Search Results

Search found 21392 results on 856 pages for 'order of operations'.

facebook twitter digg google delicious technorati stumbleupon myspace wordpress linkedin gmail igoogle windows live tumblr viadeo yahoo buzz yahoo mail yahoo bookmarks favorites email print

Page 17/856 | < Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >

  • Lightweight spinlocks built from GCC atomic operations?

    - by Thomas
    I'd like to minimize synchronization and write lock-free code when possible in a project of mine. When absolutely necessary I'd love to substitute light-weight spinlocks built from atomic operations for pthread and win32 mutex locks. My understanding is that these are system calls underneath and could cause a context switch (which may be unnecessary for very quick critical sections where simply spinning a few times would be preferable). The atomic operations I'm referring to are well documented here: http://gcc.gnu.org/onlinedocs/gcc-4.4.1/gcc/Atomic-Builtins.html Here is an example to illustrate what I'm talking about. Imagine a RB-tree with multiple readers and writers possible. RBTree::exists() is read-only and thread safe, RBTree::insert() would require exclusive access by a single writer (and no readers) to be safe. Some code: class IntSetTest { private: unsigned short lock; RBTree<int>* myset; public: // ... void add_number(int n) { // Aquire once locked==false (atomic) while (__sync_bool_compare_and_swap(&lock, 0, 0xffff) == false); // Perform a thread-unsafe operation on the set myset->insert(n); // Unlock (atomic) __sync_bool_compare_and_swap(&lock, 0xffff, 0); } bool check_number(int n) { // Increment once the lock is below 0xffff u16 savedlock = lock; while (savedlock == 0xffff || __sync_bool_compare_and_swap(&lock, savedlock, savedlock+1) == false) savedlock = lock; // Perform read-only operation bool exists = tree->exists(n); // Decrement savedlock = lock; while (__sync_bool_compare_and_swap(&lock, savedlock, savedlock-1) == false) savedlock = lock; return exists; } }; (lets assume it need not be exception-safe) Is this code indeed thread-safe? Are there any pros/cons to this idea? Any advice? Is the use of spinlocks like this a bad idea if the threads are not truly concurrent? Thanks in advance. ;)

    Read the article

  • Atomic operations on several transactionless external systems

    - by simendsjo
    Say you have an application connecting 3 different external systems. You need to update something in all 3. In case of a failure, you need to roll back the operations. This is not a hard thing to implement, but say operation 3 fails, and when rolling back, the rollback for operation 1 fails! Now the first external system is in an invalid state... I'm thinking a possible solution is to shut down the application and forcing a manual fix of the external system, but then again... It might already have used this information (and perhaps that's why it failed), or we might not have sufficient access. Or it might not even be a good way to rollback the action! Are there some good ways of handling such cases? EDIT: Some application details.. It's a multi user web application. Most of the work is done with scheduled jobs (through Quartz.Net), so most operations is run in it's own thread. Some user actions should trigger jobs that update several systems though. The external systems are somewhat unstable. I Was thinking of changing the application to use the Command and Unit Of Work pattern

    Read the article

  • Azure batch operations delete several blobs and tables

    - by reft
    I have a function that deletes every table & blob that belongs to the affected user. CloudTable uploadTable = CloudStorageServices.GetCloudUploadsTable(); TableQuery<UploadEntity> uploadQuery = uploadTable.CreateQuery<UploadEntity>(); List<UploadEntity> uploadEntity = (from e in uploadTable.ExecuteQuery(uploadQuery) where e.PartitionKey == "uploads" && e.UserName == User.Idendity.Name select e).ToList(); foreach (UploadEntity uploadTableItem in uploadEntity) { //Delete table TableOperation retrieveOperationUploads = TableOperation.Retrieve<UploadEntity>("uploads", uploadTableItem.RowKey); TableResult retrievedResultUploads = uploadTable.Execute(retrieveOperationUploads); UploadEntity deleteEntityUploads = (UploadEntity)retrievedResultUploads.Result; TableOperation deleteOperationUploads = TableOperation.Delete(deleteEntityUploads); uploadTable.Execute(deleteOperationUploads); //Delete blob CloudBlobContainer blobContainer = CloudStorageServices.GetCloudBlobsContainer(); CloudBlockBlob blob = blobContainer.GetBlockBlobReference(uploadTableItem.BlobName); blob.Delete(); } Each table got its own blob, so if the list contains 3 uploadentities, the 3 table and the 3 blobs will be deleted. I heard you can use table batch operations for reduce cost and load. I tried it, but failed miserable. Anyone intrested in helping me:)? Im guessing tablebatch operations are for tables only, so its a no go for blobs, right? How would you add tablebatchoperations for this code? Do you see any other improvements that can be done? Thanks!

    Read the article

  • When to define SDD operations System->Actor?

    - by devoured elysium
    I am having some trouble understanding how to make SDDs, as I don't fully grasp why in some cases one should define operations for System - Actor and in others don't. Here is an example: 1) The User tells the System that wants to buy some tickets, stating his client number. 2) The System confirms that the given client number is valid. 3) The User tells the System the movie that wants to see. 4) The System shows the set of available sessions and seats for that movie. 5) The System asks the user which session/seat he wants. 6) The user tells the System the chosen session/seat. This would be converted to: a) -----> tellClientNumber(clientNumber) b) <----- validClientNumber c) -----> tellMovieToSee(movie) d) <----- showsAvailableSeatsHours e) -----> tellSystemChosenSessionSeat(session, seat) I know that when we are dealing with SDD's we are still far away from coding. But I can't help trying to imagine how it how it would have been had I to convert it right away to code: I can understand 1) and 2). It's like if it was a C#/Java method with the following signature: boolean tellClientNumber(clientNumber) so I put both on the SDD. Then, we have the pair 3) 4). I can imagine that as something as: SomeDataStructureThatHoldsAvailableSessionsSeats tellSystemMovieToSee(movie) Now, the problem: From what I've come to understand, my lecturer says that we shouldn't make an operation on the SDD for 5) as we should only show operations from the Actor to the System and when the System is either presenting us data (as in c)) or validating sent data (such as in b)). I find this odd, as if I try to imagine this like a DOS app where you have to put your input sequencially, it makes sense to make an arrow even for 5). Why is this wrong? How should I try to visualize this? Thanks

    Read the article

  • Python: Behavior of object in set operations

    - by Josh Arenberg
    I'm trying to create a custom object that behaves properly in set operations. I've generally got it working, but I want to make sure I fully understand the implications. In particular, I'm interested in the behavior when there is additional data in the object that is not included in the equal / hash methods. It seems that in the 'intersection' operation, it returns the set of objects that are being compared to, where the 'union' operations returns the set of objects that are being compared. To illustrate: class MyObject: def __init__(self,value,meta): self.value = value self.meta = meta def __eq__(self,other): if self.value == other.value: return True else: return False def __hash__(self): return hash(self.value) a = MyObject('1','left') b = MyObject('1','right') c = MyObject('2','left') d = MyObject('2','right') e = MyObject('3','left') print a == b # True print a == c # False for i in set([a,c,e]).intersection(set([b,d])): print "%s %s" % (i.value,i.meta) #returns: #1 right #2 right for i in set([a,c,e]).union(set([b,d])): print "%s %s" % (i.value,i.meta) #returns: #1 left #3 left #2 left Is this behavior documented somewhere and deterministic? If so, what is the governing principle?

    Read the article

  • how to use q.js promises to work with multiple asynchronous operations

    - by kimsia
    Note: This question is also cross-posted in Q.js mailing list over here. i had a situation with multiple asynchronous operations and the answer I accepted pointed out that using Promises using a library such as q.js would be more beneficial. I am convinced to refactor my code to use Promises but because the code is pretty long, i have trimmed the irrelevant portions and exported the crucial parts into a separate repo. The repo is here and the most important file is this. The requirement is that I want pageSizes to be non-empty after traversing all the dragged'n dropped files. The problem is that the FileAPI operations inside getSizeSettingsFromPage function causes getSizeSettingsFromPage to be async. So I cannot place checkWhenReady(); like this. function traverseFiles() { for (var i=0, l=pages.length; i<l; i++) { getSizeSettingsFromPage(pages[i], calculateRatio); } checkWhenReady(); // this always returns 0. } This works, but it is not ideal. I prefer to call checkWhenReady just ONCE after all the pages have undergone this function calculateRatio successfully. function calculateRatio(width, height, filename) { // .... code pageSizes.add(filename, object); checkWhenReady(); // this works but it is not ideal. I prefer to call this method AFTER all the `pages` have undergone calculateRatio // ..... more code... } How do I refactor the code to make use of Promises in Q.js?

    Read the article

  • DRUPAL: order exposed filter items, be carefully it is not that simple (I cannot user "Sort")

    - by Patrick
    hi, DRUPAL question. I'm using Views, with an exposed filter (Taxonomy). I've downloaded the "better exposed filter" module to display it as checkbox list. a Now, how can I order the tags in the filter list ? "Views Sort" is not the solution because I only can order articles but not the filter items!! I want to add an option (checkbox) for the customer to order the tags alphabetically or leave them in the default order. thanks

    Read the article

  • How does MySQL's ORDER BY RAND() work?

    - by Eugene
    Hi, I've been doing some research and testing on how to do fast random selection in MySQL. In the process I've faced some unexpected results and now I am not fully sure I know how ORDER BY RAND() really works. I always thought that when you do ORDER BY RAND() on the table, MySQL adds a new column to the table which is filled with random values, then it sorts data by that column and then e.g. you take the above value which got there randomly. I've done lots of googling and testing and finally found that the query Jay offers in his blog is indeed the fastest solution: SELECT * FROM Table T JOIN (SELECT CEIL(MAX(ID)*RAND()) AS ID FROM Table) AS x ON T.ID >= x.ID LIMIT 1; While common ORDER BY RAND() takes 30-40 seconds on my test table, his query does the work in 0.1 seconds. He explains how this functions in the blog so I'll just skip this and finally move to the odd thing. My table is a common table with a PRIMARY KEY id and other non-indexed stuff like username, age, etc. Here's the thing I am struggling to explain SELECT * FROM table ORDER BY RAND() LIMIT 1; /*30-40 seconds*/ SELECT id FROM table ORDER BY RAND() LIMIT 1; /*0.25 seconds*/ SELECT id, username FROM table ORDER BY RAND() LIMIT 1; /*90 seconds*/ I was sort of expecting to see approximately the same time for all three queries since I am always sorting on a single column. But for some reason this didn't happen. Please let me know if you any ideas about this. I have a project where I need to do fast ORDER BY RAND() and personally I would prefer to use SELECT id FROM table ORDER BY RAND() LIMIT 1; SELECT * FROM table WHERE id=ID_FROM_PREVIOUS_QUERY LIMIT 1; which, yes, is slower than Jay's method, however it is smaller and easier to understand. My queries are rather big ones with several JOINs and with WHERE clause and while Jay's method still works, the query grows really big and complex because I need to use all the JOINs and WHERE in the JOINed (called x in his query) sub request. Thanks for your time!

    Read the article

  • sort given set of colors in VIBGYOR order

    - by Vijay Selvaraj
    Hi, how do i order a given set of colors from the rainbow in VIBGYOR order. say i input the seven colors in the following order { red, blue, green, yellow, indigo, violet, orange} and i should print the output as {violet, indigo, blue, green, yellow, orange, red}, irrespective of the order i give the output should be as above. Can someone suggest me about implementing this in java program? Thanks, -Vijay

    Read the article

  • Optimize MYSQL Query with Order by

    - by Victor
    Hello, I have seen mysql queries with order by runs slow. Is there any specific way to optimize queries which use order by ? Queries without order by run very fast but with order by its always runs slow. if any one suggest any thing on this as general solutions. Thank You

    Read the article

  • Select Top N Records Ordered by X, But Have Results in Reverse Order

    - by I. J. Kennedy
    I'm trying to get the top N records (when ordered by some column X), but have the result set in reverse order. The following statement is incorrect, but probably demonstrates what I'm after: SELECT * FROM (SELECT TOP 10 * FROM FooTable ORDER BY X DESC) ORDER BY X ASC For example, column X could be an ID or a timestamp; I want the latest 10 records but want them returned in forward chronological order.

    Read the article

  • Saving order of a List in JPA

    - by Rosen Martev
    Hello, I have the following question about JPA: Can I save the order of the elements in a java.util.List? In my application the order in which I put elements in the Lists is important but after I get those collections from the database the order is not the same (as expected). Can you show me a way to deal with this problem? P.S. There is not a field in the entities that I put in the collections by which I can order them. Rosen

    Read the article

  • NServiceBus specify order of Handlers execution

    - by John Simons
    Just wondering if this is the way to specify the order to run a handler (AuthorizationHandler) before all others? public void SpecifyOrder(Order order) { order.Specify(First<AuthorizationHandler>.Then<IHandleMessages<IMessage>>()); } It just feels odd to add Then<IHandleMessages<IMessage>>(). Is there a nicer way of saying to the Bus execute x handler before all others?

    Read the article

  • xts error - order.by requires an appropriate time-based object

    - by Samo
    I can not resolve why error in simple creation of xts object xts(rep(0, NROW(TICK.NYSE)), order.by = index(TICK.NYSE)) Error in xts(rep(0, NROW(TICK.NYSE)), order.by = index(TICK.NYSE)) : order.by requires an appropriate time-based object appeared while this was working perfectly 14 days ago when I last used the same code (since then the only difference is that TICK.NYSE grow in length since data was added since then). More details below: > Sys.getenv("TZ") [1] "EST" > tail(xts(rep(0, NROW(TICK.NYSE)), order.by = index(TICK.NYSE))) Error in xts(rep(0, NROW(TICK.NYSE)), order.by = index(TICK.NYSE)) : order.by requires an appropriate time-based object > class(index(TICK.NYSE["T09:30/T09:31"])) [1] "POSIXct" > tail(xts(rep(0, NROW(tail(TICK.NYSE))), order.by = index(tail(TICK.NYSE)))) Error in xts(rep(0, NROW(tail(TICK.NYSE))), order.by = index(tail(TICK.NYSE))) : order.by requires an appropriate time-based object > tail(TICK.NYSE) TICK-NYSE.Open TICK-NYSE.High TICK-NYSE.Low TICK-NYSE.Close 2012-03-15 14:54:00 -278 -89 -299 -89 2012-03-15 14:55:00 -89 427 -89 201 2012-03-15 14:56:00 201 318 30 234 2012-03-15 14:57:00 234 242 -222 -64 2012-03-15 14:58:00 -64 346 -82 346 2012-03-15 14:59:00 346 525 36 525 TICK-NYSE.Volume TICK-NYSE.WAP TICK-NYSE.hasGaps 2012-03-15 14:54:00 0 0 0 2012-03-15 14:55:00 0 0 0 2012-03-15 14:56:00 0 0 0 2012-03-15 14:57:00 0 0 0 2012-03-15 14:58:00 0 0 0 2012-03-15 14:59:00 0 0 0 TICK-NYSE.Count 2012-03-15 14:54:00 31 2012-03-15 14:55:00 31 2012-03-15 14:56:00 31 2012-03-15 14:57:00 31 2012-03-15 14:58:00 29 2012-03-15 14:59:00 30 > str(TICK.NYSE) An ‘xts’ object from 2011-01-18 09:30:00 to 2012-03-15 14:59:00 containing: Data: num [1:114090, 1:8] -5 -144 24 -148 -184 -77 20 121 111 -60 ... - attr(*, "dimnames")=List of 2 ..$ : NULL ..$ : chr [1:8] "TICK-NYSE.Open" "TICK-NYSE.High" "TICK-NYSE.Low" "TICK-NYSE.Close" ... Indexed by objects of class: [POSIXct,POSIXt] TZ: xts Attributes: List of 4 $ from : chr "20110119 23:59:59" $ to : chr "20110124 23:59:59" $ src : chr "IB" $ updated: POSIXct[1:1], format: "2012-01-19 02:34:52" > str(index(TICK.NYSE)) Class 'POSIXct' atomic [1:114090] 1.3e+09 1.3e+09 1.3e+09 1.3e+09 1.3e+09 ... ..- attr(*, "tzone")= chr "" ..- attr(*, "tclass")= chr [1:2] "POSIXct" "POSIXt" > Sys.getenv("TZ") [1] "EST" > tail(index(TICK.NYSE)) [1] "2012-03-15 14:54:00 EST" "2012-03-15 14:55:00 EST" [3] "2012-03-15 14:56:00 EST" "2012-03-15 14:57:00 EST" [5] "2012-03-15 14:58:00 EST" "2012-03-15 14:59:00 EST" > head(index(TICK.NYSE)) [1] "2011-01-18 09:30:00 EST" "2011-01-18 09:31:00 EST" [3] "2011-01-18 09:32:00 EST" "2011-01-18 09:33:00 EST" [5] "2011-01-18 09:34:00 EST" "2011-01-18 09:35:00 EST" > Sys.info() sysname "Linux" release "3.0.0-16-generic" version "#28-Ubuntu SMP Fri Jan 27 17:44:39 UTC 2012" > sessionInfo() R version 2.14.1 (2011-12-22) Platform: x86_64-pc-linux-gnu (64-bit) locale: [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C [3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8 [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 [7] LC_PAPER=en_US.UTF-8 LC_NAME=en_US.UTF-8 [9] LC_ADDRESS=en_US.UTF-8 LC_TELEPHONE=en_US.UTF-8 [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=en_US.UTF-8 attached base packages: [1] stats graphics grDevices utils datasets methods base other attached packages: [1] lattice_0.20-0 multicore_0.1-7 [3] doSNOW_1.0.5 snow_0.3-8 [5] doRedis_1.0.4 rredis_1.6.3 [7] foreach_1.3.2 codetools_0.2-8 [9] iterators_1.0.5 PerformanceAnalytics_1.0.3.3 [11] quantstrat_0.6.1 blotter_0.8.4 [13] twsInstrument_1.3-3 FinancialInstrument_0.10.6 [15] IBrokers_0.9-6 quantmod_0.3-18 [17] TTR_0.21-0 xts_0.8-4 [19] Defaults_1.1-1 strucchange_1.4-6 [21] sandwich_2.2-8 zoo_1.7-7 [23] rj_1.0.2-5 loaded via a namespace (and not attached): [1] grid_2.14.1 tools_2.14.1 > dput(tail(TICK.NYSE)) structure(c(-385, -213, -42, -334, -233, -111, -121, 20, -14, -125, -73, 265, -583, -269, -426, -520, -443, -440, -213, -42, -334, -233, -111, 119, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 31, 31, 30, 30, 31, 31), class = c("xts", "zoo" ), .indexCLASS = c("POSIXct", "POSIXt"), .indexTZ = "", from = "20110119 23:59:59", to = "20110124 23:59:59", src = "IB", updated = structure(1326958492.96405, class = c("POSIXct", "POSIXt")), index = structure(c(1331927640, 1331927700, 1331927760, 1331927820, 1331927880, 1331927940), tzone = "", tclass = c("POSIXct", "POSIXt")), .Dim = c(6L, 8L), .Dimnames = list(NULL, c("TICK-NYSE.Open", "TICK-NYSE.High", "TICK-NYSE.Low", "TICK-NYSE.Close", "TICK-NYSE.Volume", "TICK-NYSE.WAP", "TICK-NYSE.hasGaps", "TICK-NYSE.Count")))

    Read the article

  • How to change default boot order ubuntu 10.04 ?

    - by Sako Christian
    Hello, How can I change default boot order in Ubuntu 10.04 from Ubuntu to Windows7? However, I already checked sudo gedit /etc/default/grub and modify the grub file to be GRUB_DEFAULT=4 and update the grup sudo update-grub I even install graph software to re order the book sudo startupmanager But still after restart the default choose for boot is Ubuntu ... Thank you, Sako Christian P.S: I am using Ubuntu 10.04 with grub version 1.98

    Read the article

  • DNS Server Order Incorrect on Windows 7 via PPTP VPN to Windows 2003 Server

    - by Simon
    Hi there. When I connect a Windows XP laptop via PPTP vpn to our Windows 2003 Server, the DNS Server order is correct: 192.168.8.3 208.67.222.222 208.67.220.220 But when I connect a Windows 7 laptop via PPTP VPN to our Windows 2003 Server, the DNS order is incorrect: 208.67.222.222 208.67.220.220 192.168.8.3 What do I need to do on our Windows 2003 Server to fix this so the when I do a ping, it will work correctly?

    Read the article

  • Distributed Computing Framework (.NET) - Specifically for CPU Instensive operations

    - by StevenH
    I am currently researching the options that are available (both Open Source and Commercial) for developing a distributed application. "A distributed system consists of multiple autonomous computers that communicate through a computer network." Wikipedia The application is focused on distributing highly cpu intensive operations (as opposed to data intensive) so I'm sure MapReduce solutions don't fit the bill. Any framework that you can recommend ( + give a brief summary of any experience or comparison to other frameworks ) would be greatly appreciated. Thanks.

    Read the article

  • Composite operations in Android Canvas

    - by kayahr
    I'm just starting with Android development and I'm coming from JavaScript/HTML world so I'm currently investigating the possibilities of the Android SDK. The HTML 5 canvas supports composite operations (See here). Is this possible in an Android Canvas? I scanned the API of the Canvas class but couldn't find anything useful. I need at least the composite operation "source-in" or (if this isn't possible) "source-atop".

    Read the article

  • [VIM] How to restrict operations to certain lines?

    - by Ayman
    I have to work some relatively huge code files. How do I restrict some operations like find-next normal-n and others to a certain function / block? How would I visually know if I'm within that block or outside it? Looking and line numbers seems awkward, specially that the line numbers I need to work with are generally 5 digits long!

    Read the article

  • How to restrict operations to certain lines?

    - by Ayman
    I have to work on some relatively huge code files in vim. How do I restrict some operations like find-next normal-n and others to a certain function / block? How would I visually know if I'm within that block or outside it? Looking and line numbers seems awkward, specially that the line numbers I need to work with are generally 5 digits long!

    Read the article

  • Stack and queue operations on the same array.

    - by Passonate Learner
    Hi. I've been thinking about a program logic, but I cannot draw a conclusion to my problem. Here, I've implemented stack and queue operations to a fixed array. int A[1000]; int size=1000; int top; int front; int rear; bool StackIsEmpty() { return (top==0); } bool StackPush( int x ) { if ( top >= size ) return false; A[top++] = x; return true; } int StackTop( ) { return A[top-1]; } bool StackPop() { if ( top <= 0 ) return false; A[--top] = 0; return true; } bool QueueIsEmpty() { return (front==rear); } bool QueuePush( int x ) { if ( rear >= size ) return false; A[rear++] = x; return true; } int QueueFront( ) { return A[front]; } bool QueuePop() { if ( front >= rear ) return false; A[front++] = 0; return true; } It is presumed(or obvious) that the bottom of the stack and the front of the queue is pointing at the same location, and vice versa(top of the stack points the same location as rear of the queue). For example, integer 1 and 2 is inside an array in order of writing. And if I call StackPop(), the integer 2 will be popped out, and if I call QueuePop(), the integer 1 will be popped out. My problem is that I don't know what happens if I do both stack and queue operations on the same array. The example above is easy to work out, because there are only two values involved. But what if there are more than 2 values involved? For example, if I call StackPush(1); QueuePush(2); QueuePush(4); StackPop(); StackPush(5); QueuePop(); what values will be returned in the order of bottom(front) from the final array? I know that if I code a program, I would receive a quick answer. But the reason I'm asking this is because I want to hear a logical explanations from a human being, not a computer.

    Read the article

  • Numeric operations over SHA-1 generated keys in C#

    - by webdreamer
    I'm trying to implement a Chord distributed hash table. I want to use SHA-1 as the hash function to generate node ids and map values to the DHT. However, I'll need to use numerical operations on the SHA-1 generated key, such as a modulo, for example. I wonder in which type of variable should I put the array of bytes I get, and how can I convert from one to another.

    Read the article

  • Column-oriented DBMS and JOIN operations

    - by André
    From some of the research I've done on NoSQL, column-oriented databases (like HBase or Cassandra) seem to solve the problem of costly JOIN operations, but I don't get how this approach solves this problem. Can anyone explain it to me and/or link me to interesting documentation regarding this area? Thanks

    Read the article

facebook twitter digg google delicious technorati stumbleupon myspace wordpress linkedin gmail igoogle windows live tumblr viadeo yahoo buzz yahoo mail yahoo bookmarks favorites email print

< Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >