radiaku / csharptest-net Goto Github PK

As above really.

Just tried your sample code with 

var fcounter = new CSharpTest.Net.IO.FindFile(@"F:\", "*.docx", true, true, 
true);

and

var fcounter = new CSharpTest.Net.IO.FindFile(@"F:\", "*.cs", true, true, true);

The former returned the 2 that were in the root but there are 23 according to 
"dir *.docx /s"

The latter returned 0 files when there are 53k+!

Looks like this code is returning without checking any other folders....
(after call to FindFirstFileEx)
            if ((IntPtr.Size == 4 && hFile.ToInt32() == -1) ||
                (IntPtr.Size == 8 && hFile.ToInt64() == -1L))
            {
                Win32Error(Marshal.GetLastWin32Error());
                return;
            }

Just tried commenting it out and I'm getting the same results as dir /s

What steps will reproduce the problem?
1. install tortoiseSVN
2. install svnplugininstall.msi
3. configure svn repo properties as follows:

bugtraq:append = false 
bugtraq:label = Defect {PREFIX}- 
bugtraq:logregex = *
bugtraq:message = {PREFIX}-%BUGID% 
bugtraq:number = true 
bugtraq:warnifnoissue = true
bugtraq:url = http://jira:7070/browse/{PREFIX}-%BUGID% 
jira:url = http://jira:7070 

What is the expected output? What do you see instead?

I expected to see a "JIRA Issues" button to select issues from JIRA.  I 
only have a text box to enter issue id's.

What version of the product are you using? On what operating system?

JIRA 4.0.1
SvnPluginInstall 1.9.1016.145
TortoiseSVN 1.6.8


Please provide any additional information below.

This was a very simple installation.  I'm surprised that there is no 
screen shot of the issue selection button.

What:
Add IDL definition of RPC compatible with CSharpTest.
Value:
Entry point to write native client or server compatible with managed version.

Got for _pinnedAddresses in concurent scenatio:
System.InvalidOperationException occurred
  Message=Collection was modified; enumeration operation may not execute.
  Source=mscorlib
  StackTrace:
       at System.ThrowHelper.ThrowInvalidOperationException(ExceptionResource resource)
       at System.Collections.Generic.List`1.Enumerator.MoveNextRare()
       at System.Collections.Generic.List`1.Enumerator.MoveNext()
       at CSharpTest.Net.RpcLibrary.Interop.RpcHandle.GetPtr[T](T& value) in c:\ci\co\CommonOffice\src\csharptest-net\src\RpcLibrary\Interop\RpcHandle.cs:line 42
  InnerException: 

   internal bool GetPtr<T>(out T value)
        {
            foreach (object o in _pinnedAddresses)
            {
                if (o is T)
                {
                    value = (T) o;
                    return true;
                }
            }
            value = default(T);
            return false;
        }

Also got:
RpcInterop: ServerRegisterAuthInfo - unable to register authentication type 
RPC_C_AUTHN_NONE

System.NullReferenceException : Object reference not set to an instance of an 
object.
   at CSharpTest.Net.RpcLibrary.Interop.RpcHandle.Dispose(Boolean disposing) in RpcHandle.cs: line 84
   at CSharpTest.Net.RpcLibrary.Interop.RpcHandle.Dispose() in RpcHandle.cs: line 68
   at CSharpTest.Net.RpcLibrary.RpcClientApi.Dispose() in RpcClientApi.cs: line 56

Current implementation lacks of features and does not provide good way of 
developing them.
From my point of view it could be good
1. Move RpcLibrary into other repository specially for it.
2. Put in the same rep as Protbuf wrapper. So my WCF wrapper clone will go here 
either.
3. Make main interop stucts public and move pinvokes into NativeMethods. Is 
needed to start developing IDL generation of C# code.
4. http://code.google.com/p/csharptest-net/issues/detail?id=22
Thanks

Hi,

I have download your source code and want to start debugging your UI on my 
visual studio, but it keep complaining that i am missing a bunch of things.

Do you know how can i debug without tortoisesvn? how do you setup all those 
svn properties within the debug mode so I can just run the JIRASVN pop up?

thanks

There is a possibility that CallLevel lock would never be released. 
And this is really happening if LockTimeout expires in NodeCache.


Here is the code in the constructor of RootLock:
   _locked = _exclusive ? _tree._selfLock.TryWrite(tree._options.LockTimeout) : _tree._selfLock.TryRead(tree._options.LockTimeout);
   Assert(_locked);
   Pin = _tree._storage.LockRoot(type);


If an exception is thrown in '_tree._storage.LockRoot(type)' then 
'_tree._selfLock' will never be released. 
So the code should be enclosed by try..catch:

try
{
    Pin = _tree._storage.LockRoot(type);
}
catch
{
    if (_exclusive)
       _tree._selfLock.ReleaseWrite();
    else
       _tree._selfLock.ReleaseRead();
    throw;
}

Given I call dead server
Actual
Then got unknown SEH exception
Expected
Then got error with meaning
And can differentiate dead sever from other issues


Can use [comm_status]  to get C structure instead of C++ exception.

What steps will reproduce the problem?
1. Run the attached file

When the stream reads the data is the wrong length so de-serialization fails.  
This is using protobuf-net 2.0.0.470

Please submit any issues targeting Jira integration to the following location:

https://github.com/csharptest/JiraSVN/issues

What steps will reproduce the problem?
1. configure jira:url or parameters using a view-only account
2. open JiraSvn window in commit dialog, select any issue
3. This is the exception:

************** Exception Text **************
System.NullReferenceException: Object reference not set to an instance of 
an object.
   at CSharpTest.Net.SvnJiraIntegration.JiraConnection.GetActions
(JiraIssue issue) in D:\csharptestnet\src\Tools\SvnPlugin
\JiraConnection.cs:line 199
   at CSharpTest.Net.SvnJiraIntegration.JiraIssue.GetActions() in D:
\csharptestnet\src\Tools\SvnPlugin\JiraIssue.cs:line 80
   at CSharpTest.Net.SvnPlugin.UI.IssueItemView.GetActions() in D:
\csharptestnet\src\Tools\SvnPlugin\UI\IssueItemView.cs:line 95
   at CSharpTest.Net.SvnPlugin.UI.IssuesListView.RebuildActions() in D:
\csharptestnet\src\Tools\SvnPlugin\UI\IssuesListView.cs:line 356
   at CSharpTest.Net.SvnPlugin.UI.IssuesList.listView_ItemChecked(Object 
sender, ItemCheckedEventArgs e) in D:\csharptestnet\src\Tools\SvnPlugin\UI
\IssuesList.cs:line 129
   at System.Windows.Forms.ListView.WmReflectNotify(Message& m)
   at System.Windows.Forms.ListView.WndProc(Message& m)
   at System.Windows.Forms.Control.ControlNativeWindow.WndProc(Message& m)
   at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, 
IntPtr wparam, IntPtr lparam)
************************

This happens when loggin in on our Jira with a guest account; that account 
only has permission to view all projects and issues, but cannot do any 
modifications at all.

What steps will reproduce the problem?

from NuGet console install Google.ProtocolBuffers.Rpc

PM> Install-Package Google.ProtocolBuffers.Rpc
Attempting to resolve dependency 'CSharpTest.Net.RpcLibrary (≥ 1.11.924.348)'.
Attempting to resolve dependency 'Google.ProtocolBuffers (≥ 2.4.1.473)'.
Successfully installed 'CSharpTest.Net.RpcLibrary 1.11.924.348'.
Successfully installed 'Google.ProtocolBuffers.Rpc 1.12.919.16'.
Successfully uninstalled 'CSharpTest.Net.RpcLibrary 1.11.924.348'.
Install failed. Rolling back...
Install-Package : Could not install package 'CSharpTest.Net.RpcLibrary 
1.11.924.348'. You are trying to install this package into a project that 
targets 
'WindowsPhone,Version=v8.0', but the package does not contain any assembly 
references or content files that are compatible with that framework. For more 
information, contact the 
package author.
At line:1 char:1
+ Install-Package Google.ProtocolBuffers.Rpc
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : NotSpecified: (:) [Install-Package], InvalidOperationException
    + FullyQualifiedErrorId : NuGetCmdletUnhandledException,NuGet.PowerShell.Commands.InstallPackageCommand


What is the expected output? What do you see instead?
I need install Google.ProtocolBuffers.Rpc

What version of the product are you using? On what operating system?
OS Windows 8 64 bit
IDE MVS 2012
WP8 SDK

Please provide any additional information below.

Reported by Alex:

I using BPlusTree in my own project for storing data and its performance is 
great! But when my data size has comed nearer to 4 Gb the file is stopped to 
open. The problem is that TransactedCompoundFile has overflow error in GetFree 
method. Look at next line (931)

long position = BlocksPerSection*BlockSize*block.Section;

Even declaring left size of statement as long, don't casts right size to the 
same type. All right-hand side orerands has type of Integer, and overflow may 
occur in that place. So, you must rewrite that line to

long position = (long)BlocksPerSection*BlockSize*block.Section;
or
long position = 1L*BlocksPerSection*BlockSize*block.Section;

Add callback implementation sample like WCF callbacks
Value:
WCF users can migrate to RPC easier

This is not really a bug in the code I guess, but I'd really like to get 
it fixed. This happens when trying to access our server via https, which 
is the only way to get in from 'outside'.
The server itself is jira standalone, and proxied over apache2. Both 
tomcat and apache2 servers use SSL, but with self-signed certificates. 

When accessing jira with a browser like IE8 or Firefox you get a dialog 
box in which you can choose to go to the page, I understand this might be 
hard to do in the plugin since all communication seems to go via jira soap 
services, but I'd like to know if there is a way around? Can I tell the 
service connector somehow to trust the server's certificate, or any 
certificate? Is there a log file to investigate to finf out what exactly 
goes wrong?

What steps will reproduce the problem?
1. install TortoiseSVN 64bit client
2. install JiraSVN bia installer
3. check that this does not work: issue tracking plugin not listed

I guess this is because the provided installer is 32bit and this does not 
work with 64bit versions. This should at least be documented, yet better 
would be to provide a 64bit version of the plugin as well.

What steps will reproduce the problem?

I am trying to access the same bplustree file from multiple processes for read 
operation and getting the error "System.IO.IOException: The process cannot 
access the file 'C:\ProgramData\IEX365\tree.dat' because it is being used by 
another process.'.

What is the expected output? What do you see instead?

I want to be able to read the same data from multiple process without such 
exception.

What version of the product are you using? On what operating system?
version 2.14.126.467 on Windows 7 64bit.

Please provide any additional information below.

What steps will reproduce the problem?
1. TortoiseSVN -> Commit Files, and then commit windows show up and click 
on JIRA Issues button
2. On "Select Issue(s) window, choose the JIRA Issue number (Id) and then 
click OK
3. Trying to commit, and there is a yellow pop up show up from the Jira 
Issues button saying "Only numbers (optionally separated by commas) are 
allowed!"

What is the expected output? What do you see instead?

It suppose to just commit the changes, 

I cant commit, it show Only numbers (optionally separated by commas) are 
allowed!"

What version of the product are you using? On what operating system?

Windows XP, TortoiseSVN 

Please provide any additional information below.

What steps will reproduce the problem?
1. TortoiseSVN -> Commit Files, and then commit windows show up and click 
on JIRA Issues button
2. On "Select Issue(s) window, choose the JIRA Issue number (Id) and then 
click OK
3. Trying to commit, and there is a yellow pop up show up from the Jira 
Issues button saying "Only numbers (optionally separated by commas) are 
allowed!"

What is the expected output? What do you see instead?

Commited file :)
I cant commit, it show Only numbers (optionally separated by commas) are 
allowed!"


What version of the product are you using? On what operating system?
1.10.523
WIN7 64bit



Please provide any additional information below.
in newest tortoise 
ortoiseSVN 1.6.16, Build 21511 - 64 Bit , 2011/06/01 19:00:35
bug still exists - changing bugtraq:number = false won't help
here are my svnproperties:
    bugtraq:append = false
    bugtraq:label = Defect {PREFIX}-
    bugtraq:message = {PREFIX}-%BUGID%
    bugtraq:number = false
    bugtraq:url = https://xxx.xxx.xxx.xxx:8433/browse/{PREFIX}-%BUGID%

Use namespace to turn on custom tracers (not default static Trace).

http://msdn.microsoft.com/en-us/library/ms733025.aspx
http://blogs.msdn.com/b/mikehillberg/archive/2006/09/14/wpftracesources.aspx

I have encountered some weird deadlock condition: a deadlock through the Thread 
Pool. 
This can occur when the maximum number of threads in the pool is limited. 
The root of the problem is the asynchronous flushing in StorageCache class.


How to reproduce this problem:
First of all there should be limited number of threads in ThreadPool. It could 
be achieved by calling SetMaxThreads in standard ThreadPool or by using custom 
thread pool implementations (my case). 

Suppose we have 10 working threads.
All threads are busy. They perform some operations and sometimes call methods 
on BPlusTree. 
One of them performs the operation that invokes 'Update' in StorageCache with 
this line of code:
    this._asyncWriteBehind = this._writeBehindFunc.BeginInvoke(null, null);

But at this point we have no free working threads in the pool, so the job is 
queued.

Then one thread executes TryUpdate on BPlusTree. The writer lock is acquired 
there. 
At the same time 9 other threads trying to perform any other operations on 
BPlusTree and become locked. 
The update process goes to the final stage where it executes 'Commit' in 
StorageCache, that in turn executes 'CompleteAsync' with this code:
    this._writeBehindFunc.EndInvoke(asyncWriteBehind);

But the delegate is still inside the queue, so this thread is also locked.
And we have the situation when all 10 threads are locked.


I see 3 possible solution to this problem:
1. The thread, that performs 'CompleteAsync' should detect such situation (by 
checking some flag) and execute the flushing process in synchronous way. This 
is the best solution.
2. Increase the number of threads in ThreadPool, which reduce the probability 
of this situation. But I don't like this way.
3. Create separate thread inside BPlusTree which will perform flushing process.


I've described only the situation that has been reproduced in my application.
There can be other places in the BPlusTree with the same pattern and the same 
problem.

error: HTTP 407 proxy authentication error when calling a web service 

Solved^
In Tools\SvnPlugin\Web References\Jira\Reference.cs

added
    using System;
    using System.Xml.Serialization;
+    using System.Net;

        public JiraSoapServiceService() {
            this.Url = "http://jira.atlassian.com/jira/rpc/soap/jirasoapservice-v2";
+       this.Proxy = new WebProxy();
+            this.Proxy.Credentials = CredentialCache.DefaultCredentials;

What steps will reproduce the problem?
1. TortoiseSVN -> Commit...
2. Select "Jira Issues" Button
3. Select Issue
4. Select  "Set" and choose "Resolve Issue"
5. Enter comment, hit "ok" at the issue window and at the commit window

What is the expected output? What do you see instead?

The Issue schould be resolved now, but only the comment appears at JIRA.

What version of the product are you using? On what operating system?
JIRA 3.1.3 Pro, running on Linux
JIRA-Plugin 1.9.1016.145, running on WinXP


Please provide any additional information below.

"Parameters" at Issue Tracker Integration:
https://issues.company.com/jira

I am not sure, but do you code uses what handles?

http://msdn.microsoft.com/en-us/library/windows/desktop/aa373578.aspx
Do not use implicit handles or auto handles.

http://www.codeproject.com/Articles/4878/Introduction-to-RPC-Part-2#Multithreadi
ng11
Multithreading

The example showed in this article is not thread-safe. If two threads does 
things in the server at the same time, things could start behaving strange or 
even crash. In this simple example it does not matter, but it could easily be 
fixed by using some sort of mutual exclusive synchronization objects (critical 
sections). This will be addressed in a future article.

What steps will reproduce the problem?

1. Click the "Jira Issues" button to open the Issue selection dialog
2. Select an Issue from the list and enter some Comments "My comment"
3. Click OK
4. Update the Comments in the standard Tortoise Commit dialog, "My slightly 
better comment"
5. Click OK to Commit

What is the expected output? What do you see instead?

When looking at the issue in JIRA only the comment entered in the Issue 
selection dialog is displayed.  Any Comments or changes added after Issue 
selection are lost.


What version of the product are you using? On what operating system?
Windows XP Pro SP3

We have many JIRA filters, the most useful of which (i.e. the only one I would 
use with this plugin) is near the middle of the list.  For usability the plugin 
should pre-select the most recent filter.

I have a JIRA project with the key "FTPP". The label for the JIRA issue 
field hence reads "Defect FTPP-" which is good.

I then click the "Jira Issues" button where I see all the issues in the 
FTPP project. If I select one or more of these, however, what is returned 
is only one of the issues and in the full format, i.e. the field gets the 
value "FTPP-2" for instance which generates an error if I try to commit the 
change.

Apparently is the contents of that field not used; I have tried to enter 
"1,2" to update FTPP-1 and FTPP-2 respectively but that did not work. When 
I selected those issues in the "Jira Issues" screen they were properly 
updated, even if the field in question was empty.

All in all very confusing and I can not seem to have the top field getting 
a value that lets me commit change without an error. I have to clear it 
manually every time.

I am probably just being daft but I am stumped as to what I am doing wrong.

Current wrapper API stops all servers, need a method to stop listening one, 
still using others. Default API looks like instance class with method, by 
really some are statics....

When I close or dispose sever API I kill all host. But I have instance and want 
to close only it. Current stop listening should be internal and called on 
Process.Exit.
I spent 2 days of debugging with current API.

Steps to reproduce the problem:
1. Open the SVN commit dialog
2. Click the Jira button
3. Select an issue, set it to fixed.
4. Click OK
5. Edit the commit message (just the part before the new parts inserted by
SVNPlugIn

What is the expected output? 
I would expect the message I see in the svn commit dialog to be added to
the jira issue. It should be the same as added to the svn log, e.g. <final
svn commit message> Complete progress BUG-ID : Summary.

What do you see instead?
The message added to the jira issue uses the text that was in the svn
commit dialog before the plug in dialog was opened. The jira issue comment
is <original text before opening plug in window> revision: xxxx

What version of the product are you using?
1.9.1016.145

This is a great plugin, but is there a way to get it working with the 64bit
version?

Add async RPC support like documented on MSDN.

I installed the TortoiseSVN Plugin in 3.13.5 and everything worked fine
when clicking on the “Jira Issue” button.  However, after upgrading to Jira
4.0; clicking on the “Jira Issue” button throws the following exception. 
Any ideas, or is there a new version of the SvnPlugin coming out soon?

Our SVN server uses a common pre-commit hook to enforce a valid JIRA id at the 
*beginning* of a message.  My corresponding property is:
svn propset bugtraq:regex "^([A-Z]+-\d+)" .

I am struggling to get JiraSVN to present a legally formatted message in this 
environment?  Can we supply another control string/regex for this.  I suggest 
something in keeping with bugtraq:message but with more parameters for your 
message components:
svn propset jira:fullmessage "%BUGID% %COMMENT%\n%JIRASVNCOMMENT%" .

If the CallLevel lock cannot be acquired due to timeout then the Assertion is 
triggered. In this case it would be better to throw some specific exception 
derived from TimeoutException, so the user code can react appropriately. 

There are a couple of lines in the constructor of RootLock:
   _locked = _exclusive ? _tree._selfLock.TryWrite(tree._options.LockTimeout) : _tree._selfLock.TryRead(tree._options.LockTimeout);
   Assert(_locked);

Instead I suggest to do this:
   _locked = _exclusive ? _tree._selfLock.TryWrite(tree._options.LockTimeout) : _tree._selfLock.TryRead(tree._options.LockTimeout);
   if (!_locked)
      throw new BPlusTreeTimeoutException("...");

Struggling some issue related to concurrent creation and shutdowns of 
client/servers.

 got SEHException  from here during start of our system, trying to make tests which are simila (our system is 2 processes with several clients and servers in each)
                 try
                    {
                        result = NdrClientCall2x64(pStub.Handle, RpcApi.FUNC_FORMAT_PTR.Handle, handle.Handle,
                                                   input.Length,
                                                   pInputBuffer.Handle, out szResponse, out response);
                    }
                    catch (SEHException ex)
                    {
                        Log.Error(ex);
                        RpcException.Assert(ex.ErrorCode);
                        throw;

Issue:
Race condition in TransactionLog.cs, TransactionLog's c'tor, between the 
`File.Exists(...)` call and the `new FileInfo(_options.FileName).Length` 
statement. A file can be deleted in between leading to an exception later.

What steps will reproduce the problem?
1. Set V2 options with "CreateFile = CreatePolicy.Always"
2. Add via something like (note the `using` usage)

    public void Add(MyEntity newItem)
    {            
        // Open the BPlusTree using these options
        using (var bpt = new BPlusTree<string, KeyEntity>(_options))
        {
            bpt.Add(newItem.key, newItem.Value);
        }
     }

3. Add about 100K entities in rapid fire via the Add(MyEntity newItem) call.

There is a 1% error chance (on a high performance SSD) so by the 100th of so 
addition, you'll hit the race condition.

What is the expected output? What do you see instead?
- See above

What version of the product are you using? On what operating system?
- Tip of HG repo

Please provide any additional information below.

- Not specific to this issue but at a project level, it would be nice if you 
had a forum or something where user's of this library could talk about 
uses/features/bugs. For example: A cool idea would be extending the storage to 
use Azure Blob Storage (in Page mode). Likely extending/replicating off the 
`BTreeFileStoreV2` class. A forum could facilitate this sort of a discussion. 
Or perhaps having a multi master write model (extending on that azure thought).

What steps will reproduce the problem?
1. SVN->Commit
2. Press the Jira Issues Button and select two or more issues/bugs.
3.

What is the expected output? 

Issue/Bug text box in upper right hard of commit form should list all issue or 
bug id numbers of the selected issues. The number should be seperated by 
commas. 

Example:

id(s) selected on Slect Issues Form:
MFP-312
MFP-576
MFP-954

Text Box on Commit Form
312,576,954 or

MFP-312,MFP-576,MFP-954

What do you see instead?

Text Box on Commit Form shows only one issue id.

MFP-954

What version of the product are you using? On what operating system?

Current version as of 23 August 2010 (32 bit)
Win 7
Tortoise SVN 1.6.10

Please provide any additional information below.

Problem
- It looks like the code will try to fetch the list of all users which can be 
really huge list especially in a corporate environment (sometimes 2000+)
- It also tries to fetch all issues by using getIssuesFromTextSearch("") by 
default which can be really slow as well.

Action
- It should find out which projects the current user has right to browse via 
getProjectsNoSchemes
- It then probably should find out who are allowed to access those projects
- It should then have some sort of caching to cache users/project list so 
that it won't try to retrieve it over and over again.