So6 Repository/Synchronizer corruption

RE : So6 Repository/Synchronizer corruption

Sebastien Jourdain — 2006-01-11T12:22:56+01:00

I see what you mean but at first we only thought at the case that I explained earlier...

We didn't know that OSX could generate special characters that wasn't supported on other platform...

Regards,

Artenum Team

RE : So6 Repository/Synchronizer corruption

Rüdiger Lincke — 2006-01-11T12:22:56+01:00

Thank you for the clarification.

I guess we were a bit too naiv in the beginning. I makes sence to have a policy in the team on how to handle this.

Some warning message could be useful anyway, or information about this situation in an official handbook.

RE : So6 Repository/Synchronizer corruption

Sebastien Jourdain — 2006-01-11T12:22:56+01:00

>how does So6 handle the queue content/synchronization of the workspaces, when workspaces from different operating systems and file systems commit/upate files and folders?

I fact we use the standard java command to manage the filesystem.

>The file systems of UNIX/Linux, MacOS and Windows are pretty different, when it comes to the handling of special characters and blanks in the path. I would expect it to be pretty difficult on the first thought, if one tries to synchronize a workspace over different file systems. How is this handles in So6?

In fact we don't realy care, I mean when you do things specific to a kind of computer, you may only share your data to the same kind of computer. Otherwise you have to define rules inside your team in order to prevent specific OS limitation. That's exactly what happen to you. You have the same problem when a Unix user put in the same directory 2 files with de same name but different case and want to share its data with a Windows user.

We didn't want to limit So6 in such case.

It's the same with the encoding, each system should use the same encoding in order to prevent encoding problem. (The one specified in the So6 queue: Bug fixed in the next release)

Regards,

Artenum Team

RE : So6 Repository/Synchronizer corruption

Rüdiger Lincke — 2006-01-11T12:22:56+01:00

Hello,

how does So6 handle the queue content/synchronization of the workspaces, when workspaces from different operating systems and file systems commit/upate files and folders?

The file systems of UNIX/Linux, MacOS and Windows are pretty different, when it comes to the handling of special characters and blanks in the path. I would expect it to be pretty difficult on the first thought, if one tries to synchronize a workspace over different file systems. How is this handles in So6?

Currently I have a problem with some group of users, using Linux, MacOS and Windows for sharing the workspace. It happend that the Mac user commited some folders containing characters which are not allowed in the Windows file system, therefore making the So6 Queue unusable for the other team members. They can not update or commit anymore, since the Eclipse plug-in reports an error, when they try it. I am not sure, if the so6 web client handles this situation better, but the plug-in has a serious problem.

Maybe this issue relates to the problems discussed earlier in this thread. Is this a bug and should be reported? Are there some guidelines or limitations we should consider.

Thank you,

Rüdiger

RE : So6 Repository/Synchronizer corruption

Sebastien Jourdain — 2006-01-11T12:22:56+01:00

You should add a the root of your workspace a file named ".ignore.so6" with a line to ignore the metadata directory.

:/.metadata$

Replace / by a back-slash…

RE : So6 Repository/Synchronizer corruption

Sebastien Jourdain — 2006-01-11T12:22:56+01:00

Well I don't realy see any reason… I will still think about it...

Regards

Artenum Team

RE : So6 Repository/Synchronizer corruption

Rüdiger Lincke — 2006-01-11T12:22:56+01:00

Hello, I have some more information about the problem.

It is now commited, that it only happens under Linux (Ubuntu) using the Eclipse plug-in. Under Windows XP and Eclipse Plug-in everything is fine. Using Linux and the So6 Web Client exclusively everything works fine as well. As soon as the Eclipse plug-in is used by the particular user under Linux to share the project the local workspace gets corrupted:

Log-file of the Web Client trying to update after corruption:

<html>Your local workspace is corrupted. <br>Please restore it</html> ### The local workspace is corrupted ### # org.libresource.so6.core.WsConnection: line 628 # org.libresource.so6.core.ui.BasicProgressView: line 105 # org.libresource.so6.core.exec.ui.Update: line 85 # org.libresource.so6.core.exec.ui.Update$2: line 139 # java.lang.Thread: line 595 Update finished

*** Report *** No Report available...

Log after commiting:

---
A TTT/linuxtestfolder/testclass.class
A .metadata/.plugins/org.eclipse.core.resources/.history/a0
A .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.core.resources.prefs
R Protocols
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.markers.snap
A .metadata/.plugins/org.eclipse.jdt.core
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes/ac/history.index
A .metadata/version.ini
A TTT/.classpath
R Artifacts
R SourceCode
A .metadata/.plugins/org.eclipse.core.resources/.root/.indexes
A .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.ui.prefs
A .metadata/.plugins/org.eclipse.core.resources/.history/6
A .metadata/.plugins/org.eclipse.jdt.core/3658435329.index
A .metadata/.plugins/org.eclipse.core.resources/.history/eb
A .metadata/.plugins/org.eclipse.core.resources/.history/6/f0623bf91d56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.history/1c/602e4f591e56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources
A .metadata/.plugins/org.eclipse.core.resources/.projects
A TTT/linuxtestfolder/testclass2.java
A .metadata/.plugins/org.eclipse.team.cvs.core/.running
A .metadata/.plugins/org.eclipse.jdt.ui
A .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.jdt.core.prefs
A .metadata/.plugins
A .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.jdt.launching.prefs
A .metadata/.plugins/org.eclipse.core.resources/.history/c4/e08406b51b56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.snap
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT
A TTT/.project#1
A TTT
A .metadata/.plugins/org.eclipse.core.runtime
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes/history.index
A TTT/linuxtestfolder/testclass3.java
A TTT/.classpath#2
A .metadata/.plugins/org.eclipse.jdt.core/externalLibsTimeStamps
A .metadata/.plugins/org.eclipse.team.cvs.core
A .metadata/.plugins/org.eclipse.debug.core
A TTT/linuxtestfolder/testclass3.class
A .metadata
A .metadata/.plugins/org.eclipse.core.resources/.history/ac/80c29dd41d56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.history/13/d099bedc1d56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes/8c
A .metadata/.plugins/org.eclipse.debug.ui
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/org.eclipse.jdt.core
A .metadata/.plugins/org.eclipse.core.resources/.history/bb/c02b181d1b56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.history/e/007b23f31d56001a169f8bc7ec02ec7c
A TTT/linuxtestfolder/testclass2.class
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes
A .metadata/.plugins/org.eclipse.core.resources/.history/1c
A .metadata/.plugins/org.eclipse.core.resources/.history/eb/f0e7bedc1d56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.history/bb
A .metadata/.lock
A .metadata/.plugins/org.eclipse.core.resources/.history
A TTT/linuxtestfolder/testclass.java
A TTT/SourceCode
A .metadata/.plugins/org.eclipse.core.resources/.safetable
A .metadata/.plugins/org.eclipse.core.resources/.history/c6
A .metadata/.plugins/org.eclipse.core.resources/.history/c6/90fdbddc1d56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.history/c4
A TTT/.project#1#10
A .metadata/.plugins/org.eclipse.core.resources/.history/ac
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes/ac
A .metadata/.plugins/org.eclipse.jdt.launching
A .metadata/.plugins/org.eclipse.jdt.core/variablesAndContainers.dat
A .metadata/.plugins/org.eclipse.core.resources/.root/.markers.snap
A .metadata/.plugins/org.eclipse.core.resources/.history/13
A .metadata/.plugins/org.eclipse.ui.workbench
A .metadata/.plugins/org.eclipse.core.resources/.history/a0/c072bedc1d56001a169f8bc7ec02ec7c
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes/8c/history.index
A .metadata/.plugins/org.eclipse.core.runtime/.settings
A .metadata/.plugins/org.eclipse.core.resources/.history/e
A .metadata/.plugins/org.eclipse.ui
A .metadata/.plugins/org.eclipse.core.resources/.root
A TTT/Protocols
A .metadata/.plugins/org.eclipse.jdt.core/7465893.index
A .metadata/.plugins/org.eclipse.team.cvs.ui
A TTT/Artifacts
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.syncinfo.snap
A .metadata/.plugins/org.eclipse.ui.ide
A .metadata/.plugins/org.eclipse.jdt.launching/libraryInfos.xml
A TTT/linuxtestfolder
A .metadata/.plugins/org.eclipse.core.resources/.projects/TTT/.indexes/properties.index
A .metadata/.plugins/org.eclipse.core.resources/.root/.indexes/properties.index
A .metadata/.plugins/org.eclipse.jdt.core/savedIndexNames.txt
A TTT/.project
A .metadata/.plugins/org.eclipse.core.resources/.safetable/org.eclipse.core.resources
---

The rights for the .so6 folder are all there, also is enough space on the hard disc/partition available.

The user agreed now to use the Web Client exclusively, which works fine as a work around.

We are checking now, if the Eclipse plug-in might have been corrupted. Also there is another user using the Eclipse plug-in successfully under the same Linux disctribution, so we will try to find the difference.

For the case you have any ideas, please let us know. Userwise we can continue working like this without too much problems.

Thank you, Rüdiger

RE : So6 Repository/Synchronizer corruption

Sebastien Jourdain — 2006-01-11T12:22:56+01:00

I've just saw them...

You're welcome…

RE : So6 Repository/Synchronizer corruption

Rüdiger Lincke — 2006-01-11T12:22:56+01:00

Hello, I uploaded some screen shots in this bug tracker: Possible So6 repository corruption?

I will try to get the log from the user using the web client.

Thank you for your help!

RE : So6 Repository/Synchronizer corruption

Sebastien Jourdain — 2006-01-11T12:22:56+01:00

When you make a commit with the Web client, you can view the log and see the exception during runtime…