Community
Participate
Working Groups
Using a pserver connection, right-clicking on a checked-in project and selecting "Check out as project" works, but "Check out as..." causes a hang and a timeout after about a minute. Internally, the latter case throws org.eclipse.team.internal.ccvs.core.connection.CVSCommunicationException (Socket Exception: Connection reset) Note: This also seems to apply when To reproduce: 1) Set up a new repository as :pserver:anonymous@dev.eclipse.org:/home/tools 2) Connect, and select org.eclipse.draw2d.examples 3) Select "Check out as..." from the context menu You'll need to run with debugging enabled and watch the console log to see the exception; Eclipse reports it with a "Socket Exception: Connection reset" dialog.
This works for me on Windows XP. I'll try and find a Windows 2000 box to see if it's related to the platform. Also, could you let me know your configuration information so I can test the same setup (i.e. Eclipse build, Java VM)
(Also in bug 38342) This is 100% repeatable for me, unfortunately. I'll go install on a local WinXP box and see if it happens there. Multiple operations exhibit the same behavior: creating a new module (bug 38342), updating from a changed project, committing changes. It happens whether I'm talking to the CVS server on my local network (1.11.5 on NetBSD 1.6), or to an outside server (e.g. dev.eclipse.org). The GNU CVS tools (WinCVS and cygwin) work under the same circumstances. Eclipse version 2.1.0, build 200303272130 (2.1 release) From the System properties: java.home=C:\j2sdk1.4.1_02\jre java.io.tmpdir=C:\DOCUME~1\rdclark\LOCALS~1\Temp\ java.runtime.name=Java(TM) 2 Runtime Environment, Standard Edition java.runtime.version=1.4.1_02-b06 java.version=1.4.1_02 java.vm.info=mixed mode java.vm.name=Java HotSpot(TM) Client VM java.vm.version=1.4.1_02-b06
I am unable to reproduce this on Win2K with the same Eclipse build and Java VM. What make of network card does your computer have?
*** Bug 38432 has been marked as a duplicate of this bug. ***
VIA Rhine II Fast Ethernet Adaptor (integrated into the motherboard, a SOYO Dragon Plus.) For the record, this is a Win2K SP3 machine.
We have had previous bug reports about problems occuring in Eclipse CVS with this card (See bug 27077). We'll leave this bug open to investigate further.
We think we may have solved this problem in 3.0. Could you try using the latst 3.0 integration build (I20030528 or later) to see if the problem still occurs?
I installed a Linksys ethernet card and the bug went away. I recommend marking this as a duplicate of 27077, and I'll try the 3.0 integration build w/ the old ethernet controller. Could I ask this be added to the CVS team FAQ?
I have a related error. When I try to select an entire project (~5000 files) and merge from a branch to the trunk, or vice versa, I'll get this error (Socket Exception: Connection Reset) ... I'm on an IBM Thinkpad T40, with an Intel(r) Pro/1000 MT Mobile Connection NIC. I'm running Eclipse 2.1.1, against a CVSNT server (unsupported, I know). Multiple posts to the CVSNT & CVS mailing lists/newsgroups have indicated that this is not something that is a CVSNT/CVS specific problem - specifically others have posted that they've seen this error against main-line CVS on *nix boxes. It seems to have something to do with (a) the size of the project and (b) how deeply folders are nested. At least this has been my experience. Does it seem like this is something that might be fixed in 3.0? Note: This happens when Eclipse is interrogating CVS as to which files have changes that need to be merged, not after that, when its updating/overriding&updating, etc.
I have the same problem with a laptop (windows XP), with a Realtek RTL8139C network card. Each time I make a cvs sync, or replace with branch, I get a "Sockect Exception: Connection reset" or a "I/O has been interrupted" error, and then only way is to turn off the PC. I am using Eclipse 2.1.3
I have changed the connection type to "extssh" and now it works perfectly. So the bug is only with "pserver" type of connection. At least that's allow me to work normally.
*** Bug 37204 has been marked as a duplicate of this bug. ***
Although I have not been able to reproduce this, I have found and fixed a likely cause. The fix will be in 3.0 M9 next week in case anyone who has experienced the problem wants to verify the fix.
I tested 3.0RC2 Great! No more BSODs! Thanks a LOT! I suggest to close bug 35891 as a duplicate of this one.
