Bugzilla – Full Text Bug Listing
| Summary: | [Sync Info] Network file systems cause out-of-sync files | ||
|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Steve Northover <snorthov> |
| Component: | CVS | Assignee: | platform-cvs-inbox <platform-cvs-inbox> |
| Status: | RESOLVED WONTFIX | QA Contact: | |
| Severity: | normal | ||
| Priority: | P5 | CC: | bazzoola, Mike_Wilson |
| Version: | 3.0 | Keywords: | helpwanted |
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| Whiteboard: | |||
|
Description
Steve Northover
The reason it doesn't happen on Windows or Mac is because they have native libs for file system notifications. On Linux you're SOOL :) We could always provide an option to refresh before synchronizing. ... or you could fix the bug? Why does it think I changed random files when I didn't. Is it a file system bug/feature? You didn't say that you didn't modify any files outside of Eclipse. It would help if you gave me more details. Are the files it complains about always the same? What are they? We've got at least 4 developers using GTK daily and haven't heard of this happening before. In any case, this probably has nothing to do with CVS because this exception is generated by core.resources. They hold on to the bit that tells them if a file has been modified outside of Eclipse. Come bother me when this happens again... I'll have a look. Sorry, I was "always need to refresh for no reason" is not a good title. I have changed it. There doesn't seem to be a pattern but I fix it by refreshing and then it goes bad about 1/2 hour later when I try and synchronize. It's probably an NFS thing with Head Rat? IT is happening right now to my GTK box. I think we need to know the answer to this before M9. We can: 1) see if the problem happens on the same hardware for 2.1.x 2) see if the problem goes away if I don't run off the network Any other ideas? Strange that I am the only person in the universe that sees this. Q: Is anyone else seeing this? Yes. er... so is there a pattern? is this network installs only? novell installs only? etc... Felipe runs over the network and does not see the problem. Post 3.0 I don't see this problem when running locally. Are you still seeing this Steve? I stopped running on the network on Linux in disgust. Silenio started seeing it too and also stopped using the network, then upgraded from RH9 to Fedora 5 and it went away. Our usual practice is to move bugs to the RESOLVED REMIND state when there is not enough information provided to address the reported issue and more information has been requested. This bug has been in the RESOLVED REMIND state for quite some time so it is being CLOSED. If you can provide the requested information or you otherwise feel that this is still an issue, please reopen the bug. If you are unable to open it due to a lack or permissions, please add a comment to the bug stating that you would like to have the request reconsidered. I applogize for this blanket approach to bug screening but there are just to many bugs in the REMIND state for us to effectively triage them. Your help in this matter is appreciated. I gave up. Want me to try again? If your willing to try again, that would be great. As Jean-Michel pointed out in an earlier comment, this looks like a network filesystem issue that is throwing off the timestamp comparison in Eclipse. There is a chance that there has been a Linux kernel improvement that has fixed the problem but I suspect that the problem is still there. I'll reopen this bug since I suspect it is still a problem. *** Bug 173951 has been marked as a duplicate of this bug. *** as reported in bug [https://bugs.eclipse.org/bugs/show_bug.cgi?id=173951] I do not see this issue on all the files. I have noticed that if i check out the project using SMB on Windows and and then try mount the project using SMB on Mac the problem doesn't appear. However, if I use sshfs on Mac then some files will show +/- 1 hour difference. I can provide more info if needed. I gave up long ago. WONTFIX. |