Search Results

Search found 11936 results on 478 pages for 'computed field'.

Page 146/478 | < Previous Page | 142 143 144 145 146 147 148 149 150 151 152 153  | Next Page >

  • Partial recalculation of visibility on a 2D uniform grid

    - by Martin Källman
    Problem Imagine that we have a 2D uniform grid of dimensions N x N. For this grid we have also pre-computed a visibility look-up table, e.g. with DDA, which answers the boolean query is cell X visible from cell Y? The look-up table is a complete graph KN of the cells V in the grid, with each edge E being a binary value denoting the visibility between its vertices. Question If any given cell has its visibility modified, is it possible to extract the subset Edelta of edges which must have their visibility recomputed due to the change, so as to avoid a full-on recomputation for the entire grid? (Which is N(N-1) / 2 or N2 depending on the implementation) Update If is not possible to solve thi in closed form, then maintaining a separate mapping of each cell and every cell pair who's line intersects said cell might also be an option. This obviously consumes more memory, but the data is static. The increased memory requirement could be reduced by introducing a hierarchy, subdividing the grid into smaller parts, and by doing so the above mapping can be reused for each sub-grid. This would come at a cost in terms of increased computation relative to the number of subdivisions; also requiring a resumable ray-casting algorithm.

    Read the article

  • How lookaheads are propagated in "channel" method of building LALR parser?

    - by greenoldman
    The method is described in Dragon Book, however I read about it in ""Parsing Techniques" by D.Grune and C.J.H.Jacobs". I start from my understanding of building channels for NFA: channels are built once, they are like water channels with current you "drop" lookahead symbols in right places (sources) of the channel, and they propagate with "current" when symbol propagates, there are no barriers (the only sufficient things for propagation are presence of channel and direction/current); i.e. lookahead cannot just die out of the blue Is that right? If I am correct, then eof lookahead should be present in all states, because the source of it is the start production, and all other production states are reachable from start state. How the DFA is made out of this NFA is not perfectly clear for me -- the authors of the mentioned book write about preserving channels, but I see no purpose, if you propagated lookaheads. If the channels have to be preserved, are they cut off from the source if the DFA state does not include source NFA state? I assume no -- the channels still runs between DFA states, not only within given DFA state. In the effect eof should still be present in all items in all states. But when you take a look at DFA presented in book (pdf is from errata): DFA for LALR (fig. 9.34 in the book, p.301) you will see there are items without eof in lookahead. The grammar for this DFA is: S -> E E -> E - T E -> T T -> ( E ) T -> n So how it was computed, when eof was dropped, and on what condition? Update It is textual pdf, so two interesting states (in DFA; # is eof): State 1: S--- >•E[#] E--- >•E-T[#-] E--- >•T[#-] T--- >•n[#-] T--- >•(E)[#-] State 6: T--- >(•E)[#-)] E--- >•E-T[-)] E--- >•T[-)] T--- >•n[-)] T--- >•(E)[-)] Arc from 1 to 6 is labeled (.

    Read the article

  • learn the programming language for computing functions about integers

    - by asd
    Hi I know something about Pascal, Mathematica and Matlab, but I dont have any idea about C,C++,C# languages. I want to learn one of the languages that they they are fast and exact to compute some arithmetic functions for large numbers(for example larger than $10^3000$). I asked somebody and he said he used C++ and he said I computed this sequence in less than 10 min. I want to know C, C++, C# and visual kind of theses programs and know which is better for my goal. Let $f$ be an arithmetic function and A={k1,k2,...,kn} are integers in increasing order. Now I want to start with k1 and compare f(ki) with f(k1). If f(ki)f(k1), put ki as k1. Now start with ki, and compare f(kj) with f(ki), for ji. If f(kj)f(ki), put kj as ki, and repeat this procedure. At the end we will have a sub sequence B={L1,...,Lm} of A by this property: f(L(i+1))f(L(i)), for any 1<=i<=m-1 I have written a code for this program with Mathematica, and it take some hours to compute f of ki's or the set B for large numbers. For example, let f is the divisor function of integers. Do you know how to write the code for my purpose in Mathematica or Matlab. Mathematica is preferable.

    Read the article

  • Performing client-side OAuth authorized Twitter API calls versus server side, how much of a difference is there in terms of performance?

    - by Terence Ponce
    I'm working on a Twitter application in Ruby on Rails. One of the biggest arguments that I have with other people on the project is the method of calling the Twitter API. Before, everything was done on the server: OAuth login, updating the user's Twitter data, and retrieving tweets. Retrieving tweets was the heaviest thing to do since we don't store the tweets in our database, so viewing the tweets means that we have to call the API every time. One of the people in the project suggested that we call the tweets through Javascript instead to lessen the load on the server. We used GET search, which, correct me if I'm wrong, will be removed when v1.0 becomes completely deprecated, but that really isn't a concern now. When the Twitter API has migrated completely to v1.1 (again, correct me if I'm wrong), every calls to the API must be authenticated, so we have to authenticate our Javascript requests to the API. As said here: We don't support or recommend performing OAuth directly through Javascript -- it's insecure and puts your application at risk. The only acceptable way to perform it is if you kept all keys and secrets server-side, computed the OAuth signatures and parameters server side, then issued the request client-side from the server-generated OAuth values. If we do exactly what Twitter suggests, the only difference between this and doing everything server-side is that our server won't have to contact the Twitter API anymore every time the user wants to view tweets. Here's how I would picture what's happening every time the user makes a request: If we do it through Javascript, it would be harder on my part because I would have to create the signatures manually for every request, but I will gladly do it if the boost in performance is worth all the trouble. Doing it through Ruby on Rails would be very easy since the Twitter gem does most of the grunt work already, so I'm really encouraging the other people in the project to agree with me. Is the difference in performance trivial or is it significant enough to switch to Javascript?

    Read the article

  • Where to place the R code for R+Sweave+LaTeX workflow

    - by claytontstanley
    I spent the last week learning 3 new tools: R, Sweave, and LaTeX. One question that came to my mind though when working through my first project: Where do I place the majority of the R code? The tutorials that I read online placed the majority of the R code in the LaTeX .Rnw file. However, I find having a bunch of R calculations in the LaTeX file distracting. What I do find extremely helpful (of course) is to call out to R code in the LaTeX file and embed the result. So the workflow I've been using is to place 99% of my R code in my .R file. I run that file first, save a bunch of calculations as objects, and output the .Rout file once finished (to save the work). Then when running Sweave, I load up that .Rout file, so that I have the majority of my calculations already completed and in the Sweave R session. Then my LaTeX callouts to R are quite simple: Just give me the XTable stored in 'res.table', or give me the result of an already-computed calculation stored in the variable 'res'. So I push towards the minimal amount of R code in the LaTex file possible, to achieve the desired result (embedding stats results in the LaTeX writeup). Does anyone have any experience with this approach? I'm just worried I might run into trouble further down the line, when I start really trying to load up and leverage this workflow.

    Read the article

  • How to notify client about updated UpdatePanel content on server side

    - by csh1981
    I have a problem with UpdatePanel.Update() which works initially but then stops. I have tumbled with this problem for some time and some background is needed so please read ahead. I have an ASP.net application in which I have a subpage that display computed information in graphs. Each graph is embedded in an UpdatePanel. The graph is a user control that uses the standard asp:Chart for display. My task is to enable this page with AJAX capabilities so the page is responsive during postbacks. When I access this page from another page, during the initial page rendering, I use a wait dialog for each graph and a pageload event on the client side. In the client event, a hidden button is clicked which a server event handles (the hidden button is inside an UpdatePanel so the postback is asynchronous). Each graph is computed and the UpdatePanels are in turn updated with the Chart content. This is done using UpdatePanel.Update. And it is successful. However, I also have some RadioButtons on the page. These are dynamically created. The purpose of them is to switch graph type --- to show the same data in a different way. Same type of time consuming computation is needed in order to do so. I subscribe on each RadioButton's OnCheckedChanged event and the postback is asynchronous since the radiobuttons are inside an UpdatePanel. In the server event handler I determine the type of graph and use this as an input to the Chart control. I then remove the old Chart control from my Panel and adds new Chart and then I call UpdatePanel.Update(). But with no success. Nothing happens, no errors, nothing. Why is this?? I think this is strange because if I compute every Chart data in the initial rendering instead of using the "Wait dialog"-solution described earlier then I can select graph types successfully and all subsequent AJAX requests work as intended. Also, the same code (computing the chart, removal, and adding the Chart control to Panel and UpdatePanel.Update()) is hit during the initial rendering of the page, and it works only the first time. Here is the method that computes the graph and adds it to the panel and update the UpdatePanel: public void UpdateGraph(GraphType type, GraphMapper mapper) { //Panel is the content of UpdatePanelGraph's Panel.Controls.Clear(); chart = new Chart(type, mapper); //Computation happens inside here panel.Controls.Add(chart); //UpdatePanelGraph is in UpdateMode Conditional and has //ChildrenAsTriggers set to false UpdatePanelGraph.Update(); } I really need a way for these radiobuttons to work, possible using some clientside JavaScript or another way of handling things on the server side. I have thought about using a JavaScript postback call on the UpdatePanel instead of the UpdatePanel.Update(). However, the issue I have here is how to notify the client side when the server side is finished with computing the graph? An plausible explanation of the strange behavior is also much appreciated. Any help appreciated, thanks

    Read the article

  • Where to put business logic in MVC design?

    - by BriskLabs Pakistan
    I have created a simple MVC java application that adds records through data forms to a database. my app collects data, it also validates it and stores it. This is because the data is being sourced online from different users. the data is mostly numeric in nature. now on the numeric data being stored into database (SQL server) , i wish that my app should be able to perform computations... and display it. the user is not interested in how computations are done so they must be encapsulated. the user must only be able to view the simple computed data which for example A column data - B Column data / C column data etc... and just display it to the user... i know how to write stored procedures for same but i want a 3 tier app I want the data, that I put into the database as a record, worked upon by performing calculations on it. However, the original data should remain unaffected, while the new data, post-calculations, must be stored as a new entity record into the database. Where should I write the code for this background calculation? As it is the rules and business logic... in a new java beans files ?

    Read the article

  • XNA hlsl tex2D() only reads 3 channels from normal maps and specular maps

    - by cubrman
    Our engine uses deferred rendering and at the main draw phase gathers plenty of data from the objects it draws. In order to save on tex2D calls, we packed our objects' specular maps with all sorts of data, so three out of four channels are already taken. To make it clear: I am talking about the assets that come with the models and are stored in their material's Specular Level channel, not about the RenderTarget. So now I need another information to be stored in the alpha channel, but I cannot make the shader to read it properly! Nomatter what I write into alpha it ends up being 1 (255)! I tried: saving the textures in PNG/TGA formats. turning off pre-computed alpha in model's properties. Out of every texture available to me (we use Diffuse map, Normal Map and Specular Map) I was only able to read alpha successfully from the Diffuse Map! Here is how I add specular and normal maps to my model's material in the content processor: if (geometry.Material.Textures.ContainsKey(normalMapKey)) { ExternalReference<TextureContent> texRef = geometry.Material.Textures[normalMapKey]; geometry.Material.Textures.Remove("NormalMap"); geometry.Material.Textures.Add("NormalMap", texRef); } ... foreach (KeyValuePair<String, ExternalReference<TextureContent>> texture in material.Textures) { if ((texture.Key == "Texture") || (texture.Key == "NormalMap") || (texture.Key == "SpecularMap")) mat.Textures.Add(texture.Key, texture.Value); } In the shader I obviously use: float4 data = tex2D(specularMapSampler, TexCoords); so data.a is always 1 in my case, could you suggest a reason?

    Read the article

  • Japanese Multiplication simulation - is a program actually capable of improving calculation speed?

    - by jt0dd
    On SuperUser, I asked a (possibly silly) question about processors using mathematical shortcuts and would like to have a look at the possibility at the software application of that concept. I'd like to write a simulation of Japanese Multiplication to get benchmarks on large calculations utilizing the shortcut vs traditional CPU multiplication. I'm curious as to whether it makes sense to try this. My Question: I'd like to know whether or not a software math shortcut, as described above is actually a shortcut at all. This is a question of programming concept. By utilizing the simulation of Japanese Multiplication, is a program actually capable of improving calculation speed? Or am I doomed from the start? The answer to this question isn't required to determine whether or not the experiment will succeed, but rather whether or not it's logically possible for such a thing to occur in any program, using this concept as an example. My theory is that since addition is computed faster than multiplication, a simulation of Japanese multiplication may actually allow a program to multiply (large) numbers faster than the CPU arithmetic unit can. I think this would be a very interesting finding, if it proves to be true. If, in the multiplication of numbers of any immense size, the shortcut were to calculate the result via less instructions (or faster) than traditional ALU multiplication, I would consider the experiment a success.

    Read the article

  • Execution plan warnings–All that glitters is not gold

    - by Dave Ballantyne
    In a previous post, I showed you the new execution plan warnings related to implicit and explicit warnings.  Pretty much as soon as i hit ’post’,  I noticed something rather odd happening. This statement : select top(10) SalesOrderHeader.SalesOrderID, SalesOrderNumberfrom Sales.SalesOrderHeaderjoin Sales.SalesOrderDetail on SalesOrderHeader.SalesOrderID = SalesOrderDetail.SalesOrderID   Throws the “Type conversion may affect cardinality estimation” warning.     Ive done no such conversion in my statement why would that be ?  Well, SalesOrderNumber is a computed column , “(isnull(N'SO'+CONVERT([nvarchar](23),[SalesOrderID],0),N'*** ERROR ***'))”,  so thats where the conversion is.   Wait!!! Am i saying that every type conversion will throw the warning ?  Thankfully, no.  It only appears for columns that are used in predicates ,even if the predicate / join condition is fine ,  and the column is indexed ( and/or , presumably has statistics).    Hopefully , this wont lead to to many wild goose chases, but is definitely something to bear in mind.  If you want to see this fixed then upvote my connect item here.

    Read the article

  • Best Terminology for a particular php-based site architecture? [closed]

    - by hen3ry
    For a site... o whose overall look-and-feel is generated by one php page ("index.php"). o in which "index.php" provides for all pages served the following required components: The DOCTYPE, opening html tag, the head section, the opening body tag, the end-body tag, and end-html tag. o which uses computed hierarchical navigation menus within "index.php" to offer visitors access to the site content. o in which all content is stored in individual files that contain "headerless html". (The DOCTYPE, etc. etc. being being provided by "index.php" as described above.) Q1: what term best describes this architecture? I'm seeking a concise descriptor that is useful in conversation and definitive as a search term, in whole-web searches, and searching here on Pro Webmasters. Q2: what term best describes the individual content files? Same general goals for the descriptor as above. As you see above, I couldn't avoid using the term "headerless html", my best choice. But this term does not seem to be in general use. I've found some people use this term to describe such as my content files, but others use it quite differently.

    Read the article

  • Java code optimization leads to numerical inaccuracies and errors

    - by rano
    I'm trying to implement a version of the Fuzzy C-Means algorithm in Java and I'm trying to do some optimization by computing just once everything that can be computed just once. This is an iterative algorithm and regarding the updating of a matrix, the clusters x pixels membership matrix U, this is the update rule I want to optimize: where the x are the element of a matrix X (pixels x features) and v belongs to the matrix V (clusters x features). And m is a parameter that ranges from 1.1 to infinity. The distance used is the euclidean norm. If I had to implement this formula in a banal way I'd do: for(int i = 0; i < X.length; i++) { int count = 0; for(int j = 0; j < V.length; j++) { double num = D[i][j]; double sumTerms = 0; for(int k = 0; k < V.length; k++) { double thisDistance = D[i][k]; sumTerms += Math.pow(num / thisDistance, (1.0 / (m - 1.0))); } U[i][j] = (float) (1f / sumTerms); } } In this way some optimization is already done, I precomputed all the possible squared distances between X and V and stored them in a matrix D but that is not enough, since I'm cycling througn the elements of V two times resulting in two nested loops. Looking at the formula the numerator of the fraction is independent of the sum so I can compute numerator and denominator independently and the denominator can be computed just once for each pixel. So I came to a solution like this: int nClusters = V.length; double exp = (1.0 / (m - 1.0)); for(int i = 0; i < X.length; i++) { int count = 0; for(int j = 0; j < nClusters; j++) { double distance = D[i][j]; double denominator = D[i][nClusters]; double numerator = Math.pow(distance, exp); U[i][j] = (float) (1f / (numerator * denominator)); } } Where I precomputed the denominator into an additional column of the matrix D while I was computing the distances: for (int i = 0; i < X.length; i++) { for (int j = 0; j < V.length; j++) { double sum = 0; for (int k = 0; k < nDims; k++) { final double d = X[i][k] - V[j][k]; sum += d * d; } D[i][j] = sum; D[i][B.length] += Math.pow(1 / D[i][j], exp); } } By doing so I encounter numerical differences between the 'banal' computation and the second one that leads to different numerical value in U (not in the first iterates but soon enough). I guess that the problem is that exponentiate very small numbers to high values (the elements of U can range from 0.0 to 1.0 and exp , for m = 1.1, is 10) leads to ver y small values, whereas by dividing the numerator and the denominator and THEN exponentiating the result seems to be better numerically. The problem is it involves much more operations. Am I doing something wrong? Is there a possible solution to get both the code optimized and numerically stable? Any suggestion or criticism will be appreciated.

    Read the article

  • snapping an angle to the closest cardinal direction

    - by Josh E
    I'm developing a 2D sprite-based game, and I'm finding that I'm having trouble with making the sprites rotate correctly. In a nutshell, I've got spritesheets for each of 5 directions (the other 3 come from just flipping the sprite horizontally), and I need to clamp the velocity/rotation of the sprite to one of those directions. My sprite class has a pre-computed list of radians corresponding to the cardinal directions like this: protected readonly List<float> CardinalDirections = new List<float> { MathHelper.PiOver4, MathHelper.PiOver2, MathHelper.PiOver2 + MathHelper.PiOver4, MathHelper.Pi, -MathHelper.PiOver4, -MathHelper.PiOver2, -MathHelper.PiOver2 + -MathHelper.PiOver4, -MathHelper.Pi, }; Here's the positional update code: if (velocity == Vector2.Zero) return; var rot = ((float)Math.Atan2(velocity.Y, velocity.X)); TurretRotation = SnapPositionToGrid(rot); var snappedX = (float)Math.Cos(TurretRotation); var snappedY = (float)Math.Sin(TurretRotation); var rotVector = new Vector2(snappedX, snappedY); velocity *= rotVector; //...snip private float SnapPositionToGrid(float rotationToSnap) { if (rotationToSnap == 0) return 0.0f; var targetRotation = CardinalDirections.First(x => (x - rotationToSnap >= -0.01 && x - rotationToSnap <= 0.01)); return (float)Math.Round(targetRotation, 3); } What am I doing wrong here? I know that the SnapPositionToGrid method is far from what it needs to be - the .First(..) call is on purpose so that it throws on no match, but I have no idea how I would go about accomplishing this, and unfortunately, Google hasn't helped too much either. Am I thinking about this the wrong way, or is the answer staring at me in the face?

    Read the article

  • Mouse Speed in GLUT and OpenGL?

    - by CroCo
    I would like to simulate a point that moves in 2D. The input should be the speed of the mouse, so the new position will be computed as following new_position = old_position + delta_time*mouse_velocity As far as I know in GLUT there is no function to acquire the current speed of the mouse between each frame. What I've done so far to compute the delta_time as following void Display() { glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT); glColor3f(1.0f, 0.0f, 0.0f); static int delta_t, current_t, previous_t; current_t = glutGet(GLUT_ELAPSED_TIME); delta_t = current_t - previous_t; std::cout << delta_t << std::endl; previous_t = current_t; glutSwapBuffers(); } Where should I start from here? (Note: I have to get the speed of the mouse because I'm modeling a system) Edit: Based on the above code, delta_time fluctuates so much 34 19 2 20 1 20 0 16 1 1 10 21 0 13 1 19 34 0 13 0 6 1 14 Why does this happen?

    Read the article

  • Creating movement path displays in a top-down 2d RTS

    - by nihohit
    My game is a top-down 2d RTS coded in C# using SFML's libraries. I want that during unit selection, a unit will display it's movement path on the map. Currently, after the path is computed as a list of directions ({left, up,down, down, down, left}, as an example), it's sent to the graphical component to create it's UI equivalent, and here I'm having some problems. current, these I've checked three ways to do it: compute the size of the image (in the example above it'll be a 3*2 rectangle) and create an invisible rectangle, and then go over the directions list and mark each spot with a visible point, so as to get a continous line. This system is slightly problematic because of the amount of large images that I need to save, but mostly because I have a lot of fine detail onscreen, and a continous line obstructs the view. again, compute the size of the image, but now create several (let's say 4) invisible images of that size, and then instead of a single continous line I'll switch between the four images, in each will appear only a fourth of the spots, in a way which creates a path animation. This is nicer on the eye, but here the memory demands, and the amount of time needed to compute each such image-loop is significant. Just create a list of single markers, each on a different spot on the path. This is very quick & easy on memory, but too sparse. Is there a simple or resource-light system to create path-animations?

    Read the article

  • Context switches much slower in new linux kernels

    - by Michael Goldshteyn
    We are looking to upgrade the OS on our servers from Ubuntu 10.04 LTS to Ubuntu 12.04 LTS. Unfortunately, it seems that the latency to run a thread that has become runnable has significantly increased from the 2.6 kernel to the 3.2 kernel. In fact the latency numbers we are getting are hard to believe. Let me be more specific about the test. We have a program that has two threads. The first thread gets the current time (in ticks using RDTSC) and then signals a condition variable once a second. The second thread waits on the condition variable and wakes up when it is signaled. It then gets the current time (in ticks using RDTSC). The difference between the time in the second thread and the time in the first thread is computed and displayed on the console. After this the second thread waits on the condition variable once more. So, we get a thread to thread signaling latency measurement once a second as a result. In linux 2.6.32, this latency is somewhere on the order of 2.8-3.5 us, which is reasonable. In linux 3.2.0, this latency is somewhere on the order of 40-100 us. I have excluded any differences in hardware between the two host hosts. They run on identical hardware (dual socket X5687 {Westmere-EP} processors running at 3.6 GHz with hyperthreading, speedstep and all C states turned off). We are changing the affinity to run both threads on physical cores of the same socket (i.e., the first thread is run on Core 0 and the second thread is run on Core 1), so there is no bouncing of threads on cores or bouncing/communication between sockets. The only difference between the two hosts is that one is running Ubuntu 10.04 LTS with kernel 2.6.32-28 (the fast context switch box) and the other is running the latest Ubuntu 12.04 LTS with kernel 3.2.0-23 (the slow context switch box). Have there been any changes in the kernel that could account for this ridiculous slow down in how long it takes for a thread to be scheduled to run?

    Read the article

  • Simplifying data search using .NET

    - by Peter
    An example on the asp.net site has an example of using Linq to create a search feature on a Music album site using MVC. The code looks like this - public ActionResult Index(string movieGenre, string searchString) { var GenreLst = new List<string>(); var GenreQry = from d in db.Movies orderby d.Genre select d.Genre; GenreLst.AddRange(GenreQry.Distinct()); ViewBag.movieGenre = new SelectList(GenreLst); var movies = from m in db.Movies select m; if (!String.IsNullOrEmpty(searchString)) { movies = movies.Where(s => s.Title.Contains(searchString)); } if (!string.IsNullOrEmpty(movieGenre)) { movies = movies.Where(x => x.Genre == movieGenre); } return View(movies); } I have seen similar examples in other tutorials and I have tried them in a real-world business app that I develop/maintain. In practice this pattern doesn't seem to scale well because as the search criteria expands I keep adding more and more conditions which looks and feels unpleasant/repetitive. How can I refactor this pattern? One idea I have is to create a column in every table that is "searchable" which could be a computed column that concatenates all the data from the different columns (SQL Server 2008). So instead of having movie genre and title it would be something like. if (!String.IsNullOrEmpty(searchString)) { movies = movies.Where(s => s.SearchColumn.Contains(searchString)); } What are the performance/design/architecture implications of doing this? I have also tried using procedures that use dynamic queries but then I have just moved the ugliness to the database. E.g. CREATE PROCEDURE [dbo].[search_music] @title as varchar(50), @genre as varchar(50) AS -- set the variables to null if they are empty IF @title = '' SET @title = null IF @genre = '' SET @genre = null SELECT m.* FROM view_Music as m WHERE (title = @title OR @title IS NULL) AND (genre LIKE '%' + @genre + '%' OR @genre IS NULL) ORDER BY Id desc OPTION (RECOMPILE) Any suggestions? Tips?

    Read the article

  • Is it possible/likely to be paid fairly without a college degree? [closed]

    - by user20134
    Some back story, and then my question: I took a "break" from getting a university education last year to work full time as back end developer on a GIS application at $10.50 an hour. Later that year I was hired on by a fairly prestigious organization on their GIS application for a meager salary + rockin' benefits (not that I need them). I agreed to work on this project through Summer 2012. I don't feel like I'm being fairly compensated for my time. Other team members make between 3-5 times as much as I do, and their work isn't 3-5 times as good as mine, nor do they have 3-5 times as much output. I don't think this is a rectifiable situation within this institution. They've got a set of personnel charts and the way it gets computed, I make less money than any of the janitors (who are very good, and very nice people to boot, and I'm glad they get paid so well. I wish everyone got livable wages). I'm pretty bright, but school's a drag. I don't want mega bucks, I just want $40k/yr (localized to the southeast united states) so I can save enough money to travel, or maybe "finish [my] education". My question is this: Are people without degrees ever compensated commensurate with other people who have degrees? As a someone who never "finished their education", how badly do you think this as hurt you? How do you navigate the job seeking and hiring process? As someone who hires programmers, do you pay more for diplomas? Is that an institutional necessity, or based on your own value judgement?

    Read the article

  • Nginx and client certificates from hierarchical OpenSSL-based certification authorities

    - by Fmy Oen
    I'm trying to set up root certification authority, subordinate certification authority and to generate the client certificates signed by any of this CA that nginx 0.7.67 on Debian Squeeze will accept. My problem is that root CA signed client certificate works fine while subordinate CA signed one results in "400 Bad Request. The SSL certificate error". Step 1: nginx virtual host configuration: server { server_name test.local; access_log /var/log/nginx/test.access.log; listen 443 default ssl; keepalive_timeout 70; ssl_protocols SSLv3 TLSv1; ssl_ciphers AES128-SHA:AES256-SHA:RC4-SHA:DES-CBC3-SHA:RC4-MD5; ssl_certificate /etc/nginx/ssl/server.crt; ssl_certificate_key /etc/nginx/ssl/server.key; ssl_client_certificate /etc/nginx/ssl/client.pem; ssl_verify_client on; ssl_session_cache shared:SSL:10m; ssl_session_timeout 5m; location / { proxy_pass http://testsite.local/; } } Step 2: PKI infrastructure organization for both root and subordinate CA (based on this article): # mkdir ~/pki && cd ~/pki # mkdir rootCA subCA # cp -v /etc/ssl/openssl.cnf rootCA/ # cd rootCA/ # mkdir certs private crl newcerts; touch serial; echo 01 > serial; touch index.txt; touch crlnumber; echo 01 > crlnumber # cp -Rvp * ../subCA/ Almost no changes was made to rootCA/openssl.cnf: [ CA_default ] dir = . # Where everything is kept ... certificate = $dir/certs/rootca.crt # The CA certificate ... private_key = $dir/private/rootca.key # The private key and to subCA/openssl.cnf: [ CA_default ] dir = . # Where everything is kept ... certificate = $dir/certs/subca.crt # The CA certificate ... private_key = $dir/private/subca.key # The private key Step 3: Self-signed root CA certificate generation: # openssl genrsa -out ./private/rootca.key -des3 2048 # openssl req -x509 -new -key ./private/rootca.key -out certs/rootca.crt -config openssl.cnf Enter pass phrase for ./private/rootca.key: You are about to be asked to enter information that will be incorporated into your certificate request. What you are about to enter is what is called a Distinguished Name or a DN. There are quite a few fields but you can leave some blank For some fields there will be a default value, If you enter '.', the field will be left blank. ----- Country Name (2 letter code) [AU]: State or Province Name (full name) [Some-State]: Locality Name (eg, city) []: Organization Name (eg, company) [Internet Widgits Pty Ltd]: Organizational Unit Name (eg, section) []: Common Name (eg, YOUR name) []:rootca Email Address []: Step 4: Subordinate CA certificate generation: # cd ../subCA # openssl genrsa -out ./private/subca.key -des3 2048 # openssl req -new -key ./private/subca.key -out subca.csr -config openssl.cnf Enter pass phrase for ./private/subca.key: You are about to be asked to enter information that will be incorporated into your certificate request. What you are about to enter is what is called a Distinguished Name or a DN. There are quite a few fields but you can leave some blank For some fields there will be a default value, If you enter '.', the field will be left blank. ----- Country Name (2 letter code) [AU]: State or Province Name (full name) [Some-State]: Locality Name (eg, city) []: Organization Name (eg, company) [Internet Widgits Pty Ltd]: Organizational Unit Name (eg, section) []: Common Name (eg, YOUR name) []:subca Email Address []: Please enter the following 'extra' attributes to be sent with your certificate request A challenge password []: An optional company name []: Step 5: Subordinate CA certificate signing by root CA certificate: # cd ../rootCA/ # openssl ca -in ../subCA/subca.csr -extensions v3_ca -config openssl.cnf Using configuration from openssl.cnf Enter pass phrase for ./private/rootca.key: Check that the request matches the signature Signature ok Certificate Details: Serial Number: 1 (0x1) Validity Not Before: Feb 4 10:49:43 2013 GMT Not After : Feb 4 10:49:43 2014 GMT Subject: countryName = AU stateOrProvinceName = Some-State organizationName = Internet Widgits Pty Ltd commonName = subca X509v3 extensions: X509v3 Subject Key Identifier: C9:E2:AC:31:53:81:86:3F:CD:F8:3D:47:10:FC:E5:8E:C2:DA:A9:20 X509v3 Authority Key Identifier: keyid:E9:50:E6:BF:57:03:EA:6E:8F:21:23:86:BB:44:3D:9F:8F:4A:8B:F2 DirName:/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca serial:9F:FB:56:66:8D:D3:8F:11 X509v3 Basic Constraints: CA:TRUE Certificate is to be certified until Feb 4 10:49:43 2014 GMT (365 days) Sign the certificate? [y/n]:y 1 out of 1 certificate requests certified, commit? [y/n]y ... # cd ../subCA/ # cp -v ../rootCA/newcerts/01.pem certs/subca.crt Step 6: Server certificate generation and signing by root CA (for nginx virtual host): # cd ../rootCA # openssl genrsa -out ./private/server.key -des3 2048 # openssl req -new -key ./private/server.key -out server.csr -config openssl.cnf Enter pass phrase for ./private/server.key: You are about to be asked to enter information that will be incorporated into your certificate request. What you are about to enter is what is called a Distinguished Name or a DN. There are quite a few fields but you can leave some blank For some fields there will be a default value, If you enter '.', the field will be left blank. ----- Country Name (2 letter code) [AU]: State or Province Name (full name) [Some-State]: Locality Name (eg, city) []: Organization Name (eg, company) [Internet Widgits Pty Ltd]: Organizational Unit Name (eg, section) []: Common Name (eg, YOUR name) []:test.local Email Address []: Please enter the following 'extra' attributes to be sent with your certificate request A challenge password []: An optional company name []: # openssl ca -in server.csr -out certs/server.crt -config openssl.cnf Step 7: Client #1 certificate generation and signing by root CA: # openssl genrsa -out ./private/client1.key -des3 2048 # openssl req -new -key ./private/client1.key -out client1.csr -config openssl.cnf Enter pass phrase for ./private/client1.key: You are about to be asked to enter information that will be incorporated into your certificate request. What you are about to enter is what is called a Distinguished Name or a DN. There are quite a few fields but you can leave some blank For some fields there will be a default value, If you enter '.', the field will be left blank. ----- Country Name (2 letter code) [AU]: State or Province Name (full name) [Some-State]: Locality Name (eg, city) []: Organization Name (eg, company) [Internet Widgits Pty Ltd]: Organizational Unit Name (eg, section) []: Common Name (eg, YOUR name) []:Client #1 Email Address []: Please enter the following 'extra' attributes to be sent with your certificate request A challenge password []: An optional company name []: # openssl ca -in client1.csr -out certs/client1.crt -config openssl.cnf Step 8: Client #1 certificate converting to PKCS12 format: # openssl pkcs12 -export -out certs/client1.p12 -inkey private/client1.key -in certs/client1.crt -certfile certs/rootca.crt Step 9: Client #2 certificate generation and signing by subordinate CA: # cd ../subCA/ # openssl genrsa -out ./private/client2.key -des3 2048 # openssl req -new -key ./private/client2.key -out client2.csr -config openssl.cnf Enter pass phrase for ./private/client2.key: You are about to be asked to enter information that will be incorporated into your certificate request. What you are about to enter is what is called a Distinguished Name or a DN. There are quite a few fields but you can leave some blank For some fields there will be a default value, If you enter '.', the field will be left blank. ----- Country Name (2 letter code) [AU]: State or Province Name (full name) [Some-State]: Locality Name (eg, city) []: Organization Name (eg, company) [Internet Widgits Pty Ltd]: Organizational Unit Name (eg, section) []: Common Name (eg, YOUR name) []:Client #2 Email Address []: Please enter the following 'extra' attributes to be sent with your certificate request A challenge password []: An optional company name []: # openssl ca -in client2.csr -out certs/client2.crt -config openssl.cnf Step 10: Client #2 certificate converting to PKCS12 format: # openssl pkcs12 -export -out certs/client2.p12 -inkey private/client2.key -in certs/client2.crt -certfile certs/subca.crt Step 11: Passing server certificate and private key to nginx (performed with OS superuser privileges): # cd ../rootCA/ # cp -v certs/server.crt /etc/nginx/ssl/ # cp -v private/server.key /etc/nginx/ssl/ Step 12: Passing root and subordinate CA certificates to nginx (performed with OS superuser privileges): # cat certs/rootca.crt > /etc/nginx/ssl/client.pem # cat ../subCA/certs/subca.crt >> /etc/nginx/ssl/client.pem client.pem file look like this: # cat /etc/nginx/ssl/client.pem -----BEGIN CERTIFICATE----- MIID6TCCAtGgAwIBAgIJAJ/7VmaN048RMA0GCSqGSIb3DQEBBQUAMFYxCzAJBgNV BAYTAkFVMRMwEQYDVQQIEwpTb21lLVN0YXRlMSEwHwYDVQQKExhJbnRlcm5ldCBX aWRnaXRzIFB0eSBMdGQxDzANBgNVBAMTBnJvb3RjYTAeFw0xMzAyMDQxMDM1NTda ... -----END CERTIFICATE----- Certificate: Data: Version: 3 (0x2) Serial Number: 1 (0x1) ... -----BEGIN CERTIFICATE----- MIID4DCCAsigAwIBAgIBATANBgkqhkiG9w0BAQUFADBWMQswCQYDVQQGEwJBVTET MBEGA1UECBMKU29tZS1TdGF0ZTEhMB8GA1UEChMYSW50ZXJuZXQgV2lkZ2l0cyBQ dHkgTHRkMQ8wDQYDVQQDEwZyb290Y2EwHhcNMTMwMjA0MTA0OTQzWhcNMTQwMjA0 ... -----END CERTIFICATE----- It looks like everything is working fine: # service nginx reload # Reloading nginx configuration: Enter PEM pass phrase: # nginx. # Step 13: Installing *.p12 certificates in browser (Firefox in my case) gives the problem I've mentioned above. Client #1 = 200 OK, Client #2 = 400 Bad request/The SSL certificate error. Any ideas what should I do? Update 1: Results of SSL connection test attempts: # openssl s_client -connect test.local:443 -CAfile ~/pki/rootCA/certs/rootca.crt -cert ~/pki/rootCA/certs/client1.crt -key ~/pki/rootCA/private/client1.key -showcerts Enter pass phrase for tmp/testcert/client1.key: CONNECTED(00000003) depth=1 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = rootca verify return:1 depth=0 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = test.local verify return:1 --- Certificate chain 0 s:/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=test.local i:/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca -----BEGIN CERTIFICATE----- MIIDpjCCAo6gAwIBAgIBAjANBgkqhkiG9w0BAQUFADBWMQswCQYDVQQGEwJBVTET MBEGA1UECBMKU29tZS1TdGF0ZTEhMB8GA1UEChMYSW50ZXJuZXQgV2lkZ2l0cyBQ dHkgTHRkMQ8wDQYDVQQDEwZyb290Y2EwHhcNMTMwMjA0MTEwNjAzWhcNMTQwMjA0 ... -----END CERTIFICATE----- 1 s:/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca i:/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca -----BEGIN CERTIFICATE----- MIID6TCCAtGgAwIBAgIJAJ/7VmaN048RMA0GCSqGSIb3DQEBBQUAMFYxCzAJBgNV BAYTAkFVMRMwEQYDVQQIEwpTb21lLVN0YXRlMSEwHwYDVQQKExhJbnRlcm5ldCBX aWRnaXRzIFB0eSBMdGQxDzANBgNVBAMTBnJvb3RjYTAeFw0xMzAyMDQxMDM1NTda ... -----END CERTIFICATE----- --- Server certificate subject=/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=test.local issuer=/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca --- Acceptable client certificate CA names /C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca /C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=subca --- SSL handshake has read 3395 bytes and written 2779 bytes --- New, TLSv1/SSLv3, Cipher is AES256-SHA Server public key is 2048 bit Secure Renegotiation IS supported Compression: zlib compression Expansion: zlib compression SSL-Session: Protocol : TLSv1 Cipher : AES256-SHA Session-ID: 15BFC2029691262542FAE95A48078305E76EEE7D586400F8C4F7C516B0F9D967 Session-ID-ctx: Master-Key: 23246CF166E8F3900793F0A2561879E5DB07291F32E99591BA1CF53E6229491FEAE6858BFC9AACAF271D9C3706F139C7 Key-Arg : None PSK identity: None PSK identity hint: None SRP username: None TLS session ticket: 0000 - c2 5e 1d d2 b5 6d 40 23-b2 40 89 e4 35 75 70 07 .^...m@#[email protected]. 0010 - 1b bb 2b e6 e0 b5 ab 10-10 bf 46 6e aa 67 7f 58 ..+.......Fn.g.X 0020 - cf 0e 65 a4 67 5a 15 ba-aa 93 4e dd 3d 6e 73 4c ..e.gZ....N.=nsL 0030 - c5 56 f6 06 24 0f 48 e6-38 36 de f1 b5 31 c5 86 .V..$.H.86...1.. ... 0440 - 4c 53 39 e3 92 84 d2 d0-e5 e2 f5 8a 6a a8 86 b1 LS9.........j... Compression: 1 (zlib compression) Start Time: 1359989684 Timeout : 300 (sec) Verify return code: 0 (ok) --- Everything seems fine with Client #2 and root CA certificate but request returns 400 Bad Request error: # openssl s_client -connect test.local:443 -CAfile ~/pki/rootCA/certs/rootca.crt -cert ~/pki/subCA/certs/client2.crt -key ~/pki/subCA/private/client2.key -showcerts Enter pass phrase for tmp/testcert/client2.key: CONNECTED(00000003) depth=1 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = rootca verify return:1 depth=0 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = test.local verify return:1 ... Compression: 1 (zlib compression) Start Time: 1359989989 Timeout : 300 (sec) Verify return code: 0 (ok) --- GET / HTTP/1.0 HTTP/1.1 400 Bad Request Server: nginx/0.7.67 Date: Mon, 04 Feb 2013 15:00:43 GMT Content-Type: text/html Content-Length: 231 Connection: close <html> <head><title>400 The SSL certificate error</title></head> <body bgcolor="white"> <center><h1>400 Bad Request</h1></center> <center>The SSL certificate error</center> <hr><center>nginx/0.7.67</center> </body> </html> closed Verification fails with Client #2 certificate and subordinate CA certificate: # openssl s_client -connect test.local:443 -CAfile ~/pki/subCA/certs/subca.crt -cert ~/pki/subCA/certs/client2.crt -key ~/pki/subCA/private/client2.key -showcerts Enter pass phrase for tmp/testcert/client2.key: CONNECTED(00000003) depth=1 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = rootca verify error:num=19:self signed certificate in certificate chain verify return:0 ... Compression: 1 (zlib compression) Start Time: 1359990354 Timeout : 300 (sec) Verify return code: 19 (self signed certificate in certificate chain) --- GET / HTTP/1.0 HTTP/1.1 400 Bad Request ... Still getting 400 Bad Request error with concatenated CA certificates and Client #2 (but still everything ok with Client #1): # cat certs/rootca.crt ../subCA/certs/subca.crt > certs/concatenatedca.crt # openssl s_client -connect test.local:443 -CAfile ~/pki/rootCA/certs/concatenatedca.crt -cert ~/pki/subCA/certs/client2.crt -key ~/pki/subCA/private/client2.key -showcerts Enter pass phrase for tmp/testcert/client2.key: CONNECTED(00000003) depth=1 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = rootca verify return:1 depth=0 C = AU, ST = Some-State, O = Internet Widgits Pty Ltd, CN = test.local verify return:1 --- ... Compression: 1 (zlib compression) Start Time: 1359990772 Timeout : 300 (sec) Verify return code: 0 (ok) --- GET / HTTP/1.0 HTTP/1.1 400 Bad Request ... Update 2: I've managed to recompile nginx with enabled debug. Here is the part of successfull conection by Client #1 track: 2013/02/05 14:08:23 [debug] 38701#0: *119 accept: <MY IP ADDRESS> fd:3 2013/02/05 14:08:23 [debug] 38701#0: *119 event timer add: 3: 60000:2856497512 2013/02/05 14:08:23 [debug] 38701#0: *119 kevent set event: 3: ft:-1 fl:0025 2013/02/05 14:08:23 [debug] 38701#0: *119 malloc: 28805200:660 2013/02/05 14:08:23 [debug] 38701#0: *119 malloc: 28834400:1024 2013/02/05 14:08:23 [debug] 38701#0: *119 posix_memalign: 28860000:4096 @16 2013/02/05 14:08:23 [debug] 38701#0: *119 http check ssl handshake 2013/02/05 14:08:23 [debug] 38701#0: *119 https ssl handshake: 0x16 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL server name: "test.local" 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_do_handshake: -1 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_get_error: 2 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL handshake handler: 0 2013/02/05 14:08:23 [debug] 38701#0: *119 verify:1, error:0, depth:1, subject:"/C=AU /ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca",issuer: "/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca" 2013/02/05 14:08:23 [debug] 38701#0: *119 verify:1, error:0, depth:0, subject:"/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=Client #1",issuer: "/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca" 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_do_handshake: 1 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL: TLSv1, cipher: "AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1" 2013/02/05 14:08:23 [debug] 38701#0: *119 http process request line 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_read: -1 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_get_error: 2 2013/02/05 14:08:23 [debug] 38701#0: *119 http process request line 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_read: 1 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_read: 524 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_read: -1 2013/02/05 14:08:23 [debug] 38701#0: *119 SSL_get_error: 2 2013/02/05 14:08:23 [debug] 38701#0: *119 http request line: "GET / HTTP/1.1" And here is the part of unsuccessfull conection by Client #2 track: 2013/02/05 13:51:34 [debug] 38701#0: *112 accept: <MY_IP_ADDRESS> fd:3 2013/02/05 13:51:34 [debug] 38701#0: *112 event timer add: 3: 60000:2855488975 2013/02/05 13:51:34 [debug] 38701#0: *112 kevent set event: 3: ft:-1 fl:0025 2013/02/05 13:51:34 [debug] 38701#0: *112 malloc: 28805200:660 2013/02/05 13:51:34 [debug] 38701#0: *112 malloc: 28834400:1024 2013/02/05 13:51:34 [debug] 38701#0: *112 posix_memalign: 28860000:4096 @16 2013/02/05 13:51:34 [debug] 38701#0: *112 http check ssl handshake 2013/02/05 13:51:34 [debug] 38701#0: *112 https ssl handshake: 0x16 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL server name: "test.local" 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_do_handshake: -1 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_get_error: 2 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL handshake handler: 0 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_do_handshake: -1 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_get_error: 2 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL handshake handler: 0 2013/02/05 13:51:34 [debug] 38701#0: *112 verify:0, error:20, depth:1, subject:"/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=subca",issuer: "/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca" 2013/02/05 13:51:34 [debug] 38701#0: *112 verify:0, error:27, depth:1, subject:"/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=subca",issuer: "/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=rootca" 2013/02/05 13:51:34 [debug] 38701#0: *112 verify:1, error:27, depth:0, subject:"/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=Client #2",issuer: "/C=AU/ST=Some-State/O=Internet Widgits Pty Ltd/CN=subca" 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_do_handshake: 1 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL: TLSv1, cipher: "AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1" 2013/02/05 13:51:34 [debug] 38701#0: *112 http process request line 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_read: 1 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_read: 524 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_read: -1 2013/02/05 13:51:34 [debug] 38701#0: *112 SSL_get_error: 2 2013/02/05 13:51:34 [debug] 38701#0: *112 http request line: "GET / HTTP/1.1" So I'm getting OpenSSL error #20 and then #27. According to verify documentation: 20 X509_V_ERR_UNABLE_TO_GET_ISSUER_CERT_LOCALLY: unable to get local issuer certificate the issuer certificate could not be found: this occurs if the issuer certificate of an untrusted certificate cannot be found. 27 X509_V_ERR_CERT_UNTRUSTED: certificate not trusted the root CA is not marked as trusted for the specified purpose.

    Read the article

  • Adding Unobtrusive Validation To MVCContrib Fluent Html

    - by srkirkland
    ASP.NET MVC 3 includes a new unobtrusive validation strategy that utilizes HTML5 data-* attributes to decorate form elements.  Using a combination of jQuery validation and an unobtrusive validation adapter script that comes with MVC 3, those attributes are then turned into client side validation rules. A Quick Introduction to Unobtrusive Validation To quickly show how this works in practice, assume you have the following Order.cs class (think Northwind) [If you are familiar with unobtrusive validation in MVC 3 you can skip to the next section]: public class Order : DomainObject { [DataType(DataType.Date)] public virtual DateTime OrderDate { get; set; }   [Required] [StringLength(12)] public virtual string ShipAddress { get; set; }   [Required] public virtual Customer OrderedBy { get; set; } } .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; } Note the System.ComponentModel.DataAnnotations attributes, which provide the validation and metadata information used by ASP.NET MVC 3 to determine how to render out these properties.  Now let’s assume we have a form which can edit this Order class, specifically let’s look at the ShipAddress property: @Html.LabelFor(x => x.Order.ShipAddress) @Html.EditorFor(x => x.Order.ShipAddress) @Html.ValidationMessageFor(x => x.Order.ShipAddress) .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; } Now the Html.EditorFor() method is smart enough to look at the ShipAddress attributes and write out the necessary unobtrusive validation html attributes.  Note we could have used Html.TextBoxFor() or even Html.TextBox() and still retained the same results. If we view source on the input box generated by the Html.EditorFor() call, we get the following: <input type="text" value="Rua do Paço, 67" name="Order.ShipAddress" id="Order_ShipAddress" data-val-required="The ShipAddress field is required." data-val-length-max="12" data-val-length="The field ShipAddress must be a string with a maximum length of 12." data-val="true" class="text-box single-line input-validation-error"> .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; } As you can see, we have data-val-* attributes for both required and length, along with the proper error messages and additional data as necessary (in this case, we have the length-max=”12”). And of course, if we try to submit the form with an invalid value, we get an error on the client: Working with MvcContrib’s Fluent Html The MvcContrib project offers a fluent interface for creating Html elements which I find very expressive and useful, especially when it comes to creating select lists.  Let’s look at a few quick examples: @this.TextBox(x => x.FirstName).Class("required").Label("First Name:") @this.MultiSelect(x => x.UserId).Options(ViewModel.Users) @this.CheckBox("enabled").LabelAfter("Enabled").Title("Click to enable.").Styles(vertical_align => "middle")   @(this.Select("Order.OrderedBy").Options(Model.Customers, x => x.Id, x => x.CompanyName) .Selected(Model.Order.OrderedBy != null ? Model.Order.OrderedBy.Id : "") .FirstOption(null, "--Select A Company--") .HideFirstOptionWhen(Model.Order.OrderedBy != null) .Label("Ordered By:")) .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; } These fluent html helpers create the normal html you would expect, and I think they make life a lot easier and more readable when dealing with complex markup or select list data models (look ma: no anonymous objects for creating class names!). Of course, the problem we have now is that MvcContrib’s fluent html helpers don’t know about ASP.NET MVC 3’s unobtrusive validation attributes and thus don’t take part in client validation on your page.  This is not ideal, so I wrote a quick helper method to extend fluent html with the knowledge of what unobtrusive validation attributes to include when they are rendered. Extending MvcContrib’s Fluent Html Before posting the code, there are just a few things you need to know.  The first is that all Fluent Html elements implement the IElement interface (MvcContrib.FluentHtml.Elements.IElement), and the second is that the base System.Web.Mvc.HtmlHelper has been extended with a method called GetUnobtrusiveValidationAttributes which we can use to determine the necessary attributes to include.  With this knowledge we can make quick work of extending fluent html: public static class FluentHtmlExtensions { public static T IncludeUnobtrusiveValidationAttributes<T>(this T element, HtmlHelper htmlHelper) where T : MvcContrib.FluentHtml.Elements.IElement { IDictionary<string, object> validationAttributes = htmlHelper .GetUnobtrusiveValidationAttributes(element.GetAttr("name"));   foreach (var validationAttribute in validationAttributes) { element.SetAttr(validationAttribute.Key, validationAttribute.Value); }   return element; } } .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; } The code is pretty straight forward – basically we use a passed HtmlHelper to get a list of validation attributes for the current element and then add each of the returned attributes to the element to be rendered. The Extension In Action Now let’s get back to the earlier ShipAddress example and see what we’ve accomplished.  First we will use a fluent html helper to render out the ship address text input (this is the ‘before’ case): @this.TextBox("Order.ShipAddress").Label("Ship Address:").Class("class-name") .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; } And the resulting HTML: <label id="Order_ShipAddress_Label" for="Order_ShipAddress">Ship Address:</label> <input type="text" value="Rua do Paço, 67" name="Order.ShipAddress" id="Order_ShipAddress" class="class-name"> Now let’s do the same thing except here we’ll use the newly written extension method: @this.TextBox("Order.ShipAddress").Label("Ship Address:") .Class("class-name").IncludeUnobtrusiveValidationAttributes(Html) .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; } And the resulting HTML: <label id="Order_ShipAddress_Label" for="Order_ShipAddress">Ship Address:</label> <input type="text" value="Rua do Paço, 67" name="Order.ShipAddress" id="Order_ShipAddress" data-val-required="The ShipAddress field is required." data-val-length-max="12" data-val-length="The field ShipAddress must be a string with a maximum length of 12." data-val="true" class="class-name"> .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; } Excellent!  Now we can continue to use unobtrusive validation and have the flexibility to use ASP.NET MVC’s Html helpers or MvcContrib’s fluent html helpers interchangeably, and every element will participate in client side validation. Wrap Up Overall I’m happy with this solution, although in the best case scenario MvcContrib would know about unobtrusive validation attributes and include them automatically (of course if it is enabled in the web.config file).  I know that MvcContrib allows you to author global behaviors, but that requires changing the base class of your views, which I am not willing to do. Enjoy!

    Read the article

  • Adding Client Validation To DataAnnotations DataType Attribute

    - by srkirkland
    The System.ComponentModel.DataAnnotations namespace contains a validation attribute called DataTypeAttribute, which takes an enum specifying what data type the given property conforms to.  Here are a few quick examples: public class DataTypeEntity { [DataType(DataType.Date)] public DateTime DateTime { get; set; }   [DataType(DataType.EmailAddress)] public string EmailAddress { get; set; } } .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 attribute comes in handy when using ASP.NET MVC, because the type you specify will determine what “template” MVC uses.  Thus, for the DateTime property if you create a partial in Views/[loc]/EditorTemplates/Date.ascx (or cshtml for razor), that view will be used to render the property when using any of the Html.EditorFor() methods. One thing that the DataType() validation attribute does not do is any actual validation.  To see this, let’s take a look at the EmailAddress property above.  It turns out that regardless of the value you provide, the entity will be considered valid: //valid new DataTypeEntity {EmailAddress = "Foo"}; .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; } Hmmm.  Since DataType() doesn’t validate, that leaves us with two options: (1) Create our own attributes for each datatype to validate, like [Date], or (2) add validation into the DataType attribute directly.  In this post, I will show you how to hookup client-side validation to the existing DataType() attribute for a desired type.  From there adding server-side validation would be a breeze and even writing a custom validation attribute would be simple (more on that in future posts). Validation All The Way Down Our goal will be to leave our DataTypeEntity class (from above) untouched, requiring no reference to System.Web.Mvc.  Then we will make an ASP.NET MVC project that allows us to create a new DataTypeEntity and hookup automatic client-side date validation using the suggested “out-of-the-box” jquery.validate bits that are included with ASP.NET MVC 3.  For simplicity I’m going to focus on the only DateTime field, but the concept is generally the same for any other DataType. Building a DataTypeAttribute Adapter To start we will need to build a new validation adapter that we can register using ASP.NET MVC’s DataAnnotationsModelValidatorProvider.RegisterAdapter() method.  This method takes two Type parameters; The first is the attribute we are looking to validate with and the second is an adapter that should subclass System.Web.Mvc.ModelValidator. Since we are extending DataAnnotations we can use the subclass of ModelValidator called DataAnnotationsModelValidator<>.  This takes a generic argument of type DataAnnotations.ValidationAttribute, which lucky for us means the DataTypeAttribute will fit in nicely. So starting from there and implementing the required constructor, we get: public class DataTypeAttributeAdapter : DataAnnotationsModelValidator<DataTypeAttribute> { public DataTypeAttributeAdapter(ModelMetadata metadata, ControllerContext context, DataTypeAttribute attribute) : base(metadata, context, attribute) { } } .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; } Now you have a full-fledged validation adapter, although it doesn’t do anything yet.  There are two methods you can override to add functionality, IEnumerable<ModelValidationResult> Validate(object container) and IEnumerable<ModelClientValidationRule> GetClientValidationRules().  Adding logic to the server-side Validate() method is pretty straightforward, and for this post I’m going to focus on GetClientValidationRules(). Adding a Client Validation Rule Adding client validation is now incredibly easy because jquery.validate is very powerful and already comes with a ton of validators (including date and regular expressions for our email example).  Teamed with the new unobtrusive validation javascript support we can make short work of our ModelClientValidationDateRule: public class ModelClientValidationDateRule : ModelClientValidationRule { public ModelClientValidationDateRule(string errorMessage) { ErrorMessage = errorMessage; ValidationType = "date"; } } .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; } If your validation has additional parameters you can the ValidationParameters IDictionary<string,object> to include them.  There is a little bit of conventions magic going on here, but the distilled version is that we are defining a “date” validation type, which will be included as html5 data-* attributes (specifically data-val-date).  Then jquery.validate.unobtrusive takes this attribute and basically passes it along to jquery.validate, which knows how to handle date validation. Finishing our DataTypeAttribute Adapter Now that we have a model client validation rule, we can return it in the GetClientValidationRules() method of our DataTypeAttributeAdapter created above.  Basically I want to say if DataType.Date was provided, then return the date rule with a given error message (using ValidationAttribute.FormatErrorMessage()).  The entire adapter is below: public class DataTypeAttributeAdapter : DataAnnotationsModelValidator<DataTypeAttribute> { public DataTypeAttributeAdapter(ModelMetadata metadata, ControllerContext context, DataTypeAttribute attribute) : base(metadata, context, attribute) { }   public override System.Collections.Generic.IEnumerable<ModelClientValidationRule> GetClientValidationRules() { if (Attribute.DataType == DataType.Date) { return new[] { new ModelClientValidationDateRule(Attribute.FormatErrorMessage(Metadata.GetDisplayName())) }; }   return base.GetClientValidationRules(); } } .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; } Putting it all together Now that we have an adapter for the DataTypeAttribute, we just need to tell ASP.NET MVC to use it.  The easiest way to do this is to use the built in DataAnnotationsModelValidatorProvider by calling RegisterAdapter() in your global.asax startup method. DataAnnotationsModelValidatorProvider.RegisterAdapter(typeof(DataTypeAttribute), typeof(DataTypeAttributeAdapter)); .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; } Show and Tell Let’s see this in action using a clean ASP.NET MVC 3 project.  First make sure to reference the jquery, jquery.vaidate and jquery.validate.unobtrusive scripts that you will need for client validation. Next, let’s make a model class (note we are using the same built-in DataType() attribute that comes with System.ComponentModel.DataAnnotations). public class DataTypeEntity { [DataType(DataType.Date, ErrorMessage = "Please enter a valid date (ex: 2/14/2011)")] public DateTime DateTime { get; set; } } .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; } Then we make a create page with a strongly-typed DataTypeEntity model, the form section is shown below (notice we are just using EditorForModel): @using (Html.BeginForm()) { @Html.ValidationSummary(true) <fieldset> <legend>Fields</legend>   @Html.EditorForModel()   <p> <input type="submit" value="Create" /> </p> </fieldset> } .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; } The final step is to register the adapter in our global.asax file: DataAnnotationsModelValidatorProvider.RegisterAdapter(typeof(DataTypeAttribute), typeof(DataTypeAttributeAdapter)); Now we are ready to run the page: Looking at the datetime field’s html, we see that our adapter added some data-* validation attributes: <input type="text" value="1/1/0001" name="DateTime" id="DateTime" data-val-required="The DateTime field is required." data-val-date="Please enter a valid date (ex: 2/14/2011)" data-val="true" class="text-box single-line valid"> .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; } Here data-val-required was added automatically because DateTime is non-nullable, and data-val-date was added by our validation adapter.  Now if we try to add an invalid date: Our custom error message is displayed via client-side validation as soon as we tab out of the box.  If we didn’t include a custom validation message, the default DataTypeAttribute “The field {0} is invalid” would have been shown (of course we can change the default as well).  Note we did not specify server-side validation, but in this case we don’t have to because an invalid date will cause a server-side error during model binding. Conclusion I really like how easy it is to register new data annotations model validators, whether they are your own or, as in this post, supplements to existing validation attributes.  I’m still debating about whether adding the validation directly in the DataType attribute is the correct place to put it versus creating a dedicated “Date” validation attribute, but it’s nice to know either option is available and, as we’ve seen, simple to implement. I’m also working through the nascent stages of an open source project that will create validation attribute extensions to the existing data annotations providers using similar techniques as seen above (examples: Email, Url, EqualTo, Min, Max, CreditCard, etc).  Keep an eye on this blog and subscribe to my twitter feed (@srkirkland) if you are interested for announcements.

    Read the article

  • SQL Server &ndash; Undelete a Table and Restore a Single Table from Backup

    - by Mladen Prajdic
    This post is part of the monthly community event called T-SQL Tuesday started by Adam Machanic (blog|twitter) and hosted by someone else each month. This month the host is Sankar Reddy (blog|twitter) and the topic is Misconceptions in SQL Server. You can follow posts for this theme on Twitter by looking at #TSQL2sDay hashtag. Let me start by saying: This code is a crazy hack that is to never be used unless you really, really have to. Really! And I don’t think there’s a time when you would really have to use it for real. Because it’s a hack there are number of things that can go wrong so play with it knowing that. I’ve managed to totally corrupt one database. :) Oh… and for those saying: yeah yeah.. you have a single table in a file group and you’re restoring that, I say “nay nay” to you. As we all know SQL Server can’t do single table restores from backup. This is kind of a obvious thing due to different relational integrity (RI) concerns. Since we have to maintain that we have to restore all tables represented in a RI graph. For this exercise i say BAH! to those concerns. Note that this method “works” only for simple tables that don’t have LOB and off rows data. The code can be expanded to include those but I’ve tried to leave things “simple”. Note that for this to work our table needs to be relatively static data-wise. This doesn’t work for OLTP table. Products are a perfect example of static data. They don’t change much between backups, pretty much everything depends on them and their table is one of those tables that are relatively easy to accidentally delete everything from. This only works if the database is in Full or Bulk-Logged recovery mode for tables where the contents have been deleted or truncated but NOT when a table was dropped. Everything we’ll talk about has to be done before the data pages are reused for other purposes. After deletion or truncation the pages are marked as reusable so you have to act fast. The best thing probably is to put the database into single user mode ASAP while you’re performing this procedure and return it to multi user after you’re done. How do we do it? We will be using an undocumented but known DBCC commands: DBCC PAGE, an undocumented function sys.fn_dblog and a little known DATABASE RESTORE PAGE option. All tests will be on a copy of Production.Product table in AdventureWorks database called Production.Product1 because the original table has FK constraints that prevent us from truncating it for testing. -- create a duplicate table. This doesn't preserve indexes!SELECT *INTO AdventureWorks.Production.Product1FROM AdventureWorks.Production.Product   After we run this code take a full back to perform further testing.   First let’s see what the difference between DELETE and TRUNCATE is when it comes to logging. With DELETE every row deletion is logged in the transaction log. With TRUNCATE only whole data page deallocations are logged in the transaction log. Getting deleted data pages is simple. All we have to look for is row delete entry in the sys.fn_dblog output. But getting data pages that were truncated from the transaction log presents a bit of an interesting problem. I will not go into depths of IAM(Index Allocation Map) and PFS (Page Free Space) pages but suffice to say that every IAM page has intervals that tell us which data pages are allocated for a table and which aren’t. If we deep dive into the sys.fn_dblog output we can see that once you truncate a table all the pages in all the intervals are deallocated and this is shown in the PFS page transaction log entry as deallocation of pages. For every 8 pages in the same extent there is one PFS page row in the transaction log. This row holds information about all 8 pages in CSV format which means we can get to this data with some parsing. A great help for parsing this stuff is Peter Debetta’s handy function dbo.HexStrToVarBin that converts hexadecimal string into a varbinary value that can be easily converted to integer tus giving us a readable page number. The shortened (columns removed) sys.fn_dblog output for a PFS page with CSV data for 1 extent (8 data pages) looks like this: -- [Page ID] is displayed in hex format. -- To convert it to readable int we'll use dbo.HexStrToVarBin function found at -- http://sqlblog.com/blogs/peter_debetta/archive/2007/03/09/t-sql-convert-hex-string-to-varbinary.aspx -- This function must be installed in the master databaseSELECT Context, AllocUnitName, [Page ID], DescriptionFROM sys.fn_dblog(NULL, NULL)WHERE [Current LSN] = '00000031:00000a46:007d' The pages at the end marked with 0x00—> are pages that are allocated in the extent but are not part of a table. We can inspect the raw content of each data page with a DBCC PAGE command: -- we need this trace flag to redirect output to the query window.DBCC TRACEON (3604); -- WITH TABLERESULTS gives us data in table format instead of message format-- we use format option 3 because it's the easiest to read and manipulate further onDBCC PAGE (AdventureWorks, 1, 613, 3) WITH TABLERESULTS   Since the DBACC PAGE output can be quite extensive I won’t put it here. You can see an example of it in the link at the beginning of this section. Getting deleted data back When we run a delete statement every row to be deleted is marked as a ghost record. A background process periodically cleans up those rows. A huge misconception is that the data is actually removed. It’s not. Only the pointers to the rows are removed while the data itself is still on the data page. We just can’t access it with normal means. To get those pointers back we need to restore every deleted page using the RESTORE PAGE option mentioned above. This restore must be done from a full backup, followed by any differential and log backups that you may have. This is necessary to bring the pages up to the same point in time as the rest of the data.  However the restore doesn’t magically connect the restored page back to the original table. It simply replaces the current page with the one from the backup. After the restore we use the DBCC PAGE to read data directly from all data pages and insert that data into a temporary table. To finish the RESTORE PAGE  procedure we finally have to take a tail log backup (simple backup of the transaction log) and restore it back. We can now insert data from the temporary table to our original table by hand. Getting truncated data back When we run a truncate the truncated data pages aren’t touched at all. Even the pointers to rows stay unchanged. Because of this getting data back from truncated table is simple. we just have to find out which pages belonged to our table and use DBCC PAGE to read data off of them. No restore is necessary. Turns out that the problems we had with finding the data pages is alleviated by not having to do a RESTORE PAGE procedure. Stop stalling… show me The Code! This is the code for getting back deleted and truncated data back. It’s commented in all the right places so don’t be afraid to take a closer look. Make sure you have a full backup before trying this out. Also I suggest that the last step of backing and restoring the tail log is performed by hand. USE masterGOIF OBJECT_ID('dbo.HexStrToVarBin') IS NULL RAISERROR ('No dbo.HexStrToVarBin installed. Go to http://sqlblog.com/blogs/peter_debetta/archive/2007/03/09/t-sql-convert-hex-string-to-varbinary.aspx and install it in master database' , 18, 1) SET NOCOUNT ONBEGIN TRY DECLARE @dbName VARCHAR(1000), @schemaName VARCHAR(1000), @tableName VARCHAR(1000), @fullBackupName VARCHAR(1000), @undeletedTableName VARCHAR(1000), @sql VARCHAR(MAX), @tableWasTruncated bit; /* THE FIRST LINE ARE OUR INPUT PARAMETERS In this case we're trying to recover Production.Product1 table in AdventureWorks database. My full backup of AdventureWorks database is at e:\AW.bak */ SELECT @dbName = 'AdventureWorks', @schemaName = 'Production', @tableName = 'Product1', @fullBackupName = 'e:\AW.bak', @undeletedTableName = '##' + @tableName + '_Undeleted', @tableWasTruncated = 0, -- copy the structure from original table to a temp table that we'll fill with restored data @sql = 'IF OBJECT_ID(''tempdb..' + @undeletedTableName + ''') IS NOT NULL DROP TABLE ' + @undeletedTableName + ' SELECT *' + ' INTO ' + @undeletedTableName + ' FROM [' + @dbName + '].[' + @schemaName + '].[' + @tableName + ']' + ' WHERE 1 = 0' EXEC (@sql) IF OBJECT_ID('tempdb..#PagesToRestore') IS NOT NULL DROP TABLE #PagesToRestore /* FIND DATA PAGES WE NEED TO RESTORE*/ CREATE TABLE #PagesToRestore ([ID] INT IDENTITY(1,1), [FileID] INT, [PageID] INT, [SQLtoExec] VARCHAR(1000)) -- DBCC PACE statement to run later RAISERROR ('Looking for deleted pages...', 10, 1) -- use T-LOG direct read to get deleted data pages INSERT INTO #PagesToRestore([FileID], [PageID], [SQLtoExec]) EXEC('USE [' + @dbName + '];SELECT FileID, PageID, ''DBCC TRACEON (3604); DBCC PAGE ([' + @dbName + '], '' + FileID + '', '' + PageID + '', 3) WITH TABLERESULTS'' as SQLToExecFROM (SELECT DISTINCT LEFT([Page ID], 4) AS FileID, CONVERT(VARCHAR(100), ' + 'CONVERT(INT, master.dbo.HexStrToVarBin(SUBSTRING([Page ID], 6, 20)))) AS PageIDFROM sys.fn_dblog(NULL, NULL)WHERE AllocUnitName LIKE ''%' + @schemaName + '.' + @tableName + '%'' ' + 'AND Context IN (''LCX_MARK_AS_GHOST'', ''LCX_HEAP'') AND Operation in (''LOP_DELETE_ROWS''))t');SELECT *FROM #PagesToRestore -- if upper EXEC returns 0 rows it means the table was truncated so find truncated pages IF (SELECT COUNT(*) FROM #PagesToRestore) = 0 BEGIN RAISERROR ('No deleted pages found. Looking for truncated pages...', 10, 1) -- use T-LOG read to get truncated data pages INSERT INTO #PagesToRestore([FileID], [PageID], [SQLtoExec]) -- dark magic happens here -- because truncation simply deallocates pages we have to find out which pages were deallocated. -- we can find this out by looking at the PFS page row's Description column. -- for every deallocated extent the Description has a CSV of 8 pages in that extent. -- then it's just a matter of parsing it. -- we also remove the pages in the extent that weren't allocated to the table itself -- marked with '0x00-->00' EXEC ('USE [' + @dbName + '];DECLARE @truncatedPages TABLE(DeallocatedPages VARCHAR(8000), IsMultipleDeallocs BIT);INSERT INTO @truncatedPagesSELECT REPLACE(REPLACE(Description, ''Deallocated '', ''Y''), ''0x00-->00 '', ''N'') + '';'' AS DeallocatedPages, CHARINDEX('';'', Description) AS IsMultipleDeallocsFROM (SELECT DISTINCT LEFT([Page ID], 4) AS FileID, CONVERT(VARCHAR(100), CONVERT(INT, master.dbo.HexStrToVarBin(SUBSTRING([Page ID], 6, 20)))) AS PageID, DescriptionFROM sys.fn_dblog(NULL, NULL)WHERE Context IN (''LCX_PFS'') AND Description LIKE ''Deallocated%'' AND AllocUnitName LIKE ''%' + @schemaName + '.' + @tableName + '%'') t;SELECT FileID, PageID , ''DBCC TRACEON (3604); DBCC PAGE ([' + @dbName + '], '' + FileID + '', '' + PageID + '', 3) WITH TABLERESULTS'' as SQLToExecFROM (SELECT LEFT(PageAndFile, 1) as WasPageAllocatedToTable , SUBSTRING(PageAndFile, 2, CHARINDEX('':'', PageAndFile) - 2 ) as FileID , CONVERT(VARCHAR(100), CONVERT(INT, master.dbo.HexStrToVarBin(SUBSTRING(PageAndFile, CHARINDEX('':'', PageAndFile) + 1, LEN(PageAndFile))))) as PageIDFROM ( SELECT SUBSTRING(DeallocatedPages, delimPosStart, delimPosEnd - delimPosStart) as PageAndFile, IsMultipleDeallocs FROM ( SELECT *, CHARINDEX('';'', DeallocatedPages)*(N-1) + 1 AS delimPosStart, CHARINDEX('';'', DeallocatedPages)*N AS delimPosEnd FROM @truncatedPages t1 CROSS APPLY (SELECT TOP (case when t1.IsMultipleDeallocs = 1 then 8 else 1 end) ROW_NUMBER() OVER(ORDER BY number) as N FROM master..spt_values) t2 )t)t)tWHERE WasPageAllocatedToTable = ''Y''') SELECT @tableWasTruncated = 1 END DECLARE @lastID INT, @pagesCount INT SELECT @lastID = 1, @pagesCount = COUNT(*) FROM #PagesToRestore SELECT @sql = 'Number of pages to restore: ' + CONVERT(VARCHAR(10), @pagesCount) IF @pagesCount = 0 RAISERROR ('No data pages to restore.', 18, 1) ELSE RAISERROR (@sql, 10, 1) -- If the table was truncated we'll read the data directly from data pages without restoring from backup IF @tableWasTruncated = 0 BEGIN -- RESTORE DATA PAGES FROM FULL BACKUP IN BATCHES OF 200 WHILE @lastID <= @pagesCount BEGIN -- create CSV string of pages to restore SELECT @sql = STUFF((SELECT ',' + CONVERT(VARCHAR(100), FileID) + ':' + CONVERT(VARCHAR(100), PageID) FROM #PagesToRestore WHERE ID BETWEEN @lastID AND @lastID + 200 ORDER BY ID FOR XML PATH('')), 1, 1, '') SELECT @sql = 'RESTORE DATABASE [' + @dbName + '] PAGE = ''' + @sql + ''' FROM DISK = ''' + @fullBackupName + '''' RAISERROR ('Starting RESTORE command:' , 10, 1) WITH NOWAIT; RAISERROR (@sql , 10, 1) WITH NOWAIT; EXEC(@sql); RAISERROR ('Restore DONE' , 10, 1) WITH NOWAIT; SELECT @lastID = @lastID + 200 END /* If you have any differential or transaction log backups you should restore them here to bring the previously restored data pages up to date */ END DECLARE @dbccSinglePage TABLE ( [ParentObject] NVARCHAR(500), [Object] NVARCHAR(500), [Field] NVARCHAR(500), [VALUE] NVARCHAR(MAX) ) DECLARE @cols NVARCHAR(MAX), @paramDefinition NVARCHAR(500), @SQLtoExec VARCHAR(1000), @FileID VARCHAR(100), @PageID VARCHAR(100), @i INT = 1 -- Get deleted table columns from information_schema view -- Need sp_executeSQL because database name can't be passed in as variable SELECT @cols = 'select @cols = STUFF((SELECT '', ['' + COLUMN_NAME + '']''FROM ' + @dbName + '.INFORMATION_SCHEMA.COLUMNSWHERE TABLE_NAME = ''' + @tableName + ''' AND TABLE_SCHEMA = ''' + @schemaName + '''ORDER BY ORDINAL_POSITIONFOR XML PATH('''')), 1, 2, '''')', @paramDefinition = N'@cols nvarchar(max) OUTPUT' EXECUTE sp_executesql @cols, @paramDefinition, @cols = @cols OUTPUT -- Loop through all the restored data pages, -- read data from them and insert them into temp table -- which you can then insert into the orignial deleted table DECLARE dbccPageCursor CURSOR GLOBAL FORWARD_ONLY FOR SELECT [FileID], [PageID], [SQLtoExec] FROM #PagesToRestore ORDER BY [FileID], [PageID] OPEN dbccPageCursor; FETCH NEXT FROM dbccPageCursor INTO @FileID, @PageID, @SQLtoExec; WHILE @@FETCH_STATUS = 0 BEGIN RAISERROR ('---------------------------------------------', 10, 1) WITH NOWAIT; SELECT @sql = 'Loop iteration: ' + CONVERT(VARCHAR(10), @i); RAISERROR (@sql, 10, 1) WITH NOWAIT; SELECT @sql = 'Running: ' + @SQLtoExec RAISERROR (@sql, 10, 1) WITH NOWAIT; -- if something goes wrong with DBCC execution or data gathering, skip it but print error BEGIN TRY INSERT INTO @dbccSinglePage EXEC (@SQLtoExec) -- make the data insert magic happen here IF (SELECT CONVERT(BIGINT, [VALUE]) FROM @dbccSinglePage WHERE [Field] LIKE '%Metadata: ObjectId%') = OBJECT_ID('['+@dbName+'].['+@schemaName +'].['+@tableName+']') BEGIN DELETE @dbccSinglePage WHERE NOT ([ParentObject] LIKE 'Slot % Offset %' AND [Object] LIKE 'Slot % Column %') SELECT @sql = 'USE tempdb; ' + 'IF (OBJECTPROPERTY(object_id(''' + @undeletedTableName + '''), ''TableHasIdentity'') = 1) ' + 'SET IDENTITY_INSERT ' + @undeletedTableName + ' ON; ' + 'INSERT INTO ' + @undeletedTableName + '(' + @cols + ') ' + STUFF((SELECT ' UNION ALL SELECT ' + STUFF((SELECT ', ' + CASE WHEN VALUE = '[NULL]' THEN 'NULL' ELSE '''' + [VALUE] + '''' END FROM ( -- the unicorn help here to correctly set ordinal numbers of columns in a data page -- it's turning STRING order into INT order (1,10,11,2,21 into 1,2,..10,11...21) SELECT [ParentObject], [Object], Field, VALUE, RIGHT('00000' + O1, 6) AS ParentObjectOrder, RIGHT('00000' + REVERSE(LEFT(O2, CHARINDEX(' ', O2)-1)), 6) AS ObjectOrder FROM ( SELECT [ParentObject], [Object], Field, VALUE, REPLACE(LEFT([ParentObject], CHARINDEX('Offset', [ParentObject])-1), 'Slot ', '') AS O1, REVERSE(LEFT([Object], CHARINDEX('Offset ', [Object])-2)) AS O2 FROM @dbccSinglePage WHERE t.ParentObject = ParentObject )t)t ORDER BY ParentObjectOrder, ObjectOrder FOR XML PATH('')), 1, 2, '') FROM @dbccSinglePage t GROUP BY ParentObject FOR XML PATH('') ), 1, 11, '') + ';' RAISERROR (@sql, 10, 1) WITH NOWAIT; EXEC (@sql) END END TRY BEGIN CATCH SELECT @sql = 'ERROR!!!' + CHAR(10) + CHAR(13) + 'ErrorNumber: ' + ERROR_NUMBER() + '; ErrorMessage' + ERROR_MESSAGE() + CHAR(10) + CHAR(13) + 'FileID: ' + @FileID + '; PageID: ' + @PageID RAISERROR (@sql, 10, 1) WITH NOWAIT; END CATCH DELETE @dbccSinglePage SELECT @sql = 'Pages left to process: ' + CONVERT(VARCHAR(10), @pagesCount - @i) + CHAR(10) + CHAR(13) + CHAR(10) + CHAR(13) + CHAR(10) + CHAR(13), @i = @i+1 RAISERROR (@sql, 10, 1) WITH NOWAIT; FETCH NEXT FROM dbccPageCursor INTO @FileID, @PageID, @SQLtoExec; END CLOSE dbccPageCursor; DEALLOCATE dbccPageCursor; EXEC ('SELECT ''' + @undeletedTableName + ''' as TableName; SELECT * FROM ' + @undeletedTableName)END TRYBEGIN CATCH SELECT ERROR_NUMBER() AS ErrorNumber, ERROR_MESSAGE() AS ErrorMessage IF CURSOR_STATUS ('global', 'dbccPageCursor') >= 0 BEGIN CLOSE dbccPageCursor; DEALLOCATE dbccPageCursor; ENDEND CATCH-- if the table was deleted we need to finish the restore page sequenceIF @tableWasTruncated = 0BEGIN -- take a log tail backup and then restore it to complete page restore process DECLARE @currentDate VARCHAR(30) SELECT @currentDate = CONVERT(VARCHAR(30), GETDATE(), 112) RAISERROR ('Starting Log Tail backup to c:\Temp ...', 10, 1) WITH NOWAIT; PRINT ('BACKUP LOG [' + @dbName + '] TO DISK = ''c:\Temp\' + @dbName + '_TailLogBackup_' + @currentDate + '.trn''') EXEC ('BACKUP LOG [' + @dbName + '] TO DISK = ''c:\Temp\' + @dbName + '_TailLogBackup_' + @currentDate + '.trn''') RAISERROR ('Log Tail backup done.', 10, 1) WITH NOWAIT; RAISERROR ('Starting Log Tail restore from c:\Temp ...', 10, 1) WITH NOWAIT; PRINT ('RESTORE LOG [' + @dbName + '] FROM DISK = ''c:\Temp\' + @dbName + '_TailLogBackup_' + @currentDate + '.trn''') EXEC ('RESTORE LOG [' + @dbName + '] FROM DISK = ''c:\Temp\' + @dbName + '_TailLogBackup_' + @currentDate + '.trn''') RAISERROR ('Log Tail restore done.', 10, 1) WITH NOWAIT;END-- The last step is manual. Insert data from our temporary table to the original deleted table The misconception here is that you can do a single table restore properly in SQL Server. You can't. But with little experimentation you can get pretty close to it. One way to possible remove a dependency on a backup to retrieve deleted pages is to quickly run a similar script to the upper one that gets data directly from data pages while the rows are still marked as ghost records. It could be done if we could beat the ghost record cleanup task.

    Read the article

  • Oracle Data Integrator 11.1.1.5 Complex Files as Sources and Targets

    - by Alex Kotopoulis
    Overview ODI 11.1.1.5 adds the new Complex File technology for use with file sources and targets. The goal is to read or write file structures that are too complex to be parsed using the existing ODI File technology. This includes: Different record types in one list that use different parsing rules Hierarchical lists, for example customers with nested orders Parsing instructions in the file data, such as delimiter types, field lengths, type identifiers Complex headers such as multiple header lines or parseable information in header Skipping of lines  Conditional or choice fields Similar to the ODI File and XML File technologies, the complex file parsing is done through a JDBC driver that exposes the flat file as relational table structures. Complex files are mapped to one or more table structures, as opposed to the (simple) file technology, which always has a one-to-one relationship between file and table. The resulting set of tables follows the same concept as the ODI XML driver, table rows have additional PK-FK relationships to express hierarchy as well as order values to maintain the file order in the resulting table.   The parsing instruction format used for complex files is the nXSD (native XSD) format that is already in use with Oracle BPEL. This format extends the XML Schema standard by adding additional parsing instructions to each element. Using nXSD parsing technology, the native file is converted into an internal XML format. It is important to understand that the XML is streamed to improve performance; there is no size limitation of the native file based on memory size, the XML data is never fully materialized.  The internal XML is then converted to relational schema using the same mapping rules as the ODI XML driver. How to Create an nXSD file Complex file models depend on the nXSD schema for the given file. This nXSD file has to be created using a text editor or the Native Format Builder Wizard that is part of Oracle BPEL. BPEL is included in the ODI Suite, but not in standalone ODI Enterprise Edition. The nXSD format extends the standard XSD format through nxsd attributes. NXSD is a valid XML Schema, since the XSD standard allows extra attributes with their own namespaces. The following is a sample NXSD schema: <?xml version="1.0"?> <xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:nxsd="http://xmlns.oracle.com/pcbpel/nxsd" elementFormDefault="qualified" xmlns:tns="http://xmlns.oracle.com/pcbpel/demoSchema/csv" targetNamespace="http://xmlns.oracle.com/pcbpel/demoSchema/csv" attributeFormDefault="unqualified" nxsd:encoding="US-ASCII" nxsd:stream="chars" nxsd:version="NXSD"> <xsd:element name="Root">         <xsd:complexType><xsd:sequence>       <xsd:element name="Header">                 <xsd:complexType><xsd:sequence>                         <xsd:element name="Branch" type="xsd:string" nxsd:style="terminated" nxsd:terminatedBy=","/>                         <xsd:element name="ListDate" type="xsd:string" nxsd:style="terminated" nxsd:terminatedBy="${eol}"/>                         </xsd:sequence></xsd:complexType>                         </xsd:element>                 </xsd:sequence></xsd:complexType>         <xsd:element name="Customer" maxOccurs="unbounded">                 <xsd:complexType><xsd:sequence>                 <xsd:element name="Name" type="xsd:string" nxsd:style="terminated" nxsd:terminatedBy=","/>                         <xsd:element name="Street" type="xsd:string" nxsd:style="terminated" nxsd:terminatedBy="," />                         <xsd:element name="City" type="xsd:string" nxsd:style="terminated" nxsd:terminatedBy="${eol}" />                         </xsd:sequence></xsd:complexType>                         </xsd:element>                 </xsd:sequence></xsd:complexType> </xsd:element> </xsd:schema> The nXSD schema annotates elements to describe their position and delimiters within the flat text file. The schema above uses almost exclusively the nxsd:terminatedBy instruction to look for the next terminator chars. There are various constructs in nXSD to parse fixed length fields, look ahead in the document for string occurences, perform conditional logic, use variables to remember state, and many more. nXSD files can either be written manually using an XML Schema Editor or created using the Native Format Builder Wizard. Both Native Format Builder Wizard as well as the nXSD language are described in the Application Server Adapter Users Guide. The way to start the Native Format Builder in BPEL is to create a new File Adapter; in step 8 of the Adapter Configuration Wizard a new Schema for Native Format can be created:   The Native Format Builder guides through a number of steps to generate the nXSD based on a sample native file. If the format is complex, it is often a good idea to “approximate” it with a similar simple format and then add the complex components manually.  The resulting *.xsd file can be copied and used as the format for ODI, other BPEL constructs such as the file adapter definition are not relevant for ODI. Using this technique it is also possible to parse the same file format in SOA Suite and ODI, for example using SOA for small real-time messages, and ODI for large batches. This nXSD schema in this example describes a file with a header row containing data and 3 string fields per row delimited by commas, for example: Redwood City Downtown Branch, 06/01/2011 Ebeneezer Scrooge, Sandy Lane, Atherton Tiny Tim, Winton Terrace, Menlo Park The ODI Complex File JDBC driver exposes the file structure through a set of relational tables with PK-FK relationships. The tables for this example are: Table ROOT (1 row): ROOTPK Primary Key for root element SNPSFILENAME Name of the file SNPSFILEPATH Path of the file SNPSLOADDATE Date of load Table HEADER (1 row): ROOTFK Foreign Key to ROOT record ROWORDER Order of row in native document BRANCH Data BRANCHORDER Order of Branch within row LISTDATE Data LISTDATEORDER Order of ListDate within row Table ADDRESS (2 rows): ROOTFK Foreign Key to ROOT record ROWORDER Order of row in native document NAME Data NAMEORDER Oder of Name within row STREET Data STREETORDER Order of Street within row CITY Data CITYORDER Order of City within row Every table has PK and/or FK fields to reflect the document hierarchy through relationships. In this example this is trivial since the HEADER and all CUSTOMER records point back to the PK of ROOT. Deeper nested documents require this to identify parent elements. All tables also have a ROWORDER field to define the order of rows, as well as order fields for each column, in case the order of columns varies in the original document and needs to be maintained. If order is not relevant, these fields can be ignored. How to Create an Complex File Data Server in ODI After creating the nXSD file and a test data file, and storing it on the local file system accessible to ODI, you can go to the ODI Topology Navigator to create a Data Server and Physical Schema under the Complex File technology. This technology follows the conventions of other ODI technologies and is very similar to the XML technology. The parsing settings such as the source native file, the nXSD schema file, the root element, as well as the external database can be set in the JDBC URL: The use of an external database defined by dbprops is optional, but is strongly recommended for production use. Ideally, the staging database should be used for this. Also, when using a complex file exclusively for read purposes, it is recommended to use the ro=true property to ensure the file is not unnecessarily synchronized back from the database when the connection is closed. A data file is always required to be present  at the filename path during design-time. Without this file, operations like testing the connection, reading the model data, or reverse engineering the model will fail.  All properties of the Complex File JDBC Driver are documented in the Oracle Fusion Middleware Connectivity and Knowledge Modules Guide for Oracle Data Integrator in Appendix C: Oracle Data Integrator Driver for Complex Files Reference. David Allan has created a great viewlet Complex File Processing - 0 to 60 which shows the creation of a Complex File data server as well as a model based on this server. How to Create Models based on an Complex File Schema Once physical schema and logical schema have been created, the Complex File can be used to create a Model as if it were based on a database. When reverse-engineering the Model, data stores(tables) for each XSD element of complex type will be created. Use of complex files as sources is straightforward; when using them as targets it has to be made sure that all dependent tables have matching PK-FK pairs; the same applies to the XML driver as well. Debugging and Error Handling There are different ways to test an nXSD file. The Native Format Builder Wizard can be used even if the nXSD wasn’t created in it; it will show issues related to the schema and/or test data. In ODI, the nXSD  will be parsed and run against the existing test XML file when testing a connection in the Dataserver. If either the nXSD has an error or the data is non-compliant to the schema, an error will be displayed. Sample error message: Error while reading native data. [Line=1, Col=5] Not enough data available in the input, when trying to read data of length "19" for "element with name D1" from the specified position, using "style" as "fixedLength" and "length" as "". Ensure that there is enough data from the specified position in the input. Complex File FAQ Is the size of the native file limited by available memory? No, since the native data is streamed through the driver, only the available space in the staging database limits the size of the data. There are limits on individual field sizes, though; a single large object field needs to fit in memory. Should I always use the complex file driver instead of the file driver in ODI now? No, use the file technology for all simple file parsing tasks, for example any fixed-length or delimited files that just have one row format and can be mapped into a simple table. Because of its narrow assumptions the ODI file driver is easy to configure within ODI and can stream file data without writing it into a database. The complex file driver should be used whenever the use case cannot be handled through the file driver. Are we generating XML out of flat files before we write it into a database? We don’t materialize any XML as part of parsing a flat file, either in memory or on disk. The data produced by the XML parser is streamed in Java objects that just use XSD-derived nXSD schema as its type system. We use the nXSD schema because is the standard for describing complex flat file metadata in Oracle Fusion Middleware, and enables users to share schemas across products. Is the nXSD file interchangeable with SOA Suite? Yes, ODI can use the same nXSD files as SOA Suite, allowing mixed use cases with the same data format. Can I start the Native Format Builder from the ODI Studio? No, the Native Format Builder has to be started from a JDeveloper with BPEL instance. You can get BPEL as part of the SOA Suite bundle. Users without SOA Suite can manually develop nXSD files using XSD editors. When is the database data written back to the native file? Data is synchronized using the SYNCHRONIZE and CREATE FILE commands, and when the JDBC connection is closed. It is recommended to set the ro or read_only property to true when a file is exclusively used for reading so that no unnecessary write-backs occur. Is the nXSD metadata part of the ODI Master or Work Repository? No, the data server definition in the master repository only contains the JDBC URL with file paths; the nXSD files have to be accessible on the file systems where the JDBC driver is executed during production, either by copying or by using a network file system. Where can I find sample nXSD files? The Application Server Adapter Users Guide contains nXSD samples for various different use cases.

    Read the article

  • Building a &ldquo;real&rdquo; extension for Expression Blend

    - by Timmy Kokke
    .Last time I showed you how to get started building extensions for Expression Blend. Lets build a useful extension this time and go a bit deeper into Blend. Source of project  => here Compiled dll => here (extract into /extensions folder of Expression Blend)   The Extension When working on large Xaml files in Blend it’s often hard to find a specific control in the "Objects and Timeline Pane”. An extension that searches the active document and presents all elements that satisfy the query would be helpful. When the user starts typing a search query a search will be performed and the results are shown in the list. After the user selects an item in the results list, the control in the "Objects and Timeline Pane” will be selected. Below is a sketch of what it is going to look like. The Solution Create a new WPF User Control project as shown in the earlier tutorial in the Configuring the extension project section, but name it AdvancedSearch this time. Delete the default UserControl1.Xaml to clear the solution (a new user control will be added later thought, but adding a user control is easier then renaming one). Create the main entry point of the addin by adding a new class to the solution and naming this  AdvancedSearchPackage. Add a reference to Microsoft.Expression.Extensibility and to System.ComponentModel.Composition . Implement the IPackage interface and add the Export attribute from the MEF to the definition. While you’re at it. Add references to Microsoft.Expression.DesignSurface, Microsoft.Expression.FrameWork and Microsoft.Expression.Markup. These will be used later. The Load method from the IPackage interface is going to create a ViewModel to bind to from the UI. Add another class to the solution and name this AdvancedSearchViewModel. This class needs to implement the INotifyPropertyChanged interface to enable notifications to the view.  Add a constructor to the class that takes an IServices interface as a parameter. Create a new instance of the AdvancedSearchViewModel in the load method in the AdvanceSearchPackage class. The AdvancedSearchPackage class should looks like this now:   using System.ComponentModel.Composition; using Microsoft.Expression.Extensibility;   namespace AdvancedSearch { [Export(typeof(IPackage))] public class AdvancedSearchPackage:IPackage {   public void Load(IServices services) { new AdvancedSearchViewModel(services); }   public void Unload() { } } }   Add a new UserControl to the project and name this AdvancedSearchView. The View will be created by the ViewModel, which will pass itself to the constructor of the view. Change the constructor of the View to take a AdvancedSearchViewModel object as a parameter. Add a private field to store the ViewModel and set this field in the constructor. Point the DataContext of the view to the ViewModel. The View will look something like this now:   namespace AdvancedSearch { public partial class AdvancedSearchView:UserControl { private readonly AdvancedSearchViewModel _advancedSearchViewModel;   public AdvancedSearchView(AdvancedSearchViewModel advancedSearchViewModel) { _advancedSearchViewModel = advancedSearchViewModel; InitializeComponent(); this.DataContext = _advancedSearchViewModel; } } }   The View is going to be created in the constructor of the ViewModel and stored in a read only property.   public FrameworkElement View { get; private set; }   public AdvancedSearchViewModel(IServices services) { _services = services; View = new AdvancedSearchView(this); } The last thing the solution needs before we’ll wire things up is a new class, PossibleNode. This class will be used later to store the search results. The solution should look like this now:   Adding UI to the UI The extension should build and run now, although nothing is showing up in Blend yet. To enable the user to perform a search query add a TextBox and a ListBox to the AdvancedSearchView.xaml file. I’ve set the rows of the grid too to make them look a little better. Add the TextChanged event to the TextBox and the SelectionChanged event to the ListBox, we’ll need those later on. <Grid> <Grid.RowDefinitions> <RowDefinition Height="32" /> <RowDefinition Height="*" /> </Grid.RowDefinitions> <TextBox TextChanged="SearchQueryTextChanged" HorizontalAlignment="Stretch" Margin="4" Name="SearchQuery" VerticalAlignment="Stretch" /> <ListBox SelectionChanged="SearchResultSelectionChanged" HorizontalAlignment="Stretch" Margin="4" Name="SearchResult" VerticalAlignment="Stretch" Grid.Row="1" /> </Grid>   This will create a user interface like: To make the View show up in Blend it has to be registered with the WindowService. The GetService<T> method is used to get services from Blend, which are your entry points into Blend.When writing extensions you will encounter this method very often. In this case we’re asking for an IWindowService interface. The IWindowService interface serves events for changing windows and themes, is used for adding or removing resources and is used for registering and unregistering Palettes. All panes in Blend are palettes and are registered thru the RegisterPalette method. The first parameter passed to this method is a string containing a unique ID for the palette. This ID can be used to get access to the palette later. The second parameter is the View. The third parameter is a title for the pane. This title is shown when the pane is visible. It is also shown in the window menu of Blend. The last parameter is a KeyBinding. I have chosen Ctrl+Shift+F to call the Advanced Search pane. This value is also shown in the window menu of Blend.   services.GetService<IWindowService>().RegisterPalette( "AdvancedSearch", viewModel.View, "Advanced Search", new KeyBinding { Key = Key.F, Modifiers = ModifierKeys.Control | ModifierKeys.Shift } );   You can compiler and run now. After Blend starts you can hit Ctrl+Shift+F or go the windows menu to call the advanced search extension. Searching for controls The search has to be cleared on every change of the active document. The DocumentServices fires an event every time a new document is opened, a document is closed or another document view is selected. Add the following line to the constructor of the ViewModel to handle the ActiveDocumentChanged event:   _services.GetService<IDocumentService>().ActiveDocumentChanged += ActiveDocumentChanged;   And implement the ActiveDocumentChanged method:   private void ActiveDocumentChanged(object sender, DocumentChangedEventArgs e) { }   To get to the contents of the document we first need to get access to the “Objects and Timeline” pane. This pane is registered in the PaletteRegistry in the same way as this extension has registered itself. The palettes are accessible thru an associative array. All you need to provide is the Identifier of the palette you want. The Id of the “Objects and Timeline” pane is “Designer_TimelinePane”. I’ve included a list of the other default panes at the bottom of this article. Each palette has a Content property which can be cast to the type of the pane.   var timelinePane = (TimelinePane)_services.GetService<IWindowService>() .PaletteRegistry["Designer_TimelinePane"] .Content;   Add a private field to the top of the AdvancedSearchViewModel class to store the active SceneViewModel. The SceneViewModel is needed to set the current selection and to get the little icons for the type of control.   private SceneViewModel _activeSceneViewModel;   When the active SceneViewModel changes, the ActiveSceneViewModel is stored in this field. The list of possible nodes is cleared and an PropertyChanged event is fired for this list to notify the UI to clear the list. This will make the eventhandler look like this: private void ActiveDocumentChanged(object sender, DocumentChangedEventArgs e) { var timelinePane = (TimelinePane)_services.GetService<IWindowService>() .PaletteRegistry["Designer_TimelinePane"].Content;   _activeSceneViewModel = timelinePane.ActiveSceneViewModel; PossibleNodes = new List<PossibleNode>(); InvokePropertyChanged("PossibleNodes"); } The PossibleNode class used to store information about the controls found by the search. It’s a dumb data class with only 3 properties, the name of the control, the SceneNode and a brush used for the little icon. The SceneNode is the base class for every possible object you can create in Blend, like Brushes, Controls, Annotations, ResourceDictionaries and VisualStates. The entire PossibleNode class looks like this:   using System.Windows.Media; using Microsoft.Expression.DesignSurface.ViewModel;   namespace AdvancedSearch { public class PossibleNode { public string Name { get; set; } public SceneNode SceneNode { get; set; } public DrawingBrush IconBrush { get; set; } } }   Add these two methods to the AdvancedSearchViewModel class:   public void Search(string searchText) { } public void SelectElement(PossibleNode node){ }   Both these methods are going to be called from the view. The Search method performs the search and updates the PossibleNodes list.  The controls in the active document can be accessed thru TimeLineItemsManager class. This class contains a read only collection of TimeLineItems. By using a Linq query the possible nodes are selected and placed in the PossibleNodes list.   var timelineItemManager = new TimelineItemManager(_activeSceneViewModel); PossibleNodes = new List<PossibleNode>( (from d in timelineItemManager.ItemList where d.DisplayName.ToLowerInvariant().StartsWith( searchText.ToLowerInvariant()) select new PossibleNode() { IconBrush = d.IconBrush, SceneNode = d.SceneNode, Name = d.DisplayName }).ToList() ); InvokePropertyChanged(InternalConst.PossibleNodes);   The Select method is pretty straight forward. It contains two lines.The first to clear the selection. Otherwise the selected element would be added to the current selection. The second line selects the nodes. It is given a new array with the node to be selected.   _activeSceneViewModel.ClearSelections(); _activeSceneViewModel.SelectNodes(new[] { node.SceneNode });   The last thing that needs to be done is to wire the whole thing to the View. The two event handlers just call the Search and SelectElement methods on the ViewModel.   private void SearchQueryTextChanged(object sender, TextChangedEventArgs e) { _advancedSearchViewModel.Search(SearchQuery.Text); }   private void SearchResultSelectionChanged(object sender, SelectionChangedEventArgs e) { if(e.AddedItems.Count>0) { _advancedSearchViewModel.SelectElement(e.AddedItems[0] as PossibleNode); } }   The Listbox has to be bound to the PossibleNodes list and a simple DataTemplate is added to show the selection. The IconWithOverlay control can be found in the Microsoft.Expression.DesignSurface.UserInterface.Timeline.UI namespace in the Microsoft.Expression.DesignSurface assembly. The ListBox should look something like:   <ListBox SelectionChanged="SearchResultSelectionChanged" HorizontalAlignment="Stretch" Margin="4" Name="SearchResult" VerticalAlignment="Stretch" Grid.Row="1" ItemsSource="{Binding PossibleNodes}"> <ListBox.ItemTemplate> <DataTemplate> <StackPanel Orientation="Horizontal"> <tlui:IconWithOverlay Margin="2,0,10,0" Width="12" Height="12" SourceBrush="{Binding Path=IconBrush, Mode=OneWay}" /> <TextBlock Text="{Binding Name}"/> </StackPanel> </DataTemplate> </ListBox.ItemTemplate> </ListBox>   Compile and run. Inside Blend the extension could look something like below. What’s Next When you’ve got the extension running. Try placing breakpoints in the code and see what else is in there. There’s a lot to explore and build extension on. I personally would love an extension to search for resources. Last but not least, you can download the source of project here.  If you have any questions let me know. If you just want to use this extension, you can download the compiled dll here. Just extract the . zip into the /extensions folder of Expression Blend. Notes Target framework I ran into some issues when using the .NET Framework 4 Client Profile as a target framework. I got some strange error saying certain obvious namespaces could not be found, Microsoft.Expression in my case. If you run into something like this, try setting the target framework to .NET Framework 4 instead of the client version.   Identifiers of default panes Identifier Type Title Designer_TimelinePane TimelinePane Objects and Timeline Designer_ToolPane ToolPane Tools Designer_ProjectPane ProjectPane Projects Designer_DataPane DataPane Data Designer_ResourcePane ResourcePane Resources Designer_PropertyInspector PropertyInspector Properties Designer_TriggersPane TriggersPane Triggers Interaction_Skin SkinView States Designer_AssetPane AssetPane Assets Interaction_Parts PartsPane Parts Designer_ResultsPane ResultsPane Results

    Read the article

  • 64-bit Archives Needed

    - by user9154181
    A little over a year ago, we received a question from someone who was trying to build software on Solaris. He was getting errors from the ar command when creating an archive. At that time, the ar command on Solaris was a 32-bit command. There was more than 2GB of data, and the ar command was hitting the file size limit for a 32-bit process that doesn't use the largefile APIs. Even in 2011, 2GB is a very large amount of code, so we had not heard this one before. Most of our toolchain was extended to handle 64-bit sized data back in the 1990's, but archives were not changed, presumably because there was no perceived need for it. Since then of course, programs have continued to get larger, and in 2010, the time had finally come to investigate the issue and find a way to provide for larger archives. As part of that process, I had to do a deep dive into the archive format, and also do some Unix archeology. I'm going to record what I learned here, to document what Solaris does, and in the hope that it might help someone else trying to solve the same problem for their platform. Archive Format Details Archives are hardly cutting edge technology. They are still used of course, but their basic form hasn't changed in decades. Other than to fix a bug, which is rare, we don't tend to touch that code much. The archive file format is described in /usr/include/ar.h, and I won't repeat the details here. Instead, here is a rough overview of the archive file format, implemented by System V Release 4 (SVR4) Unix systems such as Solaris: Every archive starts with a "magic number". This is a sequence of 8 characters: "!<arch>\n". The magic number is followed by 1 or more members. A member starts with a fixed header, defined by the ar_hdr structure in/usr/include/ar.h. Immediately following the header comes the data for the member. Members must be padded at the end with newline characters so that they have even length. The requirement to pad members to an even length is a dead giveaway as to the age of the archive format. It tells you that this format dates from the 1970's, and more specifically from the era of 16-bit systems such as the PDP-11 that Unix was originally developed on. A 32-bit system would have required 4 bytes, and 64-bit systems such as we use today would probably have required 8 bytes. 2 byte alignment is a poor choice for ELF object archive members. 32-bit objects require 4 byte alignment, and 64-bit objects require 64-bit alignment. The link-editor uses mmap() to process archives, and if the members have the wrong alignment, we have to slide (copy) them to the correct alignment before we can access the ELF data structures inside. The archive format requires 2 byte padding, but it doesn't prohibit more. The Solaris ar command takes advantage of this, and pads ELF object members to 8 byte boundaries. Anything else is padded to 2 as required by the format. The archive header (ar_hdr) represents all numeric values using an ASCII text representation rather than as binary integers. This means that an archive that contains only text members can be viewed using tools such as cat, more, or a text editor. The original designers of this format clearly thought that archives would be used for many file types, and not just for objects. Things didn't turn out that way of course — nearly all archives contain relocatable objects for a single operating system and machine, and are used primarily as input to the link-editor (ld). Archives can have special members that are created by the ar command rather than being supplied by the user. These special members are all distinguished by having a name that starts with the slash (/) character. This is an unambiguous marker that says that the user could not have supplied it. The reason for this is that regular archive members are given the plain name of the file that was inserted to create them, and any path components are stripped off. Slash is the delimiter character used by Unix to separate path components, and as such cannot occur within a plain file name. The ar command hides the special members from you when you list the contents of an archive, so most users don't know that they exist. There are only two possible special members: A symbol table that maps ELF symbols to the object archive member that provides it, and a string table used to hold member names that exceed 15 characters. The '/' convention for tagging special members provides room for adding more such members should the need arise. As I will discuss below, we took advantage of this fact to add an alternate 64-bit symbol table special member which is used in archives that are larger than 4GB. When an archive contains ELF object members, the ar command builds a special archive member known as the symbol table that maps all ELF symbols in the object to the archive member that provides it. The link-editor uses this symbol table to determine which symbols are provided by the objects in that archive. If an archive has a symbol table, it will always be the first member in the archive, immediately following the magic number. Unlike member headers, symbol tables do use binary integers to represent offsets. These integers are always stored in big-endian format, even on a little endian host such as x86. The archive header (ar_hdr) provides 15 characters for representing the member name. If any member has a name that is longer than this, then the real name is written into a special archive member called the string table, and the member's name field instead contains a slash (/) character followed by a decimal representation of the offset of the real name within the string table. The string table is required to precede all normal archive members, so it will be the second member if the archive contains a symbol table, and the first member otherwise. The archive format is not designed to make finding a given member easy. Such operations move through the archive from front to back examining each member in turn, and run in O(n) time. This would be bad if archives were commonly used in that manner, but in general, they are not. Typically, the ar command is used to build an new archive from scratch, inserting all the objects in one operation, and then the link-editor accesses the members in the archive in constant time by using the offsets provided by the symbol table. Both of these operations are reasonably efficient. However, listing the contents of a large archive with the ar command can be rather slow. Factors That Limit Solaris Archive Size As is often the case, there was more than one limiting factor preventing Solaris archives from growing beyond the 32-bit limits of 2GB (32-bit signed) and 4GB (32-bit unsigned). These limits are listed in the order they are hit as archive size grows, so the earlier ones mask those that follow. The original Solaris archive file format can handle sizes up to 4GB without issue. However, the ar command was delivered as a 32-bit executable that did not use the largefile APIs. As such, the ar command itself could not create a file larger than 2GB. One can solve this by building ar with the largefile APIs which would allow it to reach 4GB, but a simpler and better answer is to deliver a 64-bit ar, which has the ability to scale well past 4GB. Symbol table offsets are stored as 32-bit big-endian binary integers, which limits the maximum archive size to 4GB. To get around this limit requires a different symbol table format, or an extension mechanism to the current one, similar in nature to the way member names longer than 15 characters are handled in member headers. The size field in the archive member header (ar_hdr) is an ASCII string capable of representing a 32-bit unsigned value. This places a 4GB size limit on the size of any individual member in an archive. In considering format extensions to get past these limits, it is important to remember that very few archives will require the ability to scale past 4GB for many years. The old format, while no beauty, continues to be sufficient for its purpose. This argues for a backward compatible fix that allows newer versions of Solaris to produce archives that are compatible with older versions of the system unless the size of the archive exceeds 4GB. Archive Format Differences Among Unix Variants While considering how to extend Solaris archives to scale to 64-bits, I wanted to know how similar archives from other Unix systems are to those produced by Solaris, and whether they had already solved the 64-bit issue. I've successfully moved archives between different Unix systems before with good luck, so I knew that there was some commonality. If it turned out that there was already a viable defacto standard for 64-bit archives, it would obviously be better to adopt that rather than invent something new. The archive file format is not formally standardized. However, the ar command and archive format were part of the original Unix from Bell Labs. Other systems started with that format, extending it in various often incompatible ways, but usually with the same common shared core. Most of these systems use the same magic number to identify their archives, despite the fact that their archives are not always fully compatible with each other. It is often true that archives can be copied between different Unix variants, and if the member names are short enough, the ar command from one system can often read archives produced on another. In practice, it is rare to find an archive containing anything other than objects for a single operating system and machine type. Such an archive is only of use on the type of system that created it, and is only used on that system. This is probably why cross platform compatibility of archives between Unix variants has never been an issue. Otherwise, the use of the same magic number in archives with incompatible formats would be a problem. I was able to find information for a number of Unix variants, described below. These can be divided roughly into three tribes, SVR4 Unix, BSD Unix, and IBM AIX. Solaris is a SVR4 Unix, and its archives are completely compatible with those from the other members of that group (GNU/Linux, HP-UX, and SGI IRIX). AIX AIX is an exception to rule that Unix archive formats are all based on the original Bell labs Unix format. It appears that AIX supports 2 formats (small and big), both of which differ in fundamental ways from other Unix systems: These formats use a different magic number than the standard one used by Solaris and other Unix variants. They include support for removing archive members from a file without reallocating the file, marking dead areas as unused, and reusing them when new archive items are inserted. They have a special table of contents member (File Member Header) which lets you find out everything that's in the archive without having to actually traverse the entire file. Their symbol table members are quite similar to those from other systems though. Their member headers are doubly linked, containing offsets to both the previous and next members. Of the Unix systems described here, AIX has the only format I saw that will have reasonable insert/delete performance for really large archives. Everyone else has O(n) performance, and are going to be slow to use with large archives. BSD BSD has gone through 4 versions of archive format, which are described in their manpage. They use the same member header as SVR4, but their symbol table format is different, and their scheme for long member names puts the name directly after the member header rather than into a string table. GNU/Linux The GNU toolchain uses the SVR4 format, and is compatible with Solaris. HP-UX HP-UX seems to follow the SVR4 model, and is compatible with Solaris. IRIX IRIX has 32 and 64-bit archives. The 32-bit format is the standard SVR4 format, and is compatible with Solaris. The 64-bit format is the same, except that the symbol table uses 64-bit integers. IRIX assumes that an archive contains objects of a single ELFCLASS/MACHINE, and any archive containing ELFCLASS64 objects receives a 64-bit symbol table. Although they only use it for 64-bit objects, nothing in the archive format limits it to ELFCLASS64. It would be perfectly valid to produce a 64-bit symbol table in an archive containing 32-bit objects, text files, or anything else. Tru64 Unix (Digital/Compaq/HP) Tru64 Unix uses a format much like ours, but their symbol table is a hash table, making specific symbol lookup much faster. The Solaris link-editor uses archives by examining the entire symbol table looking for unsatisfied symbols for the link, and not by looking up individual symbols, so there would be no benefit to Solaris from such a hash table. The Tru64 ld must use a different approach in which the hash table pays off for them. Widening the existing SVR4 archive symbol tables rather than inventing something new is the simplest path forward. There is ample precedent for this approach in the ELF world. When ELF was extended to support 64-bit objects, the approach was largely to take the existing data structures, and define 64-bit versions of them. We called the old set ELF32, and the new set ELF64. My guess is that there was no need to widen the archive format at that time, but had there been, it seems obvious that this is how it would have been done. The Implementation of 64-bit Solaris Archives As mentioned earlier, there was no desire to improve the fundamental nature of archives. They have always had O(n) insert/delete behavior, and for the most part it hasn't mattered. AIX made efforts to improve this, but those efforts did not find widespread adoption. For the purposes of link-editing, which is essentially the only thing that archives are used for, the existing format is adequate, and issues of backward compatibility trump the desire to do something technically better. Widening the existing symbol table format to 64-bits is therefore the obvious way to proceed. For Solaris 11, I implemented that, and I also updated the ar command so that a 64-bit version is run by default. This eliminates the 2 most significant limits to archive size, leaving only the limit on an individual archive member. We only generate a 64-bit symbol table if the archive exceeds 4GB, or when the new -S option to the ar command is used. This maximizes backward compatibility, as an archive produced by Solaris 11 is highly likely to be less than 4GB in size, and will therefore employ the same format understood by older versions of the system. The main reason for the existence of the -S option is to allow us to test the 64-bit format without having to construct huge archives to do so. I don't believe it will find much use outside of that. Other than the new ability to create and use extremely large archives, this change is largely invisible to the end user. When reading an archive, the ar command will transparently accept either form of symbol table. Similarly, the ELF library (libelf) has been updated to understand either format. Users of libelf (such as the link-editor ld) do not need to be modified to use the new format, because these changes are encapsulated behind the existing functions provided by libelf. As mentioned above, this work did not lift the limit on the maximum size of an individual archive member. That limit remains fixed at 4GB for now. This is not because we think objects will never get that large, for the history of computing says otherwise. Rather, this is based on an estimation that single relocatable objects of that size will not appear for a decade or two. A lot can change in that time, and it is better not to overengineer things by writing code that will sit and rot for years without being used. It is not too soon however to have a plan for that eventuality. When the time comes when this limit needs to be lifted, I believe that there is a simple solution that is consistent with the existing format. The archive member header size field is an ASCII string, like the name, and as such, the overflow scheme used for long names can also be used to handle the size. The size string would be placed into the archive string table, and its offset in the string table would then be written into the archive header size field using the same format "/ddd" used for overflowed names.

    Read the article

< Previous Page | 142 143 144 145 146 147 148 149 150 151 152 153  | Next Page >