     public static IEnumerable<int> Sequence()     {         int iteration = 0;         int first = 1;         int second = 1;         int current = 0;         while (true)         {             if (iteration++ < 2)             {                 current = 1;             }             else             {                 current = first + second;                 second = first;                 first = current;             }             yield return current;         }     } }   Whoa, you say!  Yes, that's an infinite loop!  What the heck is going on there?  Yes, that was intentional.  Would it be better to have a fibonacci sequence that returns only a specific number of items?  Perhaps, but that wouldn't give you the power to defer the execution to the caller.   The beauty of this function is it is as infinite as the sequence itself!  The fibonacci sequence is unbounded, and so is this method.  It will continue to return fibonacci numbers for as long as you ask for them.  Now that's not something you can do with a traditional method that would return a collection of ints representing each number.  In that case you would eventually run out of memory as you got to higher and higher numbers.  This method, though, never runs out of memory.   Now, that said, you do have to know when you use it that it is an infinite collection and bound it appropriately.  Fortunately, Linq provides a lot of these extension methods for you!   Let's say you only want the first 10 fibonacci numbers:       foreach(var fib in Fibonacci.Sequence().Take(10))     {         Console.WriteLine(fib);     }   Or let's say you only want the fibonacci numbers that are less than 100:       foreach(var fib in Fibonacci.Sequence().TakeWhile(f => f < 100))     {         Console.WriteLine(fib);     }   So, you see, one of the nice things about iterators is their power to work with virtually any size (even infinite) collections without adding the garbage collection overhead of making new collections.    You can also do fun things like this to make a more "fluent" interface for for loops:   // A set of integer generator extension methods public static class IntExtensions {     // Begins counting to inifity, use To() to range this.     public static IEnumerable<int> Every(this int start)     {         // deliberately avoiding condition because keeps going         // to infinity for as long as values are pulled.         for (var i = start; ; ++i)         {             yield return i;         }     }     // Begins counting to infinity by the given step value, use To() to     public static IEnumerable<int> Every(this int start, int byEvery)     {         // deliberately avoiding condition because keeps going         // to infinity for as long as values are pulled.         for (var i = start; ; i += byEvery)         {             yield return i;         }     }     // Begins counting to inifity, use To() to range this.     public static IEnumerable<int> To(this int start, int end)     {         for (var i = start; i <= end; ++i)         {             yield return i;         }     }     // Ranges the count by specifying the upper range of the count.     public static IEnumerable<int> To(this IEnumerable<int> collection, int end)     {         return collection.TakeWhile(item => item <= end);     } }   Note that there are two versions of each method.  One that starts with an int and one that starts with an IEnumerable<int>.  This is to allow more power in chaining from either an existing collection or from an int.  This lets you do things like:   // count from 1 to 30 foreach(var i in 1.To(30)) {     Console.WriteLine(i); }     // count from 1 to 10 by 2s foreach(var i in 0.Every(2).To(10)) {     Console.WriteLine(i); }     // or, if you want an infinite sequence counting by 5s until something inside breaks you out... foreach(var i in 0.Every(5)) {     if (someCondition)     {         break;     }     ... }     Yes, those are kinda play functions and not particularly useful, but they show some of the power of generators and extension methods to form a fluid interface.   So what do you think?  What are some of your favorite generators and iterators?

    Read the article

  • My raycaster is putting out strange results, how do I fix it?

    - by JamesK89
    I'm working on a raycaster in ActionScript 3.0 for the fun of it, and as a learning experience. I've got it up and running and its displaying me output as expected however I'm getting this strange bug where rays go through corners of blocks and the edges of blocks appear through walls. Maybe somebody with more experience can point out what I'm doing wrong or maybe a fresh pair of eyes can spot a tiny bug I haven't noticed. Thank you so much for your help! Screenshots: http://i55.tinypic.com/25koebm.jpg http://i51.tinypic.com/zx5jq9.jpg Relevant code: function drawScene() { rays.graphics.clear(); rays.graphics.lineStyle(1, rgba(0x00,0x66,0x00)); var halfFov = (player.fov/2); var numRays:int = ( stage.stageWidth / COLUMN_SIZE ); var prjDist = ( stage.stageWidth / 2 ) / Math.tan(toRad( halfFov )); var angStep = ( player.fov / numRays ); for( var i:int = 0; i < numRays; i++ ) { var rAng = ( ( player.angle - halfFov ) + ( angStep * i ) ) % 360; if( rAng < 0 ) rAng += 360; var ray:Object = castRay(player.position, rAng); drawRaySlice(i*COLUMN_SIZE, prjDist, player.angle, ray); } } function drawRaySlice(sx:int, prjDist, angle, ray:Object) { if( ray.distance >= MAX_DIST ) return; var height:int = int(( TILE_SIZE / (ray.distance * Math.cos(toRad(angle-ray.angle))) ) * prjDist); if( !height ) return; var yTop = int(( stage.stageHeight / 2 ) - ( height / 2 )); if( yTop < 0 ) yTop = 0; var yBot = int(( stage.stageHeight / 2 ) + ( height / 2 )); if( yBot > stage.stageHeight ) yBot = stage.stageHeight; rays.graphics.moveTo( (ray.origin.x / TILE_SIZE) * MINI_SIZE, (ray.origin.y / TILE_SIZE) * MINI_SIZE ); rays.graphics.lineTo( (ray.hit.x / TILE_SIZE) * MINI_SIZE, (ray.hit.y / TILE_SIZE) * MINI_SIZE ); for( var x:int = 0; x < COLUMN_SIZE; x++ ) { for( var y:int = yTop; y < yBot; y++ ) { buffer.setPixel(sx+x, y, clrTable[ray.tile-1] >> ( ray.horz ? 1 : 0 )); } } } function castRay(origin:Point, angle):Object { // Return values var rTexel = 0; var rHorz = false; var rTile = 0; var rDist = MAX_DIST + 1; var rMap:Point = new Point(); var rHit:Point = new Point(); // Ray angle and slope var ra = toRad(angle) % ANGLE_360; if( ra < ANGLE_0 ) ra += ANGLE_360; var rs = Math.tan(ra); var rUp = ( ra > ANGLE_0 && ra < ANGLE_180 ); var rRight = ( ra < ANGLE_90 || ra > ANGLE_270 ); // Ray position var rx = 0; var ry = 0; // Ray step values var xa = 0; var ya = 0; // Ray position, in map coordinates var mx:int = 0; var my:int = 0; var mt:int = 0; // Distance var dx = 0; var dy = 0; var ds = MAX_DIST + 1; // Horizontal intersection if( ra != ANGLE_180 && ra != ANGLE_0 && ra != ANGLE_360 ) { ya = ( rUp ? TILE_SIZE : -TILE_SIZE ); xa = ya / rs; ry = int( origin.y / TILE_SIZE ) * ( TILE_SIZE ) + ( rUp ? TILE_SIZE : -1 ); rx = origin.x + ( ry - origin.y ) / rs; mx = 0; my = 0; while( mx >= 0 && my >= 0 && mx < world.size.x && my < world.size.y ) { mx = int( rx / TILE_SIZE ); my = int( ry / TILE_SIZE ); mt = getMapTile(mx,my); if( mt > 0 && mt < 9 ) { dx = rx - origin.x; dy = ry - origin.y; ds = ( dx * dx ) + ( dy * dy ); if( rDist >= MAX_DIST || ds < rDist ) { rDist = ds; rTile = mt; rMap.x = mx; rMap.y = my; rHit.x = rx; rHit.y = ry; rHorz = true; rTexel = int(rx % TILE_SIZE) } break; } rx += xa; ry += ya; } } // Vertical intersection if( ra != ANGLE_90 && ra != ANGLE_270 ) { xa = ( rRight ? TILE_SIZE : -TILE_SIZE ); ya = xa * rs; rx = int( origin.x / TILE_SIZE ) * ( TILE_SIZE ) + ( rRight ? TILE_SIZE : -1 ); ry = origin.y + ( rx - origin.x ) * rs; mx = 0; my = 0; while( mx >= 0 && my >= 0 && mx < world.size.x && my < world.size.y ) { mx = int( rx / TILE_SIZE ); my = int( ry / TILE_SIZE ); mt = getMapTile(mx,my); if( mt > 0 && mt < 9 ) { dx = rx - origin.x; dy = ry - origin.y; ds = ( dx * dx ) + ( dy * dy ); if( rDist >= MAX_DIST || ds < rDist ) { rDist = ds; rTile = mt; rMap.x = mx; rMap.y = my; rHit.x = rx; rHit.y = ry; rHorz = false; rTexel = int(ry % TILE_SIZE); } break; } rx += xa; ry += ya; } } return { angle: angle, distance: Math.sqrt(rDist), hit: rHit, map: rMap, tile: rTile, horz: rHorz, origin: origin, texel: rTexel }; }

    Read the article

  • C# Performance Pitfall – Interop Scenarios Change the Rules

    - by Reed
    C# and .NET, overall, really do have fantastic performance in my opinion.  That being said, the performance characteristics dramatically differ from native programming, and take some relearning if you’re used to doing performance optimization in most other languages, especially C, C++, and similar.  However, there are times when revisiting tricks learned in native code play a critical role in performance optimization in C#. I recently ran across a nasty scenario that illustrated to me how dangerous following any fixed rules for optimization can be… The rules in C# when optimizing code are very different than C or C++.  Often, they’re exactly backwards.  For example, in C and C++, lifting a variable out of loops in order to avoid memory allocations often can have huge advantages.  If some function within a call graph is allocating memory dynamically, and that gets called in a loop, it can dramatically slow down a routine. This can be a tricky bottleneck to track down, even with a profiler.  Looking at the memory allocation graph is usually the key for spotting this routine, as it’s often “hidden” deep in call graph.  For example, while optimizing some of my scientific routines, I ran into a situation where I had a loop similar to: for (i=0; i<numberToProcess; ++i) { // Do some work ProcessElement(element[i]); } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } This loop was at a fairly high level in the call graph, and often could take many hours to complete, depending on the input data.  As such, any performance optimization we could achieve would be greatly appreciated by our users. After a fair bit of profiling, I noticed that a couple of function calls down the call graph (inside of ProcessElement), there was some code that effectively was doing: // Allocate some data required DataStructure* data = new DataStructure(num); // Call into a subroutine that passed around and manipulated this data highly CallSubroutine(data); // Read and use some values from here double values = data->Foo; // Cleanup delete data; // ... return bar; Normally, if “DataStructure” was a simple data type, I could just allocate it on the stack.  However, it’s constructor, internally, allocated it’s own memory using new, so this wouldn’t eliminate the problem.  In this case, however, I could change the call signatures to allow the pointer to the data structure to be passed into ProcessElement and through the call graph, allowing the inner routine to reuse the same “data” memory instead of allocating.  At the highest level, my code effectively changed to something like: DataStructure* data = new DataStructure(numberToProcess); for (i=0; i<numberToProcess; ++i) { // Do some work ProcessElement(element[i], data); } delete data; Granted, this dramatically reduced the maintainability of the code, so it wasn’t something I wanted to do unless there was a significant benefit.  In this case, after profiling the new version, I found that it increased the overall performance dramatically – my main test case went from 35 minutes runtime down to 21 minutes.  This was such a significant improvement, I felt it was worth the reduction in maintainability. In C and C++, it’s generally a good idea (for performance) to: Reduce the number of memory allocations as much as possible, Use fewer, larger memory allocations instead of many smaller ones, and Allocate as high up the call stack as possible, and reuse memory I’ve seen many people try to make similar optimizations in C# code.  For good or bad, this is typically not a good idea.  The garbage collector in .NET completely changes the rules here. In C#, reallocating memory in a loop is not always a bad idea.  In this scenario, for example, I may have been much better off leaving the original code alone.  The reason for this is the garbage collector.  The GC in .NET is incredibly effective, and leaving the allocation deep inside the call stack has some huge advantages.  First and foremost, it tends to make the code more maintainable – passing around object references tends to couple the methods together more than necessary, and overall increase the complexity of the code.  This is something that should be avoided unless there is a significant reason.  Second, (unlike C and C++) memory allocation of a single object in C# is normally cheap and fast.  Finally, and most critically, there is a large advantage to having short lived objects.  If you lift a variable out of the loop and reuse the memory, its much more likely that object will get promoted to Gen1 (or worse, Gen2).  This can cause expensive compaction operations to be required, and also lead to (at least temporary) memory fragmentation as well as more costly collections later. As such, I’ve found that it’s often (though not always) faster to leave memory allocations where you’d naturally place them – deep inside of the call graph, inside of the loops.  This causes the objects to stay very short lived, which in turn increases the efficiency of the garbage collector, and can dramatically improve the overall performance of the routine as a whole. In C#, I tend to: Keep variable declarations in the tightest scope possible Declare and allocate objects at usage While this tends to cause some of the same goals (reducing unnecessary allocations, etc), the goal here is a bit different – it’s about keeping the objects rooted for as little time as possible in order to (attempt) to keep them completely in Gen0, or worst case, Gen1.  It also has the huge advantage of keeping the code very maintainable – objects are used and “released” as soon as possible, which keeps the code very clean.  It does, however, often have the side effect of causing more allocations to occur, but keeping the objects rooted for a much shorter time. Now – nowhere here am I suggesting that these rules are hard, fast rules that are always true.  That being said, my time spent optimizing over the years encourages me to naturally write code that follows the above guidelines, then profile and adjust as necessary.  In my current project, however, I ran across one of those nasty little pitfalls that’s something to keep in mind – interop changes the rules. In this case, I was dealing with an API that, internally, used some COM objects.  In this case, these COM objects were leading to native allocations (most likely C++) occurring in a loop deep in my call graph.  Even though I was writing nice, clean managed code, the normal managed code rules for performance no longer apply.  After profiling to find the bottleneck in my code, I realized that my inner loop, a innocuous looking block of C# code, was effectively causing a set of native memory allocations in every iteration.  This required going back to a “native programming” mindset for optimization.  Lifting these variables and reusing them took a 1:10 routine down to 0:20 – again, a very worthwhile improvement. Overall, the lessons here are: Always profile if you suspect a performance problem – don’t assume any rule is correct, or any code is efficient just because it looks like it should be Remember to check memory allocations when profiling, not just CPU cycles Interop scenarios often cause managed code to act very differently than “normal” managed code. Native code can be hidden very cleverly inside of managed wrappers

    Read the article

  • Oracle Flashback Technologies - Overview

    - by Sridhar_R-Oracle
    Oracle Flashback Technologies - IntroductionIn his May 29th 2014 blog, my colleague Joe Meeks introduced Oracle Maximum Availability Architecture (MAA) and discussed both planned and unplanned outages. Let’s take a closer look at unplanned outages. These can be caused by physical failures (e.g., server, storage, network, file deletion, physical corruption, site failures) or by logical failures – cases where all components and files are physically available, but data is incorrect or corrupt. These logical failures are usually caused by human errors or application logic errors. This blog series focuses on these logical errors – what causes them and how to address and recover from them using Oracle Database Flashback. In this introductory blog post, I’ll provide an overview of the Oracle Database Flashback technologies and will discuss the features in detail in future blog posts. Let’s get started. We are all human beings (unless a machine is reading this), and making mistakes is a part of what we do…often what we do best!  We “fat finger”, we spill drinks on keyboards, unplug the wrong cables, etc.  In addition, many of us, in our lives as DBAs or developers, must have observed, caused, or corrected one or more of the following unpleasant events: Accidentally updated a table with wrong values !! Performed a batch update that went wrong - due to logical errors in the code !! Dropped a table !! How do DBAs typically recover from these types of errors? First, data needs to be restored and recovered to the point-in-time when the error occurred (incomplete or point-in-time recovery).  Moreover, depending on the type of fault, it’s possible that some services – or even the entire database – would have to be taken down during the recovery process.Apart from error conditions, there are other questions that need to be addressed as part of the investigation. For example, what did the data look like in the morning, prior to the error? What were the various changes to the row(s) between two timestamps? Who performed the transaction and how can it be reversed?  Oracle Database includes built-in Flashback technologies, with features that address these challenges and questions, and enable you to perform faster, easier, and convenient recovery from logical corruptions. HistoryFlashback Query, the first Flashback Technology, was introduced in Oracle 9i. It provides a simple, powerful and completely non-disruptive mechanism for data verification and recovery from logical errors, and enables users to view the state of data at a previous point in time.Flashback Technologies were further enhanced in Oracle 10g, to provide fast, easy recovery at the database, table, row, and even at a transaction level.Oracle Database 11g introduced an innovative method to manage and query long-term historical data with Flashback Data Archive. The 11g release also introduced Flashback Transaction, which provides an easy, one-step operation to back out a transaction. Oracle Database versions 11.2.0.2 and beyond further enhanced the performance of these features. Note that all the features listed here work without requiring any kind of restore operation.In addition, Flashback features are fully supported with the new multi-tenant capabilities introduced with Oracle Database 12c, Flashback Features Oracle Flashback Database enables point-in-time-recovery of the entire database without requiring a traditional restore and recovery operation. It rewinds the entire database to a specified point in time in the past by undoing all the changes that were made since that time.Oracle Flashback Table enables an entire table or a set of tables to be recovered to a point in time in the past.Oracle Flashback Drop enables accidentally dropped tables and all dependent objects to be restored.Oracle Flashback Query enables data to be viewed at a point-in-time in the past. This feature can be used to view and reconstruct data that was lost due to unintentional change(s) or deletion(s). This feature can also be used to build self-service error correction into applications, empowering end-users to undo and correct their errors.Oracle Flashback Version Query offers the ability to query the historical changes to data between two points in time or system change numbers (SCN) Oracle Flashback Transaction Query enables changes to be examined at the transaction level. This capability can be used to diagnose problems, perform analysis, audit transactions, and even revert the transaction by undoing SQLOracle Flashback Transaction is a procedure used to back-out a transaction and its dependent transactions.Flashback technologies eliminate the need for a traditional restore and recovery process to fix logical corruptions or make enquiries. Using these technologies, you can recover from the error in the same amount of time it took to generate the error. All the Flashback features can be accessed either via SQL command line (or) via Enterprise Manager.  Most of the Flashback technologies depend on the available UNDO to retrieve older data. The following table describes the various Flashback technologies: their purpose, dependencies and situations where each individual technology can be used.   Example Syntax Error investigation related:The purpose is to investigate what went wrong and what the values were at certain points in timeFlashback Queries  ( select .. as of SCN | Timestamp )   - Helps to see the value of a row/set of rows at a point in timeFlashback Version Queries  ( select .. versions between SCN | Timestamp and SCN | Timestamp)  - Helps determine how the value evolved between certain SCNs or between timestamps Flashback Transaction Queries (select .. XID=)   - Helps to understand how the transaction caused the changes.Error correction related:The purpose is to fix the error and correct the problems,Flashback Table  (flashback table .. to SCN | Timestamp)  - To rewind the table to a particular timestamp or SCN to reverse unwanted updates Flashback Drop (flashback table ..  to before drop )  - To undrop or undelete a table Flashback Database (flashback database to SCN  | Restore Point )  - This is the rewind button for Oracle databases. You can revert the entire database to a particular point in time. It is a fast way to perform a PITR (point-in-time recovery). Flashback Transaction (DBMS_FLASHBACK.TRANSACTION_BACKOUT(XID..))  - To reverse a transaction and its related transactions Advanced use cases Flashback technology is integrated into Oracle Recovery Manager (RMAN) and Oracle Data Guard. So, apart from the basic use cases mentioned above, the following use cases are addressed using Oracle Flashback. Block Media recovery by RMAN - to perform block level recovery Snapshot Standby - where the standby is temporarily converted to a read/write environment for testing, backup, or migration purposes Re-instate old primary in a Data Guard environment – this avoids the need to restore an old backup and perform a recovery to make it a new standby. Guaranteed Restore Points - to bring back the entire database to an older point-in-time in a guaranteed way. and so on..I hope this introductory overview helps you understand how Flashback features can be used to investigate and recover from logical errors.  As mentioned earlier, I will take a deeper-dive into to some of the critical Flashback features in my upcoming blogs and address common use cases.

    Read the article

  • Automating deployments with the SQL Compare command line

    - by Jonathan Hickford
    In my previous article, “Five Tips to Get Your Organisation Releasing Software Frequently” I looked at how teams can automate processes to speed up release frequency. In this post, I’m looking specifically at automating deployments using the SQL Compare command line. SQL Compare compares SQL Server schemas and deploys the differences. It works very effectively in scenarios where only one deployment target is required – source and target databases are specified, compared, and a change script is automatically generated and applied. But if multiple targets exist, and pressure to increase the frequency of releases builds, this solution quickly becomes unwieldy.   This is where SQL Compare’s command line comes into its own. I’ve put together a PowerShell script that loops through the Servers table and pulls out the server and database, these are then passed to sqlcompare.exe to be used as target parameters. In the example the source database is a scripts folder, a folder structure of scripted-out database objects used by both SQL Source Control and SQL Compare. The script can easily be adapted to use schema snapshots.     -- Create a DeploymentTargets database and a Servers table CREATE DATABASE DeploymentTargets GO USE DeploymentTargets GO CREATE TABLE [dbo].[Servers]( [id] [int] IDENTITY(1,1) NOT NULL, [serverName] [nvarchar](50) NULL, [environment] [nvarchar](50) NULL, [databaseName] [nvarchar](50) NULL, CONSTRAINT [PK_Servers] PRIMARY KEY CLUSTERED ([id] ASC) ) GO -- Now insert your target server and database details INSERT INTO dbo.Servers ( serverName , environment , databaseName) VALUES ( N'myserverinstance' , N'myenvironment1' , N'mydb1') INSERT INTO dbo.Servers ( serverName , environment , databaseName) VALUES ( N'myserverinstance' , N'myenvironment2' , N'mydb2') Here’s the PowerShell script you can adapt for yourself as well. # We're holding the server names and database names that we want to deploy to in a database table. # We need to connect to that server to read these details $serverName = "" $databaseName = "DeploymentTargets" $authentication = "Integrated Security=SSPI" #$authentication = "User Id=xxx;PWD=xxx" # If you are using database authentication instead of Windows authentication. # Path to the scripts folder we want to deploy to the databases $scriptsPath = "SimpleTalk" # Path to SQLCompare.exe $SQLComparePath = "C:\Program Files (x86)\Red Gate\SQL Compare 10\sqlcompare.exe" # Create SQL connection string, and connection $ServerConnectionString = "Data Source=$serverName;Initial Catalog=$databaseName;$authentication" $ServerConnection = new-object system.data.SqlClient.SqlConnection($ServerConnectionString); # Create a Dataset to hold the DataTable $dataSet = new-object "System.Data.DataSet" "ServerList" # Create a query $query = "SET NOCOUNT ON;" $query += "SELECT serverName, environment, databaseName " $query += "FROM dbo.Servers; " # Create a DataAdapter to populate the DataSet with the results $dataAdapter = new-object "System.Data.SqlClient.SqlDataAdapter" ($query, $ServerConnection) $dataAdapter.Fill($dataSet) | Out-Null # Close the connection $ServerConnection.Close() # Populate the DataTable $dataTable = new-object "System.Data.DataTable" "Servers" $dataTable = $dataSet.Tables[0] #For every row in the DataTable $dataTable | FOREACH-OBJECT { "Server Name: $($_.serverName)" "Database Name: $($_.databaseName)" "Environment: $($_.environment)" # Compare the scripts folder to the database and synchronize the database to match # NB. Have set SQL Compare to abort on medium level warnings. $arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/AbortOnWarnings:Medium") # + @("/sync" ) # Commented out the 'sync' parameter for safety, write-host $arguments & $SQLComparePath $arguments "Exit Code: $LASTEXITCODE" # Some interesting variations # Check that every database matches a folder. # For example this might be a pre-deployment step to validate everything is at the same baseline state. # Or a post deployment script to validate the deployment worked. # An exit code of 0 means the databases are identical. # # $arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/Assertidentical") # Generate a report of the difference between the folder and each database. Generate a SQL update script for each database. # For example use this after the above to generate upgrade scripts for each database # Examine the warnings and the HTML diff report to understand how the script will change objects # #$arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/ScriptFile:update_$($_.environment+"_"+$_.databaseName).sql", "/report:update_$($_.environment+"_"+$_.databaseName).html" , "/reportType:Interactive", "/showWarnings", "/include:Identical") } It’s worth noting that the above example generates the deployment scripts dynamically. This approach should be problem-free for the vast majority of changes, but it is still good practice to review and test a pre-generated deployment script prior to deployment. An alternative approach would be to pre-generate a single deployment script using SQL Compare, and run this en masse to multiple targets programmatically using sqlcmd, or using a tool like SQL Multi Script.  You can use the /ScriptFile, /report, and /showWarnings flags to generate change scripts, difference reports and any warnings.  See the commented out example in the PowerShell: #$arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/ScriptFile:update_$($_.environment+"_"+$_.databaseName).sql", "/report:update_$($_.environment+"_"+$_.databaseName).html" , "/reportType:Interactive", "/showWarnings", "/include:Identical") There is a drawback of running a pre-generated deployment script; it assumes that a given database target hasn’t drifted from its expected state. Often there are (rightly or wrongly) many individuals within an organization who have permissions to alter the production database, and changes can therefore be made outside of the prescribed development processes. The consequence is that at deployment time, the applied script has been validated against a target that no longer represents reality. The solution here would be to add a check for drift prior to running the deployment script. This is achieved by using sqlcompare.exe to compare the target against the expected schema snapshot using the /Assertidentical flag. Should this return any differences (sqlcompare.exe Exit Code 79), a drift report is outputted instead of executing the deployment script.  See the commented out example. # $arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/Assertidentical") Any checks and processes that should be undertaken prior to a manual deployment, should also be happen during an automated deployment. You might think about triggering backups prior to deployment – even better, automate the verification of the backup too.   You can use SQL Compare’s command line interface along with PowerShell to automate multiple actions and checks that you need in your deployment process. Automation is a practical solution where multiple targets and a higher release cadence come into play. As we know, with great power comes great responsibility – responsibility to ensure that the necessary checks are made so deployments remain trouble-free.  (The code sample supplied in this post automates the simple dynamic deployment case – if you are considering more advanced automation, e.g. the drift checks, script generation, deploying to large numbers of targets and backup/verification, please email me at [email protected] for further script samples or if you have further questions)

    Read the article

  • Columnstore Case Study #2: Columnstore faster than SSAS Cube at DevCon Security

    - by aspiringgeek
    Preamble This is the second in a series of posts documenting big wins encountered using columnstore indexes in SQL Server 2012 & 2014.  Many of these can be found in my big deck along with details such as internals, best practices, caveats, etc.  The purpose of sharing the case studies in this context is to provide an easy-to-consume quick-reference alternative. See also Columnstore Case Study #1: MSIT SONAR Aggregations Why Columnstore? As stated previously, If we’re looking for a subset of columns from one or a few rows, given the right indexes, SQL Server can do a superlative job of providing an answer. If we’re asking a question which by design needs to hit lots of rows—DW, reporting, aggregations, grouping, scans, etc., SQL Server has never had a good mechanism—until columnstore. Columnstore indexes were introduced in SQL Server 2012. However, they're still largely unknown. Some adoption blockers existed; yet columnstore was nonetheless a game changer for many apps.  In SQL Server 2014, potential blockers have been largely removed & they're going to profoundly change the way we interact with our data.  The purpose of this series is to share the performance benefits of columnstore & documenting columnstore is a compelling reason to upgrade to SQL Server 2014. The Customer DevCon Security provides home & business security services & has been in business for 135 years. I met DevCon personnel while speaking to the Utah County SQL User Group on 20 February 2012. (Thanks to TJ Belt (b|@tjaybelt) & Ben Miller (b|@DBADuck) for the invitation which serendipitously coincided with the height of ski season.) The App: DevCon Security Reporting: Optimized & Ad Hoc Queries DevCon users interrogate a SQL Server 2012 Analysis Services cube via SSRS. In addition, the SQL Server 2012 relational back end is the target of ad hoc queries; this DW back end is refreshed nightly during a brief maintenance window via conventional table partition switching. SSRS, SSAS, & MDX Conventional relational structures were unable to provide adequate performance for user interaction for the SSRS reports. An SSAS solution was implemented requiring personnel to ramp up technically, including learning enough MDX to satisfy requirements. Ad Hoc Queries Even though the fact table is relatively small—only 22 million rows & 33GB—the table was a typical DW table in terms of its width: 137 columns, any of which could be the target of ad hoc interrogation. As is common in DW reporting scenarios such as this, it is often nearly to optimize for such queries using conventional indexing. DevCon DBAs & developers attended PASS 2012 & were introduced to the marvels of columnstore in a session presented by Klaus Aschenbrenner (b|@Aschenbrenner) The Details Classic vs. columnstore before-&-after metrics are impressive. Scenario   Conventional Structures   Columnstore   Δ SSRS via SSAS 10 - 12 seconds 1 second >10x Ad Hoc 5-7 minutes (300 - 420 seconds) 1 - 2 seconds >100x Here are two charts characterizing this data graphically.  The first is a linear representation of Report Duration (in seconds) for Conventional Structures vs. Columnstore Indexes.  As is so often the case when we chart such significant deltas, the linear scale doesn’t expose some the dramatically improved values corresponding to the columnstore metrics.  Just to make it fair here’s the same data represented logarithmically; yet even here the values corresponding to 1 –2 seconds aren’t visible.  The Wins Performance: Even prior to columnstore implementation, at 10 - 12 seconds canned report performance against the SSAS cube was tolerable. Yet the 1 second performance afterward is clearly better. As significant as that is, imagine the user experience re: ad hoc interrogation. The difference between several minutes vs. one or two seconds is a game changer, literally changing the way users interact with their data—no mental context switching, no wondering when the results will appear, no preoccupation with the spinning mind-numbing hurry-up-&-wait indicators.  As we’ve commonly found elsewhere, columnstore indexes here provided performance improvements of one, two, or more orders of magnitude. Simplified Infrastructure: Because in this case a nonclustered columnstore index on a conventional DW table was faster than an Analysis Services cube, the entire SSAS infrastructure was rendered superfluous & was retired. PASS Rocks: Once again, the value of attending PASS is proven out. The trip to Charlotte combined with eager & enquiring minds let directly to this success story. Find out more about the next PASS Summit here, hosted this year in Seattle on November 4 - 7, 2014. DevCon BI Team Lead Nathan Allan provided this unsolicited feedback: “What we found was pretty awesome. It has been a game changer for us in terms of the flexibility we can offer people that would like to get to the data in different ways.” Summary For DW, reports, & other BI workloads, columnstore often provides significant performance enhancements relative to conventional indexing.  I have documented here, the second in a series of reports on columnstore implementations, results from DevCon Security, a live customer production app for which performance increased by factors of from 10x to 100x for all report queries, including canned queries as well as reducing time for results for ad hoc queries from 5 - 7 minutes to 1 - 2 seconds. As a result of columnstore performance, the customer retired their SSAS infrastructure. I invite you to consider leveraging columnstore in your own environment. Let me know if you have any questions.

    Read the article

  • Looking into the JQuery Image Zoom Plugin

    - by nikolaosk
    I have been using JQuery for a couple of years now and it has helped me to solve many problems on the client side of web development.  You can find all my posts about JQuery in this link. In this post I will be providing you with a hands-on example on the JQuery Image Zoom Plugin.If you want you can have a look at this post, where I describe the JQuery Cycle Plugin.You can find another post of mine talking about the JQuery Carousel Lite Plugin here.I will be writing more posts regarding the most commonly used JQuery Plugins. I have been using extensively this plugin in my websites.You can use this plugin to move mouse around an image and see a zoomed in version of a portion of it. In this hands-on example I will be using Expression Web 4.0.This application is not a free application. You can use any HTML editor you like. You can use Visual Studio 2012 Express edition. You can download it here.  You can download this plugin from this link I launch Expression Web 4.0 and then I type the following HTML markup (I am using HTML 5) <html lang="en">  <head>    <title>Liverpool Legends</title>        <meta http-equiv="Content-Type" content="text/html;charset=utf-8" >        <link rel="stylesheet" type="text/css" href="style.css">        <script type="text/javascript" src="jquery-1.8.3.min.js"> </script>     <script type="text/javascript" src="jqzoom.pack.1.0.1.js"></script>        <script type="text/javascript">        $(function () {            $(".nicezoom").jqzoom();        });    </script>       </head>  <body>    <header>        <h1>Liverpool Legends</h1>    </header>        <div id="main">            <a href="championsofeurope-large.jpg" class="nicezoom" title="Champions">        <img src="championsofeurope.jpg"  title="Champions">    </a>          </div>            <footer>        <p>All Rights Reserved</p>      </footer>     </body>  </html>   This is a very simple markup. I have added one large and one small image (make sure you use your own when trying this example) I have added references to the JQuery library (current version is 1.8.3) and the JQuery Image Zoom Plugin. Then I add 2 images in the main div element.Note the class nicezoom inside the href element. The Javascript code that makes it all happen follows.    <script type="text/javascript">        $(function () {            $(".nicezoom").jqzoom();        });    </script>     It couldn't be any simpler than that. I view my simple in Internet Explorer 10 and it works as expected. I have tested this simple solution in all major browsers and it works fine.Inside the head section we can add another Javascript script utilising some more options regarding the zoom plugin.   <script type="text/javascript">            $(function () {        var options = {                  zoomType: 'standard',                  lens:true,                  preloadImages: true,                  alwaysOn:false,                  zoomWidth: 400,                  zoomHeight: 350,                  xOffset:190,                  yOffset:80,                  position:'right'                          };          $('.nicezoom').jqzoom(options);      });         </script> I would like to explain briefly what some of those options mean. zoomType - Other admitted option values are 'reverse','drag','innerzoom' zoomWidth - The popup window width showing the zoomed area zoomHeight - The popup window height showing the zoomed area xOffset - The popup window x offset from the small image.  yOffset - The popup window y offset from the small image.  position - The popup window position.Admitted values:'right' ,'left' ,'top' ,'bottom' preloadImages - if set to true,jqzoom will preload large images. You can test it yourself and see the results in your favorite browser. Hope it helps!!!

    Read the article

  • Creating New Scripts Dynamically in Lua

    - by bazola
    Right now this is just a crazy idea that I had, but I was able to implement the code and get it working properly. I am not entirely sure of what the use cases would be just yet. What this code does is create a new Lua script file in the project directory. The ScriptWriter takes as arguments the file name, a table containing any arguments that the script should take when created, and a table containing any instance variables to create by default. My plan is to extend this code to create new functions based on inputs sent in during its creation as well. What makes this cool is that the new file is both generated and loaded dynamically on the fly. Theoretically you could get this code to generate and load any script imaginable. One use case I can think of is an AI that creates scripts to map out it's functions, and creates new scripts for new situations or environments. At this point, this is all theoretical, though. Here is the test code that is creating the new script and then immediately loading it and calling functions from it: function Card:doScriptWriterThing() local scriptName = "ScriptIAmMaking" local scripter = scriptWriter:new(scriptName, {"argumentName"}, {name = "'test'", one = 1}) scripter:makeFileForLoadedSettings() local loadedScript = require (scriptName) local scriptInstance = loadedScript:new("sayThis") print(scriptInstance:get_name()) --will print test print(scriptInstance:get_one()) -- will print 1 scriptInstance:set_one(10000) print(scriptInstance:get_one()) -- will print 10000 print(scriptInstance:get_argumentName()) -- will print sayThis scriptInstance:set_argumentName("saySomethingElse") print(scriptInstance:get_argumentName()) --will print saySomethingElse end Here is ScriptWriter.lua local ScriptWriter = {} local twoSpaceIndent = " " local equalsWithSpaces = " = " local newLine = "\n" --scriptNameToCreate must be a string --argumentsForNew and instanceVariablesToCreate must be tables and not nil function ScriptWriter:new(scriptNameToCreate, argumentsForNew, instanceVariablesToCreate) local instance = setmetatable({}, { __index = self }) instance.name = scriptNameToCreate instance.newArguments = argumentsForNew instance.instanceVariables = instanceVariablesToCreate instance.stringList = {} return instance end function ScriptWriter:makeFileForLoadedSettings() self:buildInstanceMetatable() self:buildInstanceCreationMethod() self:buildSettersAndGetters() self:buildReturn() self:writeStringsToFile() end --very first line of any script that will have instances function ScriptWriter:buildInstanceMetatable() table.insert(self.stringList, "local " .. self.name .. " = {}" .. newLine) table.insert(self.stringList, newLine) end --every script made this way needs a new method to create its instances function ScriptWriter:buildInstanceCreationMethod() --new() function declaration table.insert(self.stringList, ("function " .. self.name .. ":new(")) self:buildNewArguments() table.insert(self.stringList, ")" .. newLine) --first line inside :new() function table.insert(self.stringList, twoSpaceIndent .. "local instance = setmetatable({}, { __index = self })" .. newLine) --add designated arguments inside :new() self:buildNewArgumentVariables() --create the instance variables with the loaded values for key,value in pairs(self.instanceVariables) do table.insert(self.stringList, twoSpaceIndent .. "instance." .. key .. equalsWithSpaces .. value .. newLine) end --close the :new() function table.insert(self.stringList, twoSpaceIndent .. "return instance" .. newLine) table.insert(self.stringList, "end" .. newLine) table.insert(self.stringList, newLine) end function ScriptWriter:buildNewArguments() --if there are arguments for :new(), add them for key,value in ipairs(self.newArguments) do table.insert(self.stringList, value) table.insert(self.stringList, ", ") end if next(self.newArguments) ~= nil then --makes sure the table is not empty first table.remove(self.stringList) --remove the very last element, which will be the extra ", " end end function ScriptWriter:buildNewArgumentVariables() --add the designated arguments to :new() for key, value in ipairs(self.newArguments) do table.insert(self.stringList, twoSpaceIndent .. "instance." .. value .. equalsWithSpaces .. value .. newLine) end end --the instance variables need separate code because their names have to be the key and not the argument name function ScriptWriter:buildSettersAndGetters() for key,value in ipairs(self.newArguments) do self:buildArgumentSetter(value) self:buildArgumentGetter(value) table.insert(self.stringList, newLine) end for key,value in pairs(self.instanceVariables) do self:buildInstanceVariableSetter(key, value) self:buildInstanceVariableGetter(key, value) table.insert(self.stringList, newLine) end end --code for arguments passed in function ScriptWriter:buildArgumentSetter(variable) table.insert(self.stringList, "function " .. self.name .. ":set_" .. variable .. "(newValue)" .. newLine) table.insert(self.stringList, twoSpaceIndent .. "self." .. variable .. equalsWithSpaces .. "newValue" .. newLine) table.insert(self.stringList, "end" .. newLine) end function ScriptWriter:buildArgumentGetter(variable) table.insert(self.stringList, "function " .. self.name .. ":get_" .. variable .. "()" .. newLine) table.insert(self.stringList, twoSpaceIndent .. "return " .. "self." .. variable .. newLine) table.insert(self.stringList, "end" .. newLine) end --code for instance variable values passed in function ScriptWriter:buildInstanceVariableSetter(key, variable) table.insert(self.stringList, "function " .. self.name .. ":set_" .. key .. "(newValue)" .. newLine) table.insert(self.stringList, twoSpaceIndent .. "self." .. key .. equalsWithSpaces .. "newValue" .. newLine) table.insert(self.stringList, "end" .. newLine) end function ScriptWriter:buildInstanceVariableGetter(key, variable) table.insert(self.stringList, "function " .. self.name .. ":get_" .. key .. "()" .. newLine) table.insert(self.stringList, twoSpaceIndent .. "return " .. "self." .. key .. newLine) table.insert(self.stringList, "end" .. newLine) end --last line of any script that will have instances function ScriptWriter:buildReturn() table.insert(self.stringList, "return " .. self.name) end function ScriptWriter:writeStringsToFile() local fileName = (self.name .. ".lua") file = io.open(fileName, 'w') for key,value in ipairs(self.stringList) do file:write(value) end file:close() end return ScriptWriter And here is what the code provided will generate: local ScriptIAmMaking = {} function ScriptIAmMaking:new(argumentName) local instance = setmetatable({}, { __index = self }) instance.argumentName = argumentName instance.name = 'test' instance.one = 1 return instance end function ScriptIAmMaking:set_argumentName(newValue) self.argumentName = newValue end function ScriptIAmMaking:get_argumentName() return self.argumentName end function ScriptIAmMaking:set_name(newValue) self.name = newValue end function ScriptIAmMaking:get_name() return self.name end function ScriptIAmMaking:set_one(newValue) self.one = newValue end function ScriptIAmMaking:get_one() return self.one end return ScriptIAmMaking All of this is generated with these calls: local scripter = scriptWriter:new(scriptName, {"argumentName"}, {name = "'test'", one = 1}) scripter:makeFileForLoadedSettings() I am not sure if I am correct that this could be useful in certain situations. What I am looking for is feedback on the readability of the code, and following Lua best practices. I would also love to hear whether this approach is a valid one, and whether the way that I have done things will be extensible.

    Read the article

  • Expectations + Rewards = Innovation

    - by D'Arcy Lussier
    “Innovation” is a heavy word. We regard those that embrace it as “Innovators”. We describe organizations as being “Innovative”. We hold those associated with the word in high regard, even though its dictionary definition is very simple: Introducing something new. What our culture has done is wrapped Innovation in white robes and a gold crown. Innovation is rarely just introducing something new. Innovations and innovators are typically associated with other terms: groundbreaking, genius, industry-changing, creative, leading. Being a true innovator and creating innovations are a big deal, and something companies try to strive for…or at least say they strive for. There’s huge value in being recognized as an innovator in an industry, since the idea is that innovation equates to increased profitability. IBM ran an ad a few years back that showed what their view of innovation is: “The point of innovation is to make actual money.” If the money aspect makes you feel uneasy, consider it another way: the point of innovation is to <insert payoff here>. Companies that innovate will be more successful. Non-profits that innovate can better serve their target clients. Governments that innovate can better provide services to their citizens. True innovation is not easy to come by though. As with anything in business, how well an organization will innovate is reliant on the employees it retains, the expectations placed on those employees, and the rewards available to them. In a previous blog post I talked about one formula: Right Employees + Happy Employees = Productive Employees I want to introduce a new one, that builds upon the previous one: Expectations + Rewards = Innovation  The level of innovation your organization will realize is directly associated with the expectations you place on your staff and the rewards you make available to them. Expectations We may feel uncomfortable with the idea of placing expectations on our staff, mainly because expectation has somewhat of a negative or cold connotation to it: “I expect you to act this way or else!” The problem is in the or-else part…we focus on the negative aspects of failing to meet expectations instead of looking at the positive side. “I expect you to act this way because it will produce <insert benefit here>”. Expectations should not be set to punish but instead be set to ensure quality. At a recent conference I spoke with some Microsoft employees who told me that you have five years from starting with the company to reach a “Senior” level. If you don’t, then you’re let go. The expectation Microsoft placed on their staff is that they should be working towards improving themselves, taking more responsibility, and thus ensure that there is a constant level of quality in the workforce. Rewards Let me be clear: a paycheck is not a reward. A paycheck is simply the employer’s responsibility in the employee/employer relationship. A paycheck will never be the key motivator to drive innovation. Offering employees something over and above their required compensation can spur them to greater performance and achievement. Working in the food service industry, this tactic was used again and again: whoever has the highest sales over lunch will receive a free lunch/gift certificate/entry into a draw/etc. There was something to strive for, to try beyond the baseline of what our serving jobs were. It was through this that innovative sales techniques would be tried and honed, with key servers being top sellers time and time again. At a code camp I spoke at, I was amazed to see that all the employees from one company receive $100 Visa gift cards as a thank you for taking time to speak. Again, offering something over and above that can give that extra push for employees. Rewards work. But what about the fairness angle? In the restaurant example I gave, there were servers that would never win the competition. They just weren’t good enough at selling and never seemed to get better. So should those that did work at performing better and produce more sales for the restaurant not get rewarded because those who weren’t working at performing better might get upset? Of course not! Organizations succeed because of their top performers and those that strive to join their ranks. The Expectation/Reward Graph While the Expectations + Rewards = Innovation formula may seem like a simple mathematics formula, there’s much more going under the hood. In fact there are three different outcomes that could occur based on what you put in as values for Expectations and Rewards. Consider the graph below and the descriptions that follow: Disgruntled – High Expectation, Low Reward I worked at a company where the mantra was “Company First, Because We Pay You”. Even today I still hear stories of how this sentiment continues to be perpetuated: They provide you a paycheck and a means to live, therefore you should always put them as your top priority. Of course, this is a huge imbalance in the expectation/reward equation. Why would anyone willingly meet high expectations of availability, workload, deadlines, etc. when there is no reward other than a paycheck to show for it? Remember: paychecks are not rewards! Instead, you see employees be disgruntled which not only affects the level of production but also the level of quality within an organization. It also means that you see higher turnover. Complacent – Low Expectation, Low Reward Complacency is a systemic problem that typically exists throughout all levels of an organization. With no real expectations or rewards, nobody needs to excel. In fact, those that do try to innovate, improve, or introduce new things into the organization might be shunned or pushed out by the rest of the staff who are just doing things the same way they’ve always done it. The bigger issue for the organization with low/low values is that at best they’ll never grow beyond their current size (and may shrink actually), and at worst will cease to exist. Entitled – Low Expectation, High Reward It’s one thing to say you have the best people and reward them as such, but its another thing to actually have the best people and reward them as such. Organizations with Entitled employees are the former: their organization provides them with all types of comforts, benefits, and perks. But there’s no requirement before the rewards are dolled out, and there’s no short-list of who receives the rewards. Everyone in the company is treated the same and is given equal share of the spoils. Entitlement is actually almost identical with Complacency with one notable difference: just try to introduce higher expectations into an entitled organization! Entitled employees have been spoiled for so long that they can’t fathom having rewards taken from them, or having to achieve specific levels of performance before attaining them. Those running the organization also buy in to the Entitled sentiment, feeling that they must persist the same level of comforts to appease their staff…even though the quality of the employee pool may be suspect. Innovative – High Expectation, High Reward Finally we have the Innovative organization which places high expectations but also provides high rewards. This organization gets it: if you truly want the best employees you need to apply equal doses of pressure and praise. Realize that I’m not suggesting crazy overtime or un-realistic working conditions. I do not agree with the “Glengary-Glenross” method of encouragement. But as anyone who follows sports can tell you, the teams that win are the ones where the coaches push their players to be their best; to achieve new levels of performance that they didn’t know they could receive. And the result for the players is more money, fame, and opportunity. It’s in this environment that organizations can focus on innovation – true innovation that builds the business and allows everyone involved to truly benefit. In Closing Organizations love to use the word “Innovation” and its derivatives, but very few actually do innovate. For many, the term has just become another marketing buzzword to lump in with all the other business terms that get overused. But for those organizations that truly get the value of innovation, they will be the ones surging forward while other companies simply fade into the background. And they will be the organizations that expect more from their employees, and give them their just rewards.

    Read the article

  • How Visual Studio 2010 and Team Foundation Server enable Compliance

    - by Martin Hinshelwood
    One of the things that makes Team Foundation Server (TFS) the most powerful Application Lifecycle Management (ALM) platform is the traceability it provides to those that use it. This traceability is crucial to enable many companies to adhere to many of the Compliance regulations to which they are bound (e.g. CFR 21 Part 11 or Sarbanes–Oxley.)   From something as simple as relating Tasks to Check-in’s or being able to see the top 10 files in your codebase that are causing the most Bugs, to identifying which Bugs and Requirements are in which Release. All that information is available and more in TFS. Although all of this tradability is available within TFS you do need to understand that it is not for free. Well… I say that, but if you are using TFS properly you will have this information with no additional work except for firing up the reporting. Using Visual Studio ALM and Team Foundation Server you can relate every line of code changes all the way up to requirements and back down through Test Cases to the Test Results. Figure: The only thing missing is Build In order to build the relationship model below we need to examine how each of the relationships get there. Each member of your team from programmer to tester and Business Analyst to Business have their roll to play to knit this together. Figure: The relationships required to make this work can get a little confusing If Build is added to this to relate Work Items to Builds and with knowledge of which builds are in which environments you can easily identify what is contained within a Release. Figure: How are things progressing Along with the ability to produce the progress and trend reports the tractability that is built into TFS can be used to fulfil most audit requirements out of the box, and augmented to fulfil the rest. In order to understand the relationships, lets look at each of the important Artifacts and how they are associated with each other… Requirements – The root of all knowledge Requirements are the thing that the business cares about delivering. These could be derived as User Stories or Business Requirements Documents (BRD’s) but they should be what the Business asks for. Requirements can be related to many of the Artifacts in TFS, so lets look at the model: Figure: If the centre of the world was a requirement We can track which releases Requirements were scheduled in, but this can change over time as more details come to light. Figure: Who edited the Requirement and when There is also the ability to query Work Items based on the History of changed that were made to it. This is particularly important with Requirements. It might not be enough to say what Requirements were completed in a given but also to know which Requirements were ever assigned to a particular release. Figure: Some magic required, but result still achieved As an augmentation to this it is also possible to run a query that shows results from the past, just as if we had a time machine. You can take any Query in the system and add a “Asof” clause at the end to query historical data in the operational store for TFS. select <fields> from WorkItems [where <condition>] [order by <fields>] [asof <date>] Figure: Work Item Query Language (WIQL) format In order to achieve this you do need to save the query as a *.wiql file to your local computer and edit it in notepad, but one imported into TFS you run it any time you want. Figure: Saving Queries locally can be useful All of these Audit features are available throughout the Work Item Tracking (WIT) system within TFS. Tasks – Where the real work gets done Tasks are the work horse of the development team, but they only as useful as Excel if you do not relate them properly to other Artifacts. Figure: The Task Work Item Type has its own relationships Requirements should be broken down into Tasks that the development team work from to build what is required by the business. This may be done by a small dedicated group or by everyone that will be working on the software team but however it happens all of the Tasks create should be a Child of a Requirement Work Item Type. Figure: Tasks are related to the Requirement Tasks should be used to track the day-to-day activities of the team working to complete the software and as such they should be kept simple and short lest developers think they are more trouble than they are worth. Figure: Task Work Item Type has a narrower purpose Although the Task Work Item Type describes the work that will be done the actual development work involves making changes to files that are under Source Control. These changes are bundled together in a single atomic unit called a Changeset which is committed to TFS in a single operation. During this operation developers can associate Work Item with the Changeset. Figure: Tasks are associated with Changesets   Changesets – Who wrote this crap Changesets themselves are just an inventory of the changes that were made to a number of files to complete a Task. Figure: Changesets are linked by Tasks and Builds   Figure: Changesets tell us what happened to the files in Version Control Although comments can be changed after the fact, the inventory and Work Item associations are permanent which allows us to Audit all the way down to the individual change level. Figure: On Check-in you can resolve a Task which automatically associates it Because of this we can view the history on any file within the system and see how many changes have been made and what Changesets they belong to. Figure: Changes are tracked at the File level What would be even more powerful would be if we could view these changes super imposed over the top of the lines of code. Some people call this a blame tool because it is commonly used to find out which of the developers introduced a bug, but it can also be used as another method of Auditing changes to the system. Figure: Annotate shows the lines the Annotate functionality allows us to visualise the relationship between the individual lines of code and the Changesets. In addition to this you can create a Label and apply it to a version of your version control. The problem with Label’s is that they can be changed after they have been created with no tractability. This makes them practically useless for any sort of compliance audit. So what do you use? Branches – And why we need them Branches are a really powerful tool for development and release management, but they are most important for audits. Figure: One way to Audit releases The R1.0 branch can be created from the Label that the Build creates on the R1 line when a Release build was created. It can be created as soon as the Build has been signed of for release. However it is still possible that someone changed the Label between this time and its creation. Another better method can be to explicitly link the Build output to the Build. Builds – Lets tie some more of this together Builds are the glue that helps us enable the next level of tractability by tying everything together. Figure: The dashed pieces are not out of the box but can be enabled When the Build is called and starts it looks at what it has been asked to build and determines what code it is going to get and build. Figure: The folder identifies what changes are included in the build The Build sets a Label on the Source with the same name as the Build, but the Build itself also includes the latest Changeset ID that it will be building. At the end of the Build the Build Agent identifies the new Changesets it is building by looking at the Check-ins that have occurred since the last Build. Figure: What changes have been made since the last successful Build It will then use that information to identify the Work Items that are associated with all of the Changesets Changesets are associated with Build and change the “Integrated In” field of those Work Items . Figure: Find all of the Work Items to associate with The “Integrated In” field of all of the Work Items identified by the Build Agent as being integrated into the completed Build are updated to reflect the Build number that successfully integrated that change. Figure: Now we know which Work Items were completed in a build Now that we can link a single line of code changed all the way back through the Task that initiated the action to the Requirement that started the whole thing and back down to the Build that contains the finished Requirement. But how do we know wither that Requirement has been fully tested or even meets the original Requirements? Test Cases – How we know we are done The only way we can know wither a Requirement has been completed to the required specification is to Test that Requirement. In TFS there is a Work Item type called a Test Case Test Cases enable two scenarios. The first scenario is the ability to track and validate Acceptance Criteria in the form of a Test Case. If you agree with the Business a set of goals that must be met for a Requirement to be accepted by them it makes it both difficult for them to reject a Requirement when it passes all of the tests, but also provides a level of tractability and validation for audit that a feature has been built and tested to order. Figure: You can have many Acceptance Criteria for a single Requirement It is crucial for this to work that someone from the Business has to sign-off on the Test Case moving from the  “Design” to “Ready” states. The Second is the ability to associate an MS Test test with the Test Case thereby tracking the automated test. This is useful in the circumstance when you want to Track a test and the test results of a Unit Test designed to test the existence of and then re-existence of a a Bug. Figure: Associating a Test Case with an automated Test Although it is possible it may not make sense to track the execution of every Unit Test in your system, there are many Integration and Regression tests that may be automated that it would make sense to track in this way. Bug – Lets not have regressions In order to know wither a Bug in the application has been fixed and to make sure that it does not reoccur it needs to be tracked. Figure: Bugs are the centre of their own world If the fix to a Bug is big enough to require that it is broken down into Tasks then it is probably a Requirement. You can associate a check-in with a Bug and have it tracked against a Build. You would also have one or more Test Cases to prove the fix for the Bug. Figure: Bugs have many associations This allows you to track Bugs / Defects in your system effectively and report on them. Change Request – I am not a feature In the CMMI Process template Change Requests can also be easily tracked through the system. In some cases it can be very important to track Change Requests separately as an Auditor may want to know what was changed and who authorised it. Again and similar to Bugs, if the Change Request is big enough that it would require to be broken down into Tasks it is in reality a new feature and should be tracked as a Requirement. Figure: Make sure your Change Requests only Affect Requirements and not rewrite them Conclusion Visual Studio 2010 and Team Foundation Server together provide an exceptional Application Lifecycle Management platform that can help your team comply with even the harshest of Compliance requirements while still enabling them to be Agile. Most Audits are heavy on required documentation but most of that information is captured for you as long a you do it right. You don’t even need every team member to understand it all as each of the Artifacts are relevant to a different type of team member. Business Analysts manage Requirements and Change Requests Programmers manage Tasks and check-in against Change Requests and Bugs Testers manage Bugs and Test Cases Build Masters manage Builds Although there is some crossover there are still rolls or “hats” that are worn. Do you thing this is all achievable? Have I missed anything that you think should be there?

    Read the article

  • External File Upload Optimizations for Windows Azure

    - by rgillen
    [Cross posted from here: http://rob.gillenfamily.net/post/External-File-Upload-Optimizations-for-Windows-Azure.aspx] I’m wrapping up a bit of the work we’ve been doing on data movement optimizations for cloud computing and the latest set of data yielded some interesting points I thought I’d share. The work done here is not really rocket science but may, in some ways, be slightly counter-intuitive and therefore seemed worthy of posting. Summary: for those who don’t like to read detailed posts or don’t have time, the synopsis is that if you are uploading data to Azure, block your data (even down to 1MB) and upload in parallel. Set your block size based on your source file size, but if you must choose a fixed value, use 1MB. Following the above will result in significant performance gains… upwards of 10x-24x and a reduction in overall file transfer time of upwards of 90% (eg, uploading a 1GB file averaged 46.37 minutes prior to optimizations and averaged 1.86 minutes afterwards). Detail: For those of you who want more detail, or think that the claims at the end of the preceding paragraph are over-reaching, what follows is information and code supporting these claims. As the title would indicate, these tests were run from our research facility pointing to the Azure cloud (specifically US North Central as it is physically closest to us) and do not represent intra-cloud results… we have performed intra-cloud tests and the overall results are similar in notion but the data rates are significantly different as well as the tipping points for the various block sizes… this will be detailed separately). We started by building a very simple console application that would loop through a directory and upload each file to Azure storage. This application used the shipping storage client library from the 1.1 version of the azure tools. The only real variation from the client library is that we added code to collect and record the duration (in ms) and size (in bytes) for each file transferred. The code is available here. We then created a directory that had a collection of files for the following sizes: 2KB, 32KB, 64KB, 128KB, 512KB, 1MB, 5MB, 10MB, 25MB, 50MB, 100MB, 250MB, 500MB, 750MB, and 1GB (50 files for each size listed). These files contained randomly-generated binary data and do not benefit from compression (a separate discussion topic). Our file generation tool is available here. The baseline was established by running the application described above against the directory containing all of the data files. This application uploads the files in a random order so as to avoid transferring all of the files of a given size sequentially and thereby spreading the affects of periodic Internet delays across the collection of results.  We then ran some scripts to split the resulting data and generate some reports. The raw data collected for our non-optimized tests is available via the links in the Related Resources section at the bottom of this post. For each file size, we calculated the average upload time (and standard deviation) and the average transfer rate (and standard deviation). As you likely are aware, transferring data across the Internet is susceptible to many transient delays which can cause anomalies in the resulting data. It is for this reason that we randomized the order of source file processing as well as executed the tests 50x for each file size. We expect that these steps will yield a sufficiently balanced set of results. Once the baseline was collected and analyzed, we updated the test harness application with some methods to split the source file into user-defined block sizes and then to upload those blocks in parallel (using the PutBlock() method of Azure storage). The parallelization was handled by simply relying on the Parallel Extensions to .NET to provide a Parallel.For loop (see linked source for specific implementation details in Program.cs, line 173 and following… less than 100 lines total). Once all of the blocks were uploaded, we called PutBlockList() to assemble/commit the file in Azure storage. For each block transferred, the MD5 was calculated and sent ensuring that the bits that arrived matched was was intended. The timer for the blocked/parallelized transfer method wraps the entire process (source file splitting, block transfer, MD5 validation, file committal). A diagram of the process is as follows: We then tested the affects of blocking & parallelizing the transfers by running the updated application against the same source set and did a parameter sweep on the block size including 256KB, 512KB, 1MB, 2MB, and 4MB (our assumption was that anything lower than 256KB wasn’t worth the trouble and 4MB is the maximum size of a block supported by Azure). The raw data for the parallel tests is available via the links in the Related Resources section at the bottom of this post. This data was processed and then compared against the single-threaded / non-optimized transfer numbers and the results were encouraging. The Excel version of the results is available here. Two semi-obvious points need to be made prior to reviewing the data. The first is that if the block size is larger than the source file size you will end up with a “negative optimization” due to the overhead of attempting to block and parallelize. The second is that as the files get smaller, the clock-time cost of blocking and parallelizing (overhead) is more apparent and can tend towards negative optimizations. For this reason (and is supported in the raw data provided in the linked worksheet) the charts and dialog below ignore source file sizes less than 1MB. (click chart for full size image) The chart above illustrates some interesting points about the results: When the block size is smaller than the source file, performance increases but as the block size approaches and then passes the source file size, you see decreasing benefit to the point of negative gains (see the values for the 1MB file size) For some of the moderately-sized source files, small blocks (256KB) are best As the size of the source file gets larger (see values for 50MB and up), the smallest block size is not the most efficient (presumably due, at least in part, to the increased number of blocks, increased number of individual transfer requests, and reassembly/committal costs). Once you pass the 250MB source file size, the difference in rate for 1MB to 4MB blocks is more-or-less constant The 1MB block size gives the best average improvement (~16x) but the optimal approach would be to vary the block size based on the size of the source file.    (click chart for full size image) The above is another view of the same data as the prior chart just with the axis changed (x-axis represents file size and plotted data shows improvement by block size). It again highlights the fact that the 1MB block size is probably the best overall size but highlights the benefits of some of the other block sizes at different source file sizes. This last chart shows the change in total duration of the file uploads based on different block sizes for the source file sizes. Nothing really new here other than this view of the data highlights the negative affects of poorly choosing a block size for smaller files.   Summary What we have found so far is that blocking your file uploads and uploading them in parallel results in significant performance improvements. Further, utilizing extension methods and the Task Parallel Library (.NET 4.0) make short work of altering the shipping client library to provide this functionality while minimizing the amount of change to existing applications that might be using the client library for other interactions.   Related Resources Source code for upload test application Source code for random file generator ODatas feed of raw data from non-optimized transfer tests Experiment Metadata Experiment Datasets 2KB Uploads 32KB Uploads 64KB Uploads 128KB Uploads 256KB Uploads 512KB Uploads 1MB Uploads 5MB Uploads 10MB Uploads 25MB Uploads 50MB Uploads 100MB Uploads 250MB Uploads 500MB Uploads 750MB Uploads 1GB Uploads Raw Data OData feeds of raw data from blocked/parallelized transfer tests Experiment Metadata Experiment Datasets Raw Data 256KB Blocks 512KB Blocks 1MB Blocks 2MB Blocks 4MB Blocks Excel worksheet showing summarizations and comparisons

    Read the article

  • C#/.NET Little Wonders: Comparer&lt;T&gt;.Default

    - by James Michael Hare
    I’ve been working with a wonderful team on a major release where I work, which has had the side-effect of occupying most of my spare time preparing, testing, and monitoring.  However, I do have this Little Wonder tidbit to offer today. Introduction The IComparable<T> interface is great for implementing a natural order for a data type.  It’s a very simple interface with a single method: 1: public interface IComparer<in T> 2: { 3: // Compare two instances of same type. 4: int Compare(T x, T y); 5: }  So what do we expect for the integer return value?  It’s a pseudo-relative measure of the ordering of x and y, which returns an integer value in much the same way C++ returns an integer result from the strcmp() c-style string comparison function: If x == y, returns 0. If x > y, returns > 0 (often +1, but not guaranteed) If x < y, returns < 0 (often –1, but not guaranteed) Notice that the comparison operator used to evaluate against zero should be the same comparison operator you’d use as the comparison operator between x and y.  That is, if you want to see if x > y you’d see if the result > 0. The Problem: Comparing With null Can Be Messy This gets tricky though when you have null arguments.  According to the MSDN, a null value should be considered equal to a null value, and a null value should be less than a non-null value.  So taking this into account we’d expect this instead: If x == y (or both null), return 0. If x > y (or y only is null), return > 0. If x < y (or x only is null), return < 0. But here’s the problem – if x is null, what happens when we attempt to call CompareTo() off of x? 1: // what happens if x is null? 2: x.CompareTo(y); It’s pretty obvious we’ll get a NullReferenceException here.  Now, we could guard against this before calling CompareTo(): 1: int result; 2:  3: // first check to see if lhs is null. 4: if (x == null) 5: { 6: // if lhs null, check rhs to decide on return value. 7: if (y == null) 8: { 9: result = 0; 10: } 11: else 12: { 13: result = -1; 14: } 15: } 16: else 17: { 18: // CompareTo() should handle a null y correctly and return > 0 if so. 19: result = x.CompareTo(y); 20: } Of course, we could shorten this with the ternary operator (?:), but even then it’s ugly repetitive code: 1: int result = (x == null) 2: ? ((y == null) ? 0 : -1) 3: : x.CompareTo(y); Fortunately, the null issues can be cleaned up by drafting in an external Comparer.  The Soltuion: Comparer<T>.Default You can always develop your own instance of IComparer<T> for the job of comparing two items of the same type.  The nice thing about a IComparer is its is independent of the things you are comparing, so this makes it great for comparing in an alternative order to the natural order of items, or when one or both of the items may be null. 1: public class NullableIntComparer : IComparer<int?> 2: { 3: public int Compare(int? x, int? y) 4: { 5: return (x == null) 6: ? ((y == null) ? 0 : -1) 7: : x.Value.CompareTo(y); 8: } 9: }  Now, if you want a custom sort -- especially on large-grained objects with different possible sort fields -- this is the best option you have.  But if you just want to take advantage of the natural ordering of the type, there is an easier way.  If the type you want to compare already implements IComparable<T> or if the type is System.Nullable<T> where T implements IComparable, there is a class in the System.Collections.Generic namespace called Comparer<T> which exposes a property called Default that will create a singleton that represents the default comparer for items of that type.  For example: 1: // compares integers 2: var intComparer = Comparer<int>.Default; 3:  4: // compares DateTime values 5: var dateTimeComparer = Comparer<DateTime>.Default; 6:  7: // compares nullable doubles using the null rules! 8: var nullableDoubleComparer = Comparer<double?>.Default;  This helps you avoid having to remember the messy null logic and makes it to compare objects where you don’t know if one or more of the values is null. This works especially well when creating say an IComparer<T> implementation for a large-grained class that may or may not contain a field.  For example, let’s say you want to create a sorting comparer for a stock open price, but if the market the stock is trading in hasn’t opened yet, the open price will be null.  We could handle this (assuming a reasonable Quote definition) like: 1: public class Quote 2: { 3: // the opening price of the symbol quoted 4: public double? Open { get; set; } 5:  6: // ticker symbol 7: public string Symbol { get; set; } 8:  9: // etc. 10: } 11:  12: public class OpenPriceQuoteComparer : IComparer<Quote> 13: { 14: // Compares two quotes by opening price 15: public int Compare(Quote x, Quote y) 16: { 17: return Comparer<double?>.Default.Compare(x.Open, y.Open); 18: } 19: } Summary Defining a custom comparer is often needed for non-natural ordering or defining alternative orderings, but when you just want to compare two items that are IComparable<T> and account for null behavior, you can use the Comparer<T>.Default comparer generator and you’ll never have to worry about correct null value sorting again.     Technorati Tags: C#,.NET,Little Wonders,BlackRabbitCoder,IComparable,Comparer

    Read the article

  • I see no LOBs!

    - by Paul White
    Is it possible to see LOB (large object) logical reads from STATISTICS IO output on a table with no LOB columns? I was asked this question today by someone who had spent a good fraction of their afternoon trying to work out why this was occurring – even going so far as to re-run DBCC CHECKDB to see if any corruption had taken place.  The table in question wasn’t particularly pretty – it had grown somewhat organically over time, with new columns being added every so often as the need arose.  Nevertheless, it remained a simple structure with no LOB columns – no TEXT or IMAGE, no XML, no MAX types – nothing aside from ordinary INT, MONEY, VARCHAR, and DATETIME types.  To add to the air of mystery, not every query that ran against the table would report LOB logical reads – just sometimes – but when it did, the query often took much longer to execute. Ok, enough of the pre-amble.  I can’t reproduce the exact structure here, but the following script creates a table that will serve to demonstrate the effect: IF OBJECT_ID(N'dbo.Test', N'U') IS NOT NULL DROP TABLE dbo.Test GO CREATE TABLE dbo.Test ( row_id NUMERIC IDENTITY NOT NULL,   col01 NVARCHAR(450) NOT NULL, col02 NVARCHAR(450) NOT NULL, col03 NVARCHAR(450) NOT NULL, col04 NVARCHAR(450) NOT NULL, col05 NVARCHAR(450) NOT NULL, col06 NVARCHAR(450) NOT NULL, col07 NVARCHAR(450) NOT NULL, col08 NVARCHAR(450) NOT NULL, col09 NVARCHAR(450) NOT NULL, col10 NVARCHAR(450) NOT NULL, CONSTRAINT [PK dbo.Test row_id] PRIMARY KEY CLUSTERED (row_id) ) ; The next script loads the ten variable-length character columns with one-character strings in the first row, two-character strings in the second row, and so on down to the 450th row: WITH Numbers AS ( -- Generates numbers 1 - 450 inclusive SELECT TOP (450) n = ROW_NUMBER() OVER (ORDER BY (SELECT 0)) FROM master.sys.columns C1, master.sys.columns C2, master.sys.columns C3 ORDER BY n ASC ) INSERT dbo.Test WITH (TABLOCKX) SELECT REPLICATE(N'A', N.n), REPLICATE(N'B', N.n), REPLICATE(N'C', N.n), REPLICATE(N'D', N.n), REPLICATE(N'E', N.n), REPLICATE(N'F', N.n), REPLICATE(N'G', N.n), REPLICATE(N'H', N.n), REPLICATE(N'I', N.n), REPLICATE(N'J', N.n) FROM Numbers AS N ORDER BY N.n ASC ; Once those two scripts have run, the table contains 450 rows and 10 columns of data like this: Most of the time, when we query data from this table, we don’t see any LOB logical reads, for example: -- Find the maximum length of the data in -- column 5 for a range of rows SELECT result = MAX(DATALENGTH(T.col05)) FROM dbo.Test AS T WHERE row_id BETWEEN 50 AND 100 ; But with a different query… -- Read all the data in column 1 SELECT result = MAX(DATALENGTH(T.col01)) FROM dbo.Test AS T ; …suddenly we have 49 LOB logical reads, as well as the ‘normal’ logical reads we would expect. The Explanation If we had tried to create this table in SQL Server 2000, we would have received a warning message to say that future INSERT or UPDATE operations on the table might fail if the resulting row exceeded the in-row storage limit of 8060 bytes.  If we needed to store more data than would fit in an 8060 byte row (including internal overhead) we had to use a LOB column – TEXT, NTEXT, or IMAGE.  These special data types store the large data values in a separate structure, with just a small pointer left in the original row. Row Overflow SQL Server 2005 introduced a feature called row overflow, which allows one or more variable-length columns in a row to move to off-row storage if the data in a particular row would otherwise exceed 8060 bytes.  You no longer receive a warning when creating (or altering) a table that might need more than 8060 bytes of in-row storage; if SQL Server finds that it can no longer fit a variable-length column in a particular row, it will silently move one or more of these columns off the row into a separate allocation unit. Only variable-length columns can be moved in this way (for example the (N)VARCHAR, VARBINARY, and SQL_VARIANT types).  Fixed-length columns (like INTEGER and DATETIME for example) never move into ‘row overflow’ storage.  The decision to move a column off-row is done on a row-by-row basis – so data in a particular column might be stored in-row for some table records, and off-row for others. In general, if SQL Server finds that it needs to move a column into row-overflow storage, it moves the largest variable-length column record for that row.  Note that in the case of an UPDATE statement that results in the 8060 byte limit being exceeded, it might not be the column that grew that is moved! Sneaky LOBs Anyway, that’s all very interesting but I don’t want to get too carried away with the intricacies of row-overflow storage internals.  The point is that it is now possible to define a table with non-LOB columns that will silently exceed the old row-size limit and result in ordinary variable-length columns being moved to off-row storage.  Adding new columns to a table, expanding an existing column definition, or simply storing more data in a column than you used to – all these things can result in one or more variable-length columns being moved off the row. Note that row-overflow storage is logically quite different from old-style LOB and new-style MAX data type storage – individual variable-length columns are still limited to 8000 bytes each – you can just have more of them now.  Having said that, the physical mechanisms involved are very similar to full LOB storage – a column moved to row-overflow leaves a 24-byte pointer record in the row, and the ‘separate storage’ I have been talking about is structured very similarly to both old-style LOBs and new-style MAX types.  The disadvantages are also the same: when SQL Server needs a row-overflow column value it needs to follow the in-row pointer a navigate another chain of pages, just like retrieving a traditional LOB. And Finally… In the example script presented above, the rows with row_id values from 402 to 450 inclusive all exceed the total in-row storage limit of 8060 bytes.  A SELECT that references a column in one of those rows that has moved to off-row storage will incur one or more lob logical reads as the storage engine locates the data.  The results on your system might vary slightly depending on your settings, of course; but in my tests only column 1 in rows 402-450 moved off-row.  You might like to play around with the script – updating columns, changing data type lengths, and so on – to see the effect on lob logical reads and which columns get moved when.  You might even see row-overflow columns moving back in-row if they are updated to be smaller (hint: reduce the size of a column entry by at least 1000 bytes if you hope to see this). Be aware that SQL Server will not warn you when it moves ‘ordinary’ variable-length columns into overflow storage, and it can have dramatic effects on performance.  It makes more sense than ever to choose column data types sensibly.  If you make every column a VARCHAR(8000) or NVARCHAR(4000), and someone stores data that results in a row needing more than 8060 bytes, SQL Server might turn some of your column data into pseudo-LOBs – all without saying a word. Finally, some people make a distinction between ordinary LOBs (those that can hold up to 2GB of data) and the LOB-like structures created by row-overflow (where columns are still limited to 8000 bytes) by referring to row-overflow LOBs as SLOBs.  I find that quite appealing, but the ‘S’ stands for ‘small’, which makes expanding the whole acronym a little daft-sounding…small large objects anyone? © Paul White 2011 email: [email protected] twitter: @SQL_Kiwi

    Read the article

  • My vertex shader doesn't affect texture coords or diffuse info but works for position

    - by tina nyaa
    I am new to 3D and DirectX - in the past I have only used abstractions for 2D drawing. Over the past month I've been studying really hard and I'm trying to modify and adapt some of the shaders as part of my personal 'study project'. Below I have a shader, modified from one of the Microsoft samples. I set diffuse and tex0 vertex shader outputs to zero, but my model still shows the full texture and lighting as if I hadn't changed the values from the vertex buffer. Changing the position of the model works, but nothing else. Why is this? // // Skinned Mesh Effect file // Copyright (c) 2000-2002 Microsoft Corporation. All rights reserved. // float4 lhtDir = {0.0f, 0.0f, -1.0f, 1.0f}; //light Direction float4 lightDiffuse = {0.6f, 0.6f, 0.6f, 1.0f}; // Light Diffuse float4 MaterialAmbient : MATERIALAMBIENT = {0.1f, 0.1f, 0.1f, 1.0f}; float4 MaterialDiffuse : MATERIALDIFFUSE = {0.8f, 0.8f, 0.8f, 1.0f}; // Matrix Pallette static const int MAX_MATRICES = 100; float4x3 mWorldMatrixArray[MAX_MATRICES] : WORLDMATRIXARRAY; float4x4 mViewProj : VIEWPROJECTION; /////////////////////////////////////////////////////// struct VS_INPUT { float4 Pos : POSITION; float4 BlendWeights : BLENDWEIGHT; float4 BlendIndices : BLENDINDICES; float3 Normal : NORMAL; float3 Tex0 : TEXCOORD0; }; struct VS_OUTPUT { float4 Pos : POSITION; float4 Diffuse : COLOR; float2 Tex0 : TEXCOORD0; }; float3 Diffuse(float3 Normal) { float CosTheta; // N.L Clamped CosTheta = max(0.0f, dot(Normal, lhtDir.xyz)); // propogate scalar result to vector return (CosTheta); } VS_OUTPUT VShade(VS_INPUT i, uniform int NumBones) { VS_OUTPUT o; float3 Pos = 0.0f; float3 Normal = 0.0f; float LastWeight = 0.0f; // Compensate for lack of UBYTE4 on Geforce3 int4 IndexVector = D3DCOLORtoUBYTE4(i.BlendIndices); // cast the vectors to arrays for use in the for loop below float BlendWeightsArray[4] = (float[4])i.BlendWeights; int IndexArray[4] = (int[4])IndexVector; // calculate the pos/normal using the "normal" weights // and accumulate the weights to calculate the last weight for (int iBone = 0; iBone < NumBones-1; iBone++) { LastWeight = LastWeight + BlendWeightsArray[iBone]; Pos += mul(i.Pos, mWorldMatrixArray[IndexArray[iBone]]) * BlendWeightsArray[iBone]; Normal += mul(i.Normal, mWorldMatrixArray[IndexArray[iBone]]) * BlendWeightsArray[iBone]; } LastWeight = 1.0f - LastWeight; // Now that we have the calculated weight, add in the final influence Pos += (mul(i.Pos, mWorldMatrixArray[IndexArray[NumBones-1]]) * LastWeight); Normal += (mul(i.Normal, mWorldMatrixArray[IndexArray[NumBones-1]]) * LastWeight); // transform position from world space into view and then projection space //o.Pos = mul(float4(Pos.xyz, 1.0f), mViewProj); o.Pos = mul(float4(Pos.xyz, 1.0f), mViewProj); o.Diffuse.x = 0.0f; o.Diffuse.y = 0.0f; o.Diffuse.z = 0.0f; o.Diffuse.w = 0.0f; o.Tex0 = float2(0,0); return o; } technique t0 { pass p0 { VertexShader = compile vs_3_0 VShade(4); } } I am currently using the SlimDX .NET wrapper around DirectX, but the API is extremely similar: public void Draw() { var device = vertexBuffer.Device; device.Clear(ClearFlags.Target | ClearFlags.ZBuffer, Color.White, 1.0f, 0); device.SetRenderState(RenderState.Lighting, true); device.SetRenderState(RenderState.DitherEnable, true); device.SetRenderState(RenderState.ZEnable, true); device.SetRenderState(RenderState.CullMode, Cull.Counterclockwise); device.SetRenderState(RenderState.NormalizeNormals, true); device.SetSamplerState(0, SamplerState.MagFilter, TextureFilter.Anisotropic); device.SetSamplerState(0, SamplerState.MinFilter, TextureFilter.Anisotropic); device.SetTransform(TransformState.World, Matrix.Identity * Matrix.Translation(0, -50, 0)); device.SetTransform(TransformState.View, Matrix.LookAtLH(new Vector3(-200, 0, 0), Vector3.Zero, Vector3.UnitY)); device.SetTransform(TransformState.Projection, Matrix.PerspectiveFovLH((float)Math.PI / 4, (float)device.Viewport.Width / device.Viewport.Height, 10, 10000000)); var material = new Material(); material.Ambient = material.Diffuse = material.Emissive = material.Specular = new Color4(Color.White); material.Power = 1f; device.SetStreamSource(0, vertexBuffer, 0, vertexSize); device.VertexDeclaration = vertexDeclaration; device.Indices = indexBuffer; device.Material = material; device.SetTexture(0, texture); var param = effect.GetParameter(null, "mWorldMatrixArray"); var boneWorldTransforms = bones.OrderedBones.OrderBy(x => x.Id).Select(x => x.CombinedTransformation).ToArray(); effect.SetValue(param, boneWorldTransforms); effect.SetValue(effect.GetParameter(null, "mViewProj"), Matrix.Identity);// Matrix.PerspectiveFovLH((float)Math.PI / 4, (float)device.Viewport.Width / device.Viewport.Height, 10, 10000000)); effect.SetValue(effect.GetParameter(null, "MaterialDiffuse"), material.Diffuse); effect.SetValue(effect.GetParameter(null, "MaterialAmbient"), material.Ambient); effect.Technique = effect.GetTechnique(0); var passes = effect.Begin(FX.DoNotSaveState); for (var i = 0; i < passes; i++) { effect.BeginPass(i); device.DrawIndexedPrimitives(PrimitiveType.TriangleList, 0, 0, skin.Vertices.Length, 0, skin.Indicies.Length / 3); effect.EndPass(); } effect.End(); } Again, I set diffuse and tex0 vertex shader outputs to zero, but my model still shows the full texture and lighting as if I hadn't changed the values from the vertex buffer. Changing the position of the model works, but nothing else. Why is this? Also, whatever I set in the bone transformation matrices doesn't seem to have an effect on my model. If I set every bone transformation to a zero matrix, the model still shows up as if nothing had happened, but changing the Pos field in shader output makes the model disappear. I don't understand why I'm getting this kind of behaviour. Thank you!

    Read the article

  • Developing Schema Compare for Oracle (Part 1)

    - by Simon Cooper
    SQL Compare is one of Red Gate's most successful SQL Server tools; it allows developers and DBAs to compare and synchronize the contents of their databases. Although similar tools exist for Oracle, they are quite noticeably lacking in the usability and stability that SQL Compare is known for in the SQL Server world. We could see a real need for a usable schema comparison tools for Oracle, and so the Schema Compare for Oracle project was born. Over the next few weeks, as we come up to release of v1, I'll be doing a series of posts on the development of Schema Compare for Oracle. For the first post, I thought I would start with the main pitfalls that we stumbled across when developing the product, especially from a SQL Server background. 1. Schemas and Databases The most obvious difference is that the concept of a 'database' is quite different between Oracle and SQL Server. On SQL Server, one server instance has multiple databases, each with separate schemas. There is typically little communication between separate databases, and most databases are no more than about 1000-2000 objects. This means SQL Compare can register an entire database in a reasonable amount of time, and cross-database dependencies probably won't be an issue. It is a quite different scene under Oracle, however. The terms 'database' and 'instance' are used interchangeably, (although technically 'database' refers to the datafiles on disk, and 'instance' the running Oracle process that reads & writes to the database), and a database is a single conceptual entity. This immediately presents problems, as it is infeasible to register an entire database as we do in SQL Compare; in my Oracle install, using the standard recommended options, there are 63975 system objects. If we tried to register all those, not only would it take hours, but the client would probably run out of memory before we finished. As a result, we had to allow people to specify what schemas they wanted to register. This decision had quite a few knock-on effects for the design, which I will cover in a future post. 2. Connecting to Oracle The next obvious difference is in actually connecting to Oracle – in SQL Server, you can specify a server and database, and off you go. On Oracle things are slightly more complicated. SIDs, Service Names, and TNS A database (the files on disk) must have a unique identifier for the databases on the system, called the SID. It also has a global database name, which consists of a name (which doesn't have to match the SID) and a domain. Alternatively, you can identify a database using a service name, which normally has a 1-to-1 relationship with instances, but may not if, for example, using RAC (Real Application Clusters) for redundancy and failover. You specify the computer and instance you want to connect to using TNS (Transparent Network Substrate). The user-visible parts are a config file (tnsnames.ora) on the client machine that specifies how to connect to an instance. For example, the entry for one of my test instances is: SC_11GDB1 = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = simonctest)(PORT = 1521)) ) (CONNECT_DATA = (SID = 11gR1db1) ) ) This gives the hostname, port, and SID of the instance I want to connect to, and associates it with a name (SC_11GDB1). The tnsnames syntax also allows you to specify failover, multiple descriptions and address lists, and client load balancing. You can then specify this TNS identifier as the data source in a connection string. Although using ODP.NET (the .NET dlls provided by Oracle) was fine for internal prototype builds, once we released the EAP we discovered that this simply wasn't an acceptable solution for installs on other people's machines. Due to .NET assembly strong naming, users had to have installed on their machines the exact same version of the ODP.NET dlls as we had on our build server. We couldn't ship the ODP.NET dlls with our installer as the Oracle license agreement prohibited this, and we didn't want to force users to install another Oracle client just so they can run our program. To be able to list the TNS entries in the connection dialog, we also had to locate and parse the tnsnames.ora file, which was complicated by users with several Oracle client installs and intricate TNS entries. After much swearing at our computers, we eventually decided to use a third party Oracle connection library from Devart that we could ship with our program; this could use whatever client version was installed, parse the TNS entries for us, and also had the nice feature of being able to connect to an Oracle server without having any client installed at all. Unfortunately, their current license agreement prevents us from shipping an Oracle SDK, but that's a bridge we'll cross when we get to it. 3. Running synchronization scripts The most important difference is that in Oracle, DDL is non-transactional; you cannot rollback DDL statements like you can on SQL Server. Although we considered various solutions to this, including using the flashback archive or recycle bin, or generating an undo script, no reliable method of completely undoing a half-executed sync script has yet been found; so in this case we simply have to trust that the DBA or developer will check and verify the script before running it. However, before we got to that stage, we had to get the scripts to run in the first place... To run a synchronization script from SQL Compare we essentially pass the script over to the SqlCommand.ExecuteNonQuery method. However, when we tried to do the same for an OracleConnection we got a very strange error – 'ORA-00911: invalid character', even when running the most basic CREATE TABLE command. After much hair-pulling and Googling, we discovered that Oracle has got some very strange behaviour with semicolons at the end of statements. To understand what's going on, we need to take a quick foray into SQL and PL/SQL. PL/SQL is not T-SQL In SQL Server, T-SQL is the language used to interface with the database. It has DDL, DML, control flow, and many other nice features (like Turing-completeness) that you can mix and match in the same script. In Oracle, DDL SQL and PL/SQL are two completely separate languages, with different syntax, different datatypes and different execution engines within the instance. Oracle SQL is much more like 'pure' ANSI SQL, with no state, no control flow, and only the basic DML commands. PL/SQL is the Turing-complete language, but can only do DML and DCL (i.e. BEGIN TRANSATION commands). Any DDL or SQL commands that aren't recognised by the PL/SQL engine have to be passed back to the SQL engine via an EXECUTE IMMEDIATE command. In PL/SQL, a semicolons is a valid token used to delimit the end of a statement. In SQL, a semicolon is not a valid token (even though the Oracle documentation gives them at the end of the syntax diagrams) . When you execute the command CREATE TABLE table1 (COL1 NUMBER); in SQL*Plus the semicolon on the end is a command to SQL*Plus to execute the preceding statement on the server; it strips off the semicolon before passing it on. SQL Developer does a similar thing. When executing a PL/SQL block, however, the syntax is like so: BEGIN INSERT INTO table1 VALUES (1); INSERT INTO table1 VALUES (2); END; / In this case, the semicolon is accepted by the PL/SQL engine as a statement delimiter, and instead the / is the command to SQL*Plus to execute the current block. This explains the ORA-00911 error we got when trying to run the CREATE TABLE command – the server is complaining about the semicolon on the end. This also means that there is no SQL syntax to execute more than one DDL command in the same OracleCommand. Therefore, we would have to do a round-trip to the server for every command we want to execute. Obviously, this would cause lots of network traffic and be very slow on slow or congested networks. Our first attempt at a solution was to wrap every SQL statement (without semicolon) inside an EXECUTE IMMEDIATE command in a PL/SQL block and pass that to the server to execute. One downside of this solution is that we get no feedback as to how the script execution is going; we're currently evaluating better solutions to this thorny issue. Next up: Dependencies; how we solved the problem of being unable to register the entire database, and the knock-on effects to the whole product.

    Read the article

  • Behavior Driven Development (BDD) and DevExpress XAF

    - by Patrick Liekhus
    So in my previous posts I showed you how I used EDMX to quickly build my business objects within XPO and XAF.  But how do you test whether your business objects are actually doing what you want and verify that your business logic is correct?  Well I was reading my monthly MSDN magazine last last year and came across an article about using SpecFlow and WatiN to build BDD tests.  So why not use these same techniques to write SpecFlow style scripts and have them generate EasyTest scripts for use with XAF.  Let me outline and show a few things below.  I plan on releasing this code in a short while, I just wanted to preview what I was thinking. Before we begin… First, if you have not read the article in MSDN, here is the link to the article that I found my inspiration.  It covers the overview of BDD vs. TDD, how to write some of the SpecFlow syntax and how use the “Steps” logic to create your own tests. Second, if you have not heard of EasyTest from DevExpress I strongly recommend you review it here.  It basically takes the power of XAF and the beauty of your application and allows you to create text based files to execute automated commands within your application. Why would we do this?  Because as you will see below, the cucumber syntax is easier for business analysts to interpret and digest the business rules from.  You can find most of the information you will need on Cucumber syntax within The Secret Ninja Cucumber Scrolls located here.  The basics of the syntax are that Given X When Y Then Z.  For example, Given I am at the login screen When I enter my login credentials Then I expect to see the home screen.  Pretty easy syntax to follow. Finally, we will need to download and install SpecFlow.  You can find it on their website here.  Once you have this installed then let’s write our first test. Let’s get started… So where to start.  Create a new testing project within your solution.  I typically call this with a similar naming convention as used by XAF, my project name .FunctionalTests (i.e.  AlbumManager.FunctionalTests).  Remove the basic test that is created for you.  We will not use the default test but rather create our own SpecFlow “Feature” files.  Add a new item to your project and select the SpecFlow Feature file under C#.  Name your feature file as you do your class files after the test they are performing. Now you can crack open your new feature file and write the actual test.  Make sure to have your Ninja Scrolls from above as it provides valuable resources on how to write your test syntax.  In this test below you can see how I defined the documentation in the Feature section.  This is strictly for our purposes of readability and do not effect the test.  The next section is the Scenario Outline which is considered a test template.  You can see the brackets <> around the fields that will be filled in for each test.  So in the example below you can see that Given I am starting a new test and the application is open.  This means I want a new EasyTest file and the windows application generated by XAF is open.  Next When I am at the Albums screen tells XAF to navigate to the Albums list view.  And I click the New:Album button, tells XAF to click the new button on the list grid.  And I enter the following information tells XAF which fields to complete with the mapped values.  And I click the Save and Close button causes the record to be saved and the detail form to be closed.  Then I verify results tests the input data against what is visible in the grid to ensure that your record was created. The Scenarios section gives each test a unique name and then fills in the values for each test.  This way you can use the same test to make multiple passes with different data. Almost there.  Now we must save the feature file and the BDD tests will be written using standard unit test syntax.  This is all handled for you by SpecFlow so just save the file.  What you will see in your Test List Editor is a unit test for each of the above scenarios you just built. You can now use standard unit testing frameworks to execute the test as you desire.  As you would expect then, these BDD SpecFlow tests can be automated into your build process to ensure that your business requirements are satisfied each and every time. How does it work? What we have done is to intercept the testing logic at runtime to interpret the SpecFlow syntax into EasyTest syntax.  This is the basic StepDefinitions that we are working on now.  We expect to put these on CodePlex within the next few days.  You can always override and make your own rules as you see fit for your project.  Follow the MSDN magazine above to start your own.  You can see part of our implementation below. As you can gather from the MSDN article and the code sample below, we have created our own common rules to build the above syntax. The code implementation for these rules basically saves your information from the feature file into an EasyTest file format.  It then executes the EasyTest file and parses the XML results of the test.  If the test succeeds the test is passed.  If the test fails, the EasyTest failure message is logged and the screen shot (as captured by EasyTest) is saved for your review. Again we are working on getting this code ready for mass consumption, but at this time it is not ready.  We will post another message when it is ready with all details about usage and setup. Thanks

    Read the article

  • DTracing TCP congestion control

    - by user12820842
    In a previous post, I showed how we can use DTrace to probe TCP receive and send window events. TCP receive and send windows are in effect both about flow-controlling how much data can be received - the receive window reflects how much data the local TCP is prepared to receive, while the send window simply reflects the size of the receive window of the peer TCP. Both then represent flow control as imposed by the receiver. However, consider that without the sender imposing flow control, and a slow link to a peer, TCP will simply fill up it's window with sent segments. Dealing with multiple TCP implementations filling their peer TCP's receive windows in this manner, busy intermediate routers may drop some of these segments, leading to timeout and retransmission, which may again lead to drops. This is termed congestion, and TCP has multiple congestion control strategies. We can see that in this example, we need to have some way of adjusting how much data we send depending on how quickly we receive acknowledgement - if we get ACKs quickly, we can safely send more segments, but if acknowledgements come slowly, we should proceed with more caution. More generally, we need to implement flow control on the send side also. Slow Start and Congestion Avoidance From RFC2581, let's examine the relevant variables: "The congestion window (cwnd) is a sender-side limit on the amount of data the sender can transmit into the network before receiving an acknowledgment (ACK). Another state variable, the slow start threshold (ssthresh), is used to determine whether the slow start or congestion avoidance algorithm is used to control data transmission" Slow start is used to probe the network's ability to handle transmission bursts both when a connection is first created and when retransmission timers fire. The latter case is important, as the fact that we have effectively lost TCP data acts as a motivator for re-probing how much data the network can handle from the sending TCP. The congestion window (cwnd) is initialized to a relatively small value, generally a low multiple of the sending maximum segment size. When slow start kicks in, we will only send that number of bytes before waiting for acknowledgement. When acknowledgements are received, the congestion window is increased in size until cwnd reaches the slow start threshold ssthresh value. For most congestion control algorithms the window increases exponentially under slow start, assuming we receive acknowledgements. We send 1 segment, receive an ACK, increase the cwnd by 1 MSS to 2*MSS, send 2 segments, receive 2 ACKs, increase the cwnd by 2*MSS to 4*MSS, send 4 segments etc. When the congestion window exceeds the slow start threshold, congestion avoidance is used instead of slow start. During congestion avoidance, the congestion window is generally updated by one MSS for each round-trip-time as opposed to each ACK, and so cwnd growth is linear instead of exponential (we may receive multiple ACKs within a single RTT). This continues until congestion is detected. If a retransmit timer fires, congestion is assumed and the ssthresh value is reset. It is reset to a fraction of the number of bytes outstanding (unacknowledged) in the network. At the same time the congestion window is reset to a single max segment size. Thus, we initiate slow start until we start receiving acknowledgements again, at which point we can eventually flip over to congestion avoidance when cwnd ssthresh. Congestion control algorithms differ most in how they handle the other indication of congestion - duplicate ACKs. A duplicate ACK is a strong indication that data has been lost, since they often come from a receiver explicitly asking for a retransmission. In some cases, a duplicate ACK may be generated at the receiver as a result of packets arriving out-of-order, so it is sensible to wait for multiple duplicate ACKs before assuming packet loss rather than out-of-order delivery. This is termed fast retransmit (i.e. retransmit without waiting for the retransmission timer to expire). Note that on Oracle Solaris 11, the congestion control method used can be customized. See here for more details. In general, 3 or more duplicate ACKs indicate packet loss and should trigger fast retransmit . It's best not to revert to slow start in this case, as the fact that the receiver knew it was missing data suggests it has received data with a higher sequence number, so we know traffic is still flowing. Falling back to slow start would be excessive therefore, so fast recovery is used instead. Observing slow start and congestion avoidance The following script counts TCP segments sent when under slow start (cwnd ssthresh). #!/usr/sbin/dtrace -s #pragma D option quiet tcp:::connect-request / start[args[1]-cs_cid] == 0/ { start[args[1]-cs_cid] = 1; } tcp:::send / start[args[1]-cs_cid] == 1 && args[3]-tcps_cwnd tcps_cwnd_ssthresh / { @c["Slow start", args[2]-ip_daddr, args[4]-tcp_dport] = count(); } tcp:::send / start[args[1]-cs_cid] == 1 && args[3]-tcps_cwnd args[3]-tcps_cwnd_ssthresh / { @c["Congestion avoidance", args[2]-ip_daddr, args[4]-tcp_dport] = count(); } As we can see the script only works on connections initiated since it is started (using the start[] associative array with the connection ID as index to set whether it's a new connection (start[cid] = 1). From there we simply differentiate send events where cwnd ssthresh (congestion avoidance). Here's the output taken when I accessed a YouTube video (where rport is 80) and from an FTP session where I put a large file onto a remote system. # dtrace -s tcp_slow_start.d ^C ALGORITHM RADDR RPORT #SEG Slow start 10.153.125.222 20 6 Slow start 138.3.237.7 80 14 Slow start 10.153.125.222 21 18 Congestion avoidance 10.153.125.222 20 1164 We see that in the case of the YouTube video, slow start was exclusively used. Most of the segments we sent in that case were likely ACKs. Compare this case - where 14 segments were sent using slow start - to the FTP case, where only 6 segments were sent before we switched to congestion avoidance for 1164 segments. In the case of the FTP session, the FTP data on port 20 was predominantly sent with congestion avoidance in operation, while the FTP session relied exclusively on slow start. For the default congestion control algorithm - "newreno" - on Solaris 11, slow start will increase the cwnd by 1 MSS for every acknowledgement received, and by 1 MSS for each RTT in congestion avoidance mode. Different pluggable congestion control algorithms operate slightly differently. For example "highspeed" will update the slow start cwnd by the number of bytes ACKed rather than the MSS. And to finish, here's a neat oneliner to visually display the distribution of congestion window values for all TCP connections to a given remote port using a quantization. In this example, only port 80 is in use and we see the majority of cwnd values for that port are in the 4096-8191 range. # dtrace -n 'tcp:::send { @q[args[4]-tcp_dport] = quantize(args[3]-tcps_cwnd); }' dtrace: description 'tcp:::send ' matched 10 probes ^C 80 value ------------- Distribution ------------- count -1 | 0 0 |@@@@@@ 5 1 | 0 2 | 0 4 | 0 8 | 0 16 | 0 32 | 0 64 | 0 128 | 0 256 | 0 512 | 0 1024 | 0 2048 |@@@@@@@@@ 8 4096 |@@@@@@@@@@@@@@@@@@@@@@@@@@ 23 8192 | 0

    Read the article

  • Parallel Classloading Revisited: Fully Concurrent Loading

    - by davidholmes
    Java 7 introduced support for parallel classloading. A description of that project and its goals can be found here: http://openjdk.java.net/groups/core-libs/ClassLoaderProposal.html The solution for parallel classloading was to add to each class loader a ConcurrentHashMap, referenced through a new field, parallelLockMap. This contains a mapping from class names to Objects to use as a classloading lock for that class name. This was then used in the following way: protected Class loadClass(String name, boolean resolve) throws ClassNotFoundException { synchronized (getClassLoadingLock(name)) { // First, check if the class has already been loaded Class c = findLoadedClass(name); if (c == null) { long t0 = System.nanoTime(); try { if (parent != null) { c = parent.loadClass(name, false); } else { c = findBootstrapClassOrNull(name); } } catch (ClassNotFoundException e) { // ClassNotFoundException thrown if class not found // from the non-null parent class loader } if (c == null) { // If still not found, then invoke findClass in order // to find the class. long t1 = System.nanoTime(); c = findClass(name); // this is the defining class loader; record the stats sun.misc.PerfCounter.getParentDelegationTime().addTime(t1 - t0); sun.misc.PerfCounter.getFindClassTime().addElapsedTimeFrom(t1); sun.misc.PerfCounter.getFindClasses().increment(); } } if (resolve) { resolveClass(c); } return c; } } Where getClassLoadingLock simply does: protected Object getClassLoadingLock(String className) { Object lock = this; if (parallelLockMap != null) { Object newLock = new Object(); lock = parallelLockMap.putIfAbsent(className, newLock); if (lock == null) { lock = newLock; } } return lock; } This approach is very inefficient in terms of the space used per map and the number of maps. First, there is a map per-classloader. As per the code above under normal delegation the current classloader creates and acquires a lock for the given class, checks if it is already loaded, then asks its parent to load it; the parent in turn creates another lock in its own map, checks if the class is already loaded and then delegates to its parent and so on till the boot loader is invoked for which there is no map and no lock. So even in the simplest of applications, you will have two maps (in the system and extensions loaders) for every class that has to be loaded transitively from the application's main class. If you knew before hand which loader would actually load the class the locking would only need to be performed in that loader. As it stands the locking is completely unnecessary for all classes loaded by the boot loader. Secondly, once loading has completed and findClass will return the class, the lock and the map entry is completely unnecessary. But as it stands, the lock objects and their associated entries are never removed from the map. It is worth understanding exactly what the locking is intended to achieve, as this will help us understand potential remedies to the above inefficiencies. Given this is the support for parallel classloading, the class loader itself is unlikely to need to guard against concurrent load attempts - and if that were not the case it is likely that the classloader would need a different means to protect itself rather than a lock per class. Ultimately when a class file is located and the class has to be loaded, defineClass is called which calls into the VM - the VM does not require any locking at the Java level and uses its own mutexes for guarding its internal data structures (such as the system dictionary). The classloader locking is primarily needed to address the following situation: if two threads attempt to load the same class, one will initiate the request through the appropriate loader and eventually cause defineClass to be invoked. Meanwhile the second attempt will block trying to acquire the lock. Once the class is loaded the first thread will release the lock, allowing the second to acquire it. The second thread then sees that the class has now been loaded and will return that class. Neither thread can tell which did the loading and they both continue successfully. Consider if no lock was acquired in the classloader. Both threads will eventually locate the file for the class, read in the bytecodes and call defineClass to actually load the class. In this case the first to call defineClass will succeed, while the second will encounter an exception due to an attempted redefinition of an existing class. It is solely for this error condition that the lock has to be used. (Note that parallel capable classloaders should not need to be doing old deadlock-avoidance tricks like doing a wait() on the lock object\!). There are a number of obvious things we can try to solve this problem and they basically take three forms: Remove the need for locking. This might be achieved by having a new version of defineClass which acts like defineClassIfNotPresent - simply returning an existing Class rather than triggering an exception. Increase the coarseness of locking to reduce the number of lock objects and/or maps. For example, using a single shared lockMap instead of a per-loader lockMap. Reduce the lifetime of lock objects so that entries are removed from the map when no longer needed (eg remove after loading, use weak references to the lock objects and cleanup the map periodically). There are pros and cons to each of these approaches. Unfortunately a significant "con" is that the API introduced in Java 7 to support parallel classloading has essentially mandated that these locks do in fact exist, and they are accessible to the application code (indirectly through the classloader if it exposes them - which a custom loader might do - and regardless they are accessible to custom classloaders). So while we can reason that we could do parallel classloading with no locking, we can not implement this without breaking the specification for parallel classloading that was put in place for Java 7. Similarly we might reason that we can remove a mapping (and the lock object) because the class is already loaded, but this would again violate the specification because it can be reasoned that the following assertion should hold true: Object lock1 = loader.getClassLoadingLock(name); loader.loadClass(name); Object lock2 = loader.getClassLoadingLock(name); assert lock1 == lock2; Without modifying the specification, or at least doing some creative wordsmithing on it, options 1 and 3 are precluded. Even then there are caveats, for example if findLoadedClass is not atomic with respect to defineClass, then you can have concurrent calls to findLoadedClass from different threads and that could be expensive (this is also an argument against moving findLoadedClass outside the locked region - it may speed up the common case where the class is already loaded, but the cost of re-executing after acquiring the lock could be prohibitive. Even option 2 might need some wordsmithing on the specification because the specification for getClassLoadingLock states "returns a dedicated object associated with the specified class name". The question is, what does "dedicated" mean here? Does it mean unique in the sense that the returned object is only associated with the given class in the current loader? Or can the object actually guard loading of multiple classes, possibly across different class loaders? So it seems that changing the specification will be inevitable if we wish to do something here. In which case lets go for something that more cleanly defines what we want to be doing: fully concurrent class-loading. Note: defineClassIfNotPresent is already implemented in the VM as find_or_define_class. It is only used if the AllowParallelDefineClass flag is set. This gives us an easy hook into existing VM mechanics. Proposal: Fully Concurrent ClassLoaders The proposal is that we expand on the notion of a parallel capable class loader and define a "fully concurrent parallel capable class loader" or fully concurrent loader, for short. A fully concurrent loader uses no synchronization in loadClass and the VM uses the "parallel define class" mechanism. For a fully concurrent loader getClassLoadingLock() can return null (or perhaps not - it doesn't matter as we won't use the result anyway). At present we have not made any changes to this method. All the parallel capable JDK classloaders become fully concurrent loaders. This doesn't require any code re-design as none of the mechanisms implemented rely on the per-name locking provided by the parallelLockMap. This seems to give us a path to remove all locking at the Java level during classloading, while retaining full compatibility with Java 7 parallel capable loaders. Fully concurrent loaders will still encounter the performance penalty associated with concurrent attempts to find and prepare a class's bytecode for definition by the VM. What this penalty is depends on the number of concurrent load attempts possible (a function of the number of threads and the application logic, and dependent on the number of processors), and the costs associated with finding and preparing the bytecodes. This obviously has to be measured across a range of applications. Preliminary webrevs: http://cr.openjdk.java.net/~dholmes/concurrent-loaders/webrev.hotspot/ http://cr.openjdk.java.net/~dholmes/concurrent-loaders/webrev.jdk/ Please direct all comments to the mailing list [email protected].

    Read the article

  • Scheduling thread tiles with C++ AMP

    - by Daniel Moth
    This post assumes you are totally comfortable with, what some of us call, the simple model of C++ AMP, i.e. you could write your own matrix multiplication. We are now ready to explore the tiled model, which builds on top of the non-tiled one. Tiling the extent We know that when we pass a grid (which is just an extent under the covers) to the parallel_for_each call, it determines the number of threads to schedule and their index values (including dimensionality). For the single-, two-, and three- dimensional cases you can go a step further and subdivide the threads into what we call tiles of threads (others may call them thread groups). So here is a single-dimensional example: extent<1> e(20); // 20 units in a single dimension with indices from 0-19 grid<1> g(e);      // same as extent tiled_grid<4> tg = g.tile<4>(); …on the 3rd line we subdivided the single-dimensional space into 5 single-dimensional tiles each having 4 elements, and we captured that result in a concurrency::tiled_grid (a new class in amp.h). Let's move on swiftly to another example, in pictures, this time 2-dimensional: So we start on the left with a grid of a 2-dimensional extent which has 8*6=48 threads. We then have two different examples of tiling. In the first case, in the middle, we subdivide the 48 threads into tiles where each has 4*3=12 threads, hence we have 2*2=4 tiles. In the second example, on the right, we subdivide the original input into tiles where each has 2*2=4 threads, hence we have 4*3=12 tiles. Notice how you can play with the tile size and achieve different number of tiles. The numbers you pick must be such that the original total number of threads (in our example 48), remains the same, and every tile must have the same size. Of course, you still have no clue why you would do that, but stick with me. First, we should see how we can use this tiled_grid, since the parallel_for_each function that we know expects a grid. Tiled parallel_for_each and tiled_index It turns out that we have additional overloads of parallel_for_each that accept a tiled_grid instead of a grid. However, those overloads, also expect that the lambda you pass in accepts a concurrency::tiled_index (new in amp.h), not an index<N>. So how is a tiled_index different to an index? A tiled_index object, can have only 1 or 2 or 3 dimensions (matching exactly the tiled_grid), and consists of 4 index objects that are accessible via properties: global, local, tile_origin, and tile. The global index is the same as the index we know and love: the global thread ID. The local index is the local thread ID within the tile. The tile_origin index returns the global index of the thread that is at position 0,0 of this tile, and the tile index is the position of the tile in relation to the overall grid. Confused? Here is an example accompanied by a picture that hopefully clarifies things: array_view<int, 2> data(8, 6, p_my_data); parallel_for_each(data.grid.tile<2,2>(), [=] (tiled_index<2,2> t_idx) restrict(direct3d) { /* todo */ }); Given the code above and the picture on the right, what are the values of each of the 4 index objects that the t_idx variables exposes, when the lambda is executed by T (highlighted in the picture on the right)? If you can't work it out yourselves, the solution follows: t_idx.global       = index<2> (6,3) t_idx.local          = index<2> (0,1) t_idx.tile_origin = index<2> (6,2) t_idx.tile             = index<2> (3,1) Don't move on until you are comfortable with this… the picture really helps, so use it. Tiled Matrix Multiplication Example – part 1 Let's paste here the C++ AMP matrix multiplication example, bolding the lines we are going to change (can you guess what the changes will be?) 01: void MatrixMultiplyTiled_Part1(vector<float>& vC, const vector<float>& vA, const vector<float>& vB, int M, int N, int W) 02: { 03: 04: array_view<const float,2> a(M, W, vA); 05: array_view<const float,2> b(W, N, vB); 06: array_view<writeonly<float>,2> c(M, N, vC); 07: parallel_for_each(c.grid, 08: [=](index<2> idx) restrict(direct3d) { 09: 10: int row = idx[0]; int col = idx[1]; 11: float sum = 0.0f; 12: for(int i = 0; i < W; i++) 13: sum += a(row, i) * b(i, col); 14: c[idx] = sum; 15: }); 16: } To turn this into a tiled example, first we need to decide our tile size. Let's say we want each tile to be 16*16 (which assumes that we'll have at least 256 threads to process, and that c.grid.extent.size() is divisible by 256, and moreover that c.grid.extent[0] and c.grid.extent[1] are divisible by 16). So we insert at line 03 the tile size (which must be a compile time constant). 03: static const int TS = 16; ...then we need to tile the grid to have tiles where each one has 16*16 threads, so we change line 07 to be as follows 07: parallel_for_each(c.grid.tile<TS,TS>(), ...that means that our index now has to be a tiled_index with the same characteristics as the tiled_grid, so we change line 08 08: [=](tiled_index<TS, TS> t_idx) restrict(direct3d) { ...which means, without changing our core algorithm, we need to be using the global index that the tiled_index gives us access to, so we insert line 09 as follows 09: index<2> idx = t_idx.global; ...and now this code just works and it is tiled! Closing thoughts on part 1 The process we followed just shows the mechanical transformation that can take place from the simple model to the tiled model (think of this as step 1). In fact, when we wrote the matrix multiplication example originally, the compiler was doing this mechanical transformation under the covers for us (and it has additional smarts to deal with the cases where the total number of threads scheduled cannot be divisible by the tile size). The point is that the thread scheduling is always tiled, even when you use the non-tiled model. But with this mechanical transformation, we haven't gained anything… Hint: our goal with explicitly using the tiled model is to gain even more performance. In the next post, we'll evolve this further (beyond what the compiler can automatically do for us, in this first release), so you can see the full usage of the tiled model and its benefits… Comments about this post by Daniel Moth welcome at the original blog.

    Read the article

  • We've completed the first iteration

    - by CliveT
    There are a lot of features in C# that are implemented by the compiler and not by the underlying platform. One such feature is a lambda expression. Since local variables cannot be accessed once the current method activation finishes, the compiler has to go out of its way to generate a new class which acts as a home for any variable whose lifetime needs to be extended past the activation of the procedure. Take the following example:     Random generator = new Random();     Func func = () = generator.Next(10); In this case, the compiler generates a new class called c_DisplayClass1 which is marked with the CompilerGenerated attribute. [CompilerGenerated] private sealed class c__DisplayClass1 {     // Fields     public Random generator;     // Methods     public int b__0()     {         return this.generator.Next(10);     } } Two quick comments on this: (i)    A display was the means that compilers for languages like Algol recorded the various lexical contours of the nested procedure activations on the stack. I imagine that this is what has led to the name. (ii)    It is a shame that the same attribute is used to mark all compiler generated classes as it makes it hard to figure out what they are being used for. Indeed, you could imagine optimisations that the runtime could perform if it knew that classes corresponded to certain high level concepts. We can see that the local variable generator has been turned into a field in the class, and the body of the lambda expression has been turned into a method of the new class. The code that builds the Func object simply constructs an instance of this class and initialises the fields to their initial values.     c__DisplayClass1 class2 = new c__DisplayClass1();     class2.generator = new Random();     Func func = new Func(class2.b__0); Reflector already contains code to spot this pattern of code and reproduce the form containing the lambda expression, so this is example is correctly decompiled. The use of compiler generated code is even more spectacular in the case of iterators. C# introduced the idea of a method that could automatically store its state between calls, so that it can pick up where it left off. The code can express the logical flow with yield return and yield break denoting places where the method should return a particular value and be prepared to resume.         {             yield return 1;             yield return 2;             yield return 3;         } Of course, there was already a .NET pattern for expressing the idea of returning a sequence of values with the computation proceeding lazily (in the sense that the work for the next value is executed on demand). This is expressed by the IEnumerable interface with its Current property for fetching the current value and the MoveNext method for forcing the computation of the next value. The sequence is terminated when this method returns false. The C# compiler links these two ideas together so that an IEnumerator returning method using the yield keyword causes the compiler to produce the implementation of an Iterator. Take the following piece of code.         IEnumerable GetItems()         {             yield return 1;             yield return 2;             yield return 3;         } The compiler implements this by defining a new class that implements a state machine. This has an integer state that records which yield point we should go to if we are resumed. It also has a field that records the Current value of the enumerator and a field for recording the thread. This latter value is used for optimising the creation of iterator instances. [CompilerGenerated] private sealed class d__0 : IEnumerable, IEnumerable, IEnumerator, IEnumerator, IDisposable {     // Fields     private int 1__state;     private int 2__current;     public Program 4__this;     private int l__initialThreadId; The body gets converted into the code to construct and initialize this new class. private IEnumerable GetItems() {     d__0 d__ = new d__0(-2);     d__.4__this = this;     return d__; } When the class is constructed we set the state, which was passed through as -2 and the current thread. public d__0(int 1__state) {     this.1__state = 1__state;     this.l__initialThreadId = Thread.CurrentThread.ManagedThreadId; } The state needs to be set to 0 to represent a valid enumerator and this is done in the GetEnumerator method which optimises for the usual case where the returned enumerator is only used once. IEnumerator IEnumerable.GetEnumerator() {     if ((Thread.CurrentThread.ManagedThreadId == this.l__initialThreadId)               && (this.1__state == -2))     {         this.1__state = 0;         return this;     } The state machine itself is implemented inside the MoveNext method. private bool MoveNext() {     switch (this.1__state)     {         case 0:             this.1__state = -1;             this.2__current = 1;             this.1__state = 1;             return true;         case 1:             this.1__state = -1;             this.2__current = 2;             this.1__state = 2;             return true;         case 2:             this.1__state = -1;             this.2__current = 3;             this.1__state = 3;             return true;         case 3:             this.1__state = -1;             break;     }     return false; } At each stage, the current value of the state is used to determine how far we got, and then we generate the next value which we return after recording the next state. Finally we return false from the MoveNext to signify the end of the sequence. Of course, that example was really simple. The original method body didn't have any local variables. Any local variables need to live between the calls to MoveNext and so they need to be transformed into fields in much the same way that we did in the case of the lambda expression. More complicated MoveNext methods are required to deal with resources that need to be disposed when the iterator finishes, and sometimes the compiler uses a temporary variable to hold the return value. Why all of this explanation? We've implemented the de-compilation of iterators in the current EAP version of Reflector (7). This contrasts with previous version where all you could do was look at the MoveNext method and try to figure out the control flow. There's a fair amount of things we have to do. We have to spot the use of a CompilerGenerated class which implements the Enumerator pattern. We need to go to the class and figure out the fields corresponding to the local variables. We then need to go to the MoveNext method and try to break it into the various possible states and spot the state transitions. We can then take these pieces and put them back together into an object model that uses yield return to show the transition points. After that Reflector can carry on optimising using its usual optimisations. The pattern matching is currently a little too sensitive to changes in the code generation, and we only do a limited analysis of the MoveNext method to determine use of the compiler generated fields. In some ways, it is a pity that iterators are compiled away and there is no metadata that reflects the original intent. Without it, we are always going to dependent on our knowledge of the compiler's implementation. For example, we have noticed that the Async CTP changes the way that iterators are code generated, so we'll have to do some more work to support that. However, with that warning in place, we seem to do a reasonable job of decompiling the iterators that are built into the framework. Hopefully, the EAP will give us a chance to find examples where we don't spot the pattern correctly or regenerate the wrong code, and we can improve things. Please give it a go, and report any problems.

    Read the article

  • Execution plan warnings–The final chapter

    - by Dave Ballantyne
    In my previous posts (here and here), I showed examples of some of the execution plan warnings that have been added to SQL Server 2012.  There is one other warning that is of interest to me : “Unmatched Indexes”. Firstly, how do I know this is the final one ?  The plan is an XML document, right ? So that means that it can have an accompanying XSD.  As an XSD is a schema definition, we can poke around inside it to find interesting things that *could* be in the final XML file. The showplan schema is stored in the folder Microsoft SQL Server\110\Tools\Binn\schemas\sqlserver\2004\07\showplan and by comparing schemas over releases you can get a really good idea of any new functionality that has been added. Here is the section of the Sql Server 2012 showplan schema that has been interesting me so far : <xsd:complexType name="AffectingConvertWarningType"> <xsd:annotation> <xsd:documentation>Warning information for plan-affecting type conversion</xsd:documentation> </xsd:annotation> <xsd:sequence> <!-- Additional information may go here when available --> </xsd:sequence> <xsd:attribute name="ConvertIssue" use="required"> <xsd:simpleType> <xsd:restriction base="xsd:string"> <xsd:enumeration value="Cardinality Estimate" /> <xsd:enumeration value="Seek Plan" /> <!-- to be extended here --> </xsd:restriction> </xsd:simpleType> </xsd:attribute> <xsd:attribute name="Expression" type ="xsd:string" use="required" /></xsd:complexType><xsd:complexType name="WarningsType"> <xsd:annotation> <xsd:documentation>List of all possible iterator or query specific warnings (e.g. hash spilling, no join predicate)</xsd:documentation> </xsd:annotation> <xsd:choice minOccurs="1" maxOccurs="unbounded"> <xsd:element name="ColumnsWithNoStatistics" type="shp:ColumnReferenceListType" minOccurs="0" maxOccurs="1" /> <xsd:element name="SpillToTempDb" type="shp:SpillToTempDbType" minOccurs="0" maxOccurs="unbounded" /> <xsd:element name="Wait" type="shp:WaitWarningType" minOccurs="0" maxOccurs="unbounded" /> <xsd:element name="PlanAffectingConvert" type="shp:AffectingConvertWarningType" minOccurs="0" maxOccurs="unbounded" /> </xsd:choice> <xsd:attribute name="NoJoinPredicate" type="xsd:boolean" use="optional" /> <xsd:attribute name="SpatialGuess" type="xsd:boolean" use="optional" /> <xsd:attribute name="UnmatchedIndexes" type="xsd:boolean" use="optional" /> <xsd:attribute name="FullUpdateForOnlineIndexBuild" type="xsd:boolean" use="optional" /></xsd:complexType> I especially like the “to be extended here” comment,  high hopes that we will see more of these in the future.   So “Unmatched Indexes” was a warning that I couldn’t get and many thanks must go to Fabiano Amorim (b|t) for showing me the way.   Filtered indexes were introduced in Sql Server 2008 and are really useful if you only need to index only a portion of the data within a table.  However,  if your SQL code uses a variable as a predicate on the filtered data that matches the filtered condition, then the filtered index cannot be used as, naturally,  the value in the variable may ( and probably will ) change and therefore will need to read data outside the index.  As an aside,  you could use option(recompile) here , in which case the optimizer will build a plan specific to the variable values and use the filtered index,  but that can bring about other problems.   To demonstrate this warning, we need to generate some test data :   DROP TABLE #TestTab1GOCREATE TABLE #TestTab1 (Col1 Int not null, Col2 Char(7500) not null, Quantity Int not null)GOINSERT INTO #TestTab1 VALUES (1,1,1),(1,2,5),(1,2,10),(1,3,20), (2,1,101),(2,2,105),(2,2,110),(2,3,120)GO and then add a filtered index CREATE INDEX ixFilter ON #TestTab1 (Col1)WHERE Quantity = 122 Now if we execute SELECT COUNT(*) FROM #TestTab1 WHERE Quantity = 122 We will see the filtered index being scanned But if we parameterize the query DECLARE @i INT = 122SELECT COUNT(*) FROM #TestTab1 WHERE Quantity = @i The plan is very different a table scan, as the value of the variable used in the predicate can change at run time, and also we see the familiar warning triangle. If we now look at the properties pane, we will see two pieces of information “Warnings” and “UnmatchedIndexes”. So, handily, we are being told which filtered index is not being used due to parameterization.

    Read the article

  • Of transactions and Mongo

    - by Nuri Halperin
    Originally posted on: http://geekswithblogs.net/nuri/archive/2014/05/20/of-transactions-and-mongo-again.aspxWhat's the first thing you hear about NoSQL databases? That they lose your data? That there's no transactions? No joins? No hope for "real" applications? Well, you *should* be wondering whether a certain of database is the right one for your job. But if you do so, you should be wondering that about "traditional" databases as well! In the spirit of exploration let's take a look at a common challenge: You are a bank. You have customers with accounts. Customer A wants to pay B. You want to allow that only if A can cover the amount being transferred. Let's looks at the problem without any context of any database engine in mind. What would you do? How would you ensure that the amount transfer is done "properly"? Would you prevent a "transaction" from taking place unless A can cover the amount? There are several options: Prevent any change to A's account while the transfer is taking place. That boils down to locking. Apply the change, and allow A's balance to go below zero. Charge person A some interest on the negative balance. Not friendly, but certainly a choice. Don't do either. Options 1 and 2 are difficult to attain in the NoSQL world. Mongo won't save you headaches here either. Option 3 looks a bit harsh. But here's where this can go: ledger. See, and account doesn't need to be represented by a single row in a table of all accounts with only the current balance on it. More often than not, accounting systems use ledgers. And entries in ledgers - as it turns out – don't actually get updated. Once a ledger entry is written, it is not removed or altered. A transaction is represented by an entry in the ledger stating and amount withdrawn from A's account and an entry in the ledger stating an addition of said amount to B's account. For sake of space-saving, that entry in the ledger can happen using one entry. Think {Timestamp, FromAccountId, ToAccountId, Amount}. The implication of the original question – "how do you enforce non-negative balance rule" then boils down to: Insert entry in ledger Run validation of recent entries Insert reverse entry to roll back transaction if validation failed. What is validation? Sum up the transactions that A's account has (all deposits and debits), and ensure the balance is positive. For sake of efficiency, one can roll up transactions and "close the book" on transactions with a pseudo entry stating balance as of midnight or something. This lets you avoid doing math on the fly on too many transactions. You simply run from the latest "approved balance" marker to date. But that's an optimization, and premature optimizations are the root of (some? most?) evil.. Back to some nagging questions though: "But mongo is only eventually consistent!" Well, yes, kind of. It's not actually true that Mongo has not transactions. It would be more descriptive to say that Mongo's transaction scope is a single document in a single collection. A write to a Mongo document happens completely or not at all. So although it is true that you can't update more than one documents "at the same time" under a "transaction" umbrella as an atomic update, it is NOT true that there' is no isolation. So a competition between two concurrent updates is completely coherent and the writes will be serialized. They will not scribble on the same document at the same time. In our case - in choosing a ledger approach - we're not even trying to "update" a document, we're simply adding a document to a collection. So there goes the "no transaction" issue. Now let's turn our attention to consistency. What you should know about mongo is that at any given moment, only on member of a replica set is writable. This means that the writable instance in a set of replicated instances always has "the truth". There could be a replication lag such that a reader going to one of the replicas still sees "old" state of a collection or document. But in our ledger case, things fall nicely into place: Run your validation against the writable instance. It is guaranteed to have a ledger either with (after) or without (before) the ledger entry got written. No funky states. Again, the ledger writing *adds* a document, so there's no inconsistent document state to be had either way. Next, we might worry about data loss. Here, mongo offers several write-concerns. Write-concern in Mongo is a mode that marshals how uptight you want the db engine to be about actually persisting a document write to disk before it reports to the application that it is "done". The most volatile, is to say you don't care. In that case, mongo would just accept your write command and say back "thanks" with no guarantee of persistence. If the server loses power at the wrong moment, it may have said "ok" but actually no written the data to disk. That's kind of bad. Don't do that with data you care about. It may be good for votes on a pole regarding how cute a furry animal is, but not so good for business. There are several other write-concerns varying from flushing the write to the disk of the writable instance, flushing to disk on several members of the replica set, a majority of the replica set or all of the members of a replica set. The former choice is the quickest, as no network coordination is required besides the main writable instance. The others impose extra network and time cost. Depending on your tolerance for latency and read-lag, you will face a choice of what works for you. It's really important to understand that no data loss occurs once a document is flushed to an instance. The record is on disk at that point. From that point on, backup strategies and disaster recovery are your worry, not loss of power to the writable machine. This scenario is not different from a relational database at that point. Where does this leave us? Oh, yes. Eventual consistency. By now, we ensured that the "source of truth" instance has the correct data, persisted and coherent. But because of lag, the app may have gone to the writable instance, performed the update and then gone to a replica and looked at the ledger there before the transaction replicated. Here are 2 options to deal with this. Similar to write concerns, mongo support read preferences. An app may choose to read only from the writable instance. This is not an awesome choice to make for every ready, because it just burdens the one instance, and doesn't make use of the other read-only servers. But this choice can be made on a query by query basis. So for the app that our person A is using, we can have person A issue the transfer command to B, and then if that same app is going to immediately as "are we there yet?" we'll query that same writable instance. But B and anyone else in the world can just chill and read from the read-only instance. They have no basis to expect that the ledger has just been written to. So as far as they know, the transaction hasn't happened until they see it appear later. We can further relax the demand by creating application UI that reacts to a write command with "thank you, we will post it shortly" instead of "thank you, we just did everything and here's the new balance". This is a very powerful thing. UI design for highly scalable systems can't insist that the all databases be locked just to paint an "all done" on screen. People understand. They were trained by many online businesses already that your placing of an order does not mean that your product is already outside your door waiting (yes, I know, large retailers are working on it... but were' not there yet). The second thing we can do, is add some artificial delay to a transaction's visibility on the ledger. The way that works is simply adding some logic such that the query against the ledger never nets a transaction for customers newer than say 15 minutes and who's validation flag is not set. This buys us time 2 ways: Replication can catch up to all instances by then, and validation rules can run and determine if this transaction should be "negated" with a compensating transaction. In case we do need to "roll back" the transaction, the backend system can place the timestamp of the compensating transaction at the exact same time or 1ms after the original one. Effectively, once A or B visits their ledger, both transactions would be visible and the overall balance "as of now" would reflect no change.  The 2 transactions (attempted/ reverted) would be visible , since we do actually account for the attempt. Hold on a second. There's a hole in the story: what if several transfers from A to some accounts are registered, and 2 independent validators attempt to compute the balance concurrently? Is there a chance that both would conclude non-sufficient-funds even though rolling back transaction 100 would free up enough for transaction 117 (some random later transaction)? Yes. there is that chance. But the integrity of the business rule is not compromised, since the prime rule is don't dispense money you don't have. To minimize or eliminate this scenario, we can also assign a single validation process per origin account. This may seem non-scalable, but it can easily be done as a "sharded" distribution. Say we have 11 validation threads (or processing nodes etc.). We divide the account number space such that each validator is exclusively responsible for a certain range of account numbers. Sounds cunningly similar to Mongo's sharding strategy, doesn't it? Each validator then works in isolation. More capacity needed? Chop the account space into more chunks. So where  are we now with the nagging questions? "No joins": Huh? What are those for? "No transactions": You mean no cross-collection and no cross-document transactions? Granted - but don't always need them either. "No hope for real applications": well... There are more issues and edge cases to slog through, I'm sure. But hopefully this gives you some ideas of how to solve common problems without distributed locking and relational databases. But then again, you can choose relational databases if they suit your problem.

    Read the article

  • Query optimization using composite indexes

    - by xmarch
    Many times, during the process of creating a new Coherence application, developers do not pay attention to the way cache queries are constructed; they only check that these queries comply with functional specs. Later, performance testing shows that these perform poorly and it is then when developers start working on improvements until the non-functional performance requirements are met. This post describes the optimization process of a real-life scenario, where using a composite attribute index has brought a radical improvement in query execution times.  The execution times went down from 4 seconds to 2 milliseconds! E-commerce solution based on Oracle ATG – Endeca In the context of a new e-commerce solution based on Oracle ATG – Endeca, Oracle Coherence has been used to calculate and store SKU prices. In this architecture, a Coherence cache stores the final SKU prices used for Endeca baseline indexing. Each SKU price is calculated from a base SKU price and a series of calculations based on information from corporate global discounts. Corporate global discounts information is stored in an auxiliary Coherence cache with over 800.000 entries. In particular, to obtain each price the process needs to execute six queries over the global discount cache. After the implementation was finished, we discovered that the most expensive steps in the price calculation discount process were the global discounts cache query. This query has 10 parameters and is executed 6 times for each SKU price calculation. The steps taken to optimise this query are described below; Starting point Initial query was: String filter = "levelId = :iLevelId AND  salesCompanyId = :iSalesCompanyId AND salesChannelId = :iSalesChannelId "+ "AND departmentId = :iDepartmentId AND familyId = :iFamilyId AND brand = :iBrand AND manufacturer = :iManufacturer "+ "AND areaId = :iAreaId AND endDate >=  :iEndDate AND startDate <= :iStartDate"; Map<String, Object> params = new HashMap<String, Object>(10); // Fill all parameters. params.put("iLevelId", xxxx); // Executing filter. Filter globalDiscountsFilter = QueryHelper.createFilter(filter, params); NamedCache globalDiscountsCache = CacheFactory.getCache(CacheConstants.GLOBAL_DISCOUNTS_CACHE_NAME); Set applicableDiscounts = globalDiscountsCache.entrySet(globalDiscountsFilter); With the small dataset used for development the cache queries performed very well. However, when carrying out performance testing with a real-world sample size of 800,000 entries, each query execution was taking more than 4 seconds. First round of optimizations The first optimisation step was the creation of separate Coherence index for each of the 10 attributes used by the filter. This avoided object deserialization while executing the query. Each index was created as follows: globalDiscountsCache.addIndex(new ReflectionExtractor("getXXX" ) , false, null); After adding these indexes the query execution time was reduced to between 450 ms and 1s. However, these execution times were still not good enough.  Second round of optimizations In this optimisation phase a Coherence query explain plan was used to identify how many entires each index reduced the results set by, along with the cost in ms of executing that part of the query. Though the explain plan showed that all the indexes for the query were being used, it also showed that the ordering of the query parameters was "sub-optimal".  Parameters associated to object attributes with high-cardinality should appear at the beginning of the filter, or more specifically, the attributes that filters out the highest of number records should be placed at the beginning. But examining corporate global discount data we realized that depending on the values of the parameters used in the query the “good” order for the attributes was different. In particular, if the attributes brand and family had specific values it was more optimal to have a different query changing the order of the attributes. Ultimately, we ended up with three different optimal variants of the query that were used in its relevant cases: String filter = "brand = :iBrand AND familyId = :iFamilyId AND departmentId = :iDepartmentId AND levelId = :iLevelId "+ "AND manufacturer = :iManufacturer AND endDate >= :iEndDate AND salesCompanyId = :iSalesCompanyId "+ "AND areaId = :iAreaId AND salesChannelId = :iSalesChannelId AND startDate <= :iStartDate"; String filter = "familyId = :iFamilyId AND departmentId = :iDepartmentId AND levelId = :iLevelId AND brand = :iBrand "+ "AND manufacturer = :iManufacturer AND endDate >=  :iEndDate AND salesCompanyId = :iSalesCompanyId "+ "AND areaId = :iAreaId  AND salesChannelId = :iSalesChannelId AND startDate <= :iStartDate"; String filter = "brand = :iBrand AND departmentId = :iDepartmentId AND familyId = :iFamilyId AND levelId = :iLevelId "+ "AND manufacturer = :iManufacturer AND endDate >= :iEndDate AND salesCompanyId = :iSalesCompanyId "+ "AND areaId = :iAreaId AND salesChannelId = :iSalesChannelId AND startDate <= :iStartDate"; Using the appropriate query depending on the value of brand and family parameters the query execution time dropped to between 100 ms and 150 ms. But these these execution times were still not good enough and the solution was cumbersome. Third and last round of optimizations The third and final optimization was to introduce a composite index. However, this did mean that it was not possible to use the Coherence Query Language (CohQL), as composite indexes are not currently supporte in CohQL. As the original query had 8 parameters using EqualsFilter, 1 using GreaterEqualsFilter and 1 using LessEqualsFilter, the composite index was built for the 8 attributes using EqualsFilter. The final query had an EqualsFilter for the multiple extractor, a GreaterEqualsFilter and a LessEqualsFilter for the 2 remaining attributes.  All individual indexes were dropped except the ones being used for LessEqualsFilter and GreaterEqualsFilter. We were now running in an scenario with an 8-attributes composite filter and 2 single attribute filters. The composite index created was as follows: ValueExtractor[] ve = { new ReflectionExtractor("getSalesChannelId" ), new ReflectionExtractor("getLevelId" ),    new ReflectionExtractor("getAreaId" ), new ReflectionExtractor("getDepartmentId" ),    new ReflectionExtractor("getFamilyId" ), new ReflectionExtractor("getManufacturer" ),    new ReflectionExtractor("getBrand" ), new ReflectionExtractor("getSalesCompanyId" )}; MultiExtractor me = new MultiExtractor(ve); NamedCache globalDiscountsCache = CacheFactory.getCache(CacheConstants.GLOBAL_DISCOUNTS_CACHE_NAME); globalDiscountsCache.addIndex(me, false, null); And the final query was: ValueExtractor[] ve = { new ReflectionExtractor("getSalesChannelId" ), new ReflectionExtractor("getLevelId" ),    new ReflectionExtractor("getAreaId" ), new ReflectionExtractor("getDepartmentId" ),    new ReflectionExtractor("getFamilyId" ), new ReflectionExtractor("getManufacturer" ),    new ReflectionExtractor("getBrand" ), new ReflectionExtractor("getSalesCompanyId" )}; MultiExtractor me = new MultiExtractor(ve); // Fill composite parameters.String SalesCompanyId = xxxx;...AndFilter composite = new AndFilter(new EqualsFilter(me,                   Arrays.asList(iSalesChannelId, iLevelId, iAreaId, iDepartmentId, iFamilyId, iManufacturer, iBrand, SalesCompanyId)),                                     new GreaterEqualsFilter(new ReflectionExtractor("getEndDate" ), iEndDate)); AndFilter finalFilter = new AndFilter(composite, new LessEqualsFilter(new ReflectionExtractor("getStartDate" ), iStartDate)); NamedCache globalDiscountsCache = CacheFactory.getCache(CacheConstants.GLOBAL_DISCOUNTS_CACHE_NAME); Set applicableDiscounts = globalDiscountsCache.entrySet(finalFilter);      Using this composite index the query improved dramatically and the execution time dropped to between 2 ms and  4 ms.  These execution times completely met the non-functional performance requirements . It should be noticed than when using the composite index the order of the attributes inside the ValueExtractor was not relevant.

    Read the article

  • Controlling the Sizing of the af:messages Dialog

    - by Duncan Mills
    Over the last day or so a small change in behaviour between 11.1.2.n releases of ADF and earlier versions has come to my attention. This has concerned the default sizing of the dialog that the framework automatically generates to handle the display of JSF messages being handled by the <af:messages> component. Unlike a normal popup, you don't have a physical <af:dialog> or <af:window> to set the sizing on in your page definition, so you're at the mercy of what the framework provides. In this case the framework now defines a fixed 250x250 pixel content area dialog for these messages, which can look a bit weird if the message is either very short, or very long. Unfortunately this is not something that you can control through the skin, instead you have to be a little more creative. Here's the solution I've come up with.  Unfortunately, I've not found a supportable way to reset the dialog so as to say  just size yourself based on your contents, it is actually possible to do this by tweaking the correct DOM objects, but I wanted to start with a mostly supportable solution that only uses the best practice of working through the ADF client side APIs. The Technique The basic approach I've taken is really very simple.  The af:messages dialog is just a normal richDialog object, it just happens to be one that is pre-defined for you with a particular known name "msgDlg" (which hopefully won't change). Knowing this, you can call the accepted APIs to control the content width and height of that dialog, as our meerkat friends would say, "simples" 1 The JavaScript For this example I've defined three JavaScript functions.   The first does all the hard work and is designed to be called from server side Java or from a page load event to set the default. The second is a utility function used by the first to validate the values you're about to use for height and width. The final function is one that can be called from the page load event to set an initial default sizing if that's all you need to do. Function resizeDefaultMessageDialog() /**  * Function that actually resets the default message dialog sizing.  * Note that the width and height supplied define the content area  * So the actual physical dialog size will be larger to account for  * the chrome containing the header / footer etc.  * @param docId Faces component id of the document  * @param contentWidth - new content width you need  * @param contentHeight - new content height  */ function resizeDefaultMessageDialog(docId, contentWidth, contentHeight) {   // Warning this value may change from release to release   var defMDName = "::msgDlg";   //Find the default messages dialog   msgDialogComponent = AdfPage.PAGE.findComponentByAbsoluteId(docId + defMDName); // In your version add a check here to ensure we've found the right object!   // Check the new width is supplied and is a positive number, if so apply it.   if (dimensionIsValid(contentWidth)){       msgDialogComponent.setContentWidth(contentWidth);   }   // Check the new height is supplied and is a positive number, if so apply it.   if (dimensionIsValid(contentHeight)){       msgDialogComponent.setContentHeight(contentHeight);   } }  Function dimensionIsValid()  /**  * Simple function to check that sensible numeric values are   * being proposed for a dimension  * @param sampleDimension   * @return booolean  */ function dimensionIsValid(sampleDimension){     return (!isNaN(sampleDimension) && sampleDimension > 0); } Function  initializeDefaultMessageDialogSize() /**  * This function will re-define the default sizing applied by the framework   * in 11.1.2.n versions  * It is designed to be called with the document onLoad event  */ function initializeDefaultMessageDialogSize(loadEvent){   //get the configuration information   var documentId = loadEvent.getSource().getProperty('documentId');   var newWidth = loadEvent.getSource().getProperty('defaultMessageDialogContentWidth');   var newHeight = loadEvent.getSource().getProperty('defaultMessageDialogContentHeight');   resizeDefaultMessageDialog(documentId, newWidth, newHeight); } Wiring in the Functions As usual, the first thing we need to do when using JavaScript with ADF is to define an af:resource  in the document metaContainer facet <af:document>   ....     <f:facet name="metaContainer">     <af:resource type="javascript" source="/resources/js/hackMessagedDialog.js"/>    </f:facet> </af:document> This makes the script functions available to call.  Next if you want to use the option of defining an initial default size for the dialog you use a combination of <af:clientListener> and <af:clientAttribute> tags like this. <af:document title="MyApp" id="doc1">   <af:clientListener method="initializeDefaultMessageDialogSize" type="load"/>   <af:clientAttribute name="documentId" value="doc1"/>   <af:clientAttribute name="defaultMessageDialogContentWidth" value="400"/>   <af:clientAttribute name="defaultMessageDialogContentHeight" value="150"/>  ...   Just in Time Dialog Sizing  So  what happens if you have a variety of messages that you might add and in some cases you need a small dialog and an other cases a large one? Well in that case you can re-size these dialogs just before you submit the message. Here's some example Java code: FacesContext ctx = FacesContext.getCurrentInstance();          //reset the default dialog size for this message ExtendedRenderKitService service =              Service.getRenderKitService(ctx, ExtendedRenderKitService.class); service.addScript(ctx, "resizeDefaultMessageDialog('doc1',100,50);");          FacesMessage msg = new FacesMessage("Short message"); msg.setSeverity(FacesMessage.SEVERITY_ERROR); ctx.addMessage(null, msg);  So there you have it. This technique should, at least, allow you to control the dialog sizing just enough to stop really objectionable whitespace or scrollbars. 1 Don't worry if you don't get the reference, lest's just say my kids watch too many adverts.

    Read the article

< Previous Page | 370 371 372 373 374 375 376 377 378 379 380 381  | Next Page >