3

I'm using the latest versions of Eclipse (4.2.1) and EGit (2.3.1) to push code to git repositories that are hosted at Assembla. Lately I start getting those messages: "Invalid channel 117". The push seems to go trough, so it doesn't really hurt, but it's annoying because I cannot see the result of the push. Any ideas? Thanks!

Error Log:

An exception occurred during push on URI git@git.assembla.com:cool_awesome_app.git: Invalid channel 117
org.eclipse.jgit.api.errors.TransportException: Invalid channel 117
at org.eclipse.jgit.api.PushCommand.call(PushCommand.java:157)
at org.eclipse.egit.core.op.PushOperation.run(PushOperation.java:215)
at org.eclipse.egit.ui.internal.push.PushOperationUI.execute(PushOperationUI.java:147)
at org.eclipse.egit.ui.internal.push.PushOperationUI$1.run(PushOperationUI.java:209)
at org.eclipse.core.internal.jobs.Worker.run(Worker.java:53)
Caused by: org.eclipse.jgit.errors.PackProtocolException: Invalid channel 117
at org.eclipse.jgit.transport.SideBandInputStream.needDataPacket(SideBandInputStream.java:175)
at org.eclipse.jgit.transport.SideBandInputStream.read(SideBandInputStream.java:136)
at org.eclipse.jgit.util.IO.readFully(IO.java:246)
at org.eclipse.jgit.transport.PacketLineIn.readLength(PacketLineIn.java:186)
at org.eclipse.jgit.transport.PacketLineIn.readString(PacketLineIn.java:138)
at org.eclipse.jgit.transport.BasePackPushConnection.readStringLongTimeout(BasePackPushConnection.java:352)
at org.eclipse.jgit.transport.BasePackPushConnection.readStatusReport(BasePackPushConnection.java:293)
at org.eclipse.jgit.transport.BasePackPushConnection.doPush(BasePackPushConnection.java:187)
at org.eclipse.jgit.transport.BasePackPushConnection.push(BasePackPushConnection.java:142)
at org.eclipse.jgit.transport.PushProcess.execute(PushProcess.java:141)
at org.eclipse.jgit.transport.Transport.push(Transport.java:1162)
at org.eclipse.jgit.api.PushCommand.call(PushCommand.java:153)
... 4 more

eclipse.buildId=M20130204-1200
java.version=1.6.0_43
java.vendor=Apple Inc.
Klemens Zleptnig
  • 1,698
  • 20
  • 36
  • Is there a stack trace somewhere, e.g. in the Error Log view? By the way, the version of EGit is probably 2.3.1, not 2.1.3, right? – robinst Apr 14 '13 at 12:24
  • I'm having the very same problem with EGit version 2.2.0.201212191850-r. – Gilberto Torrezan Apr 15 '13 at 19:01
  • I wonder if this happens with an earlier version of egit? Seems like it does not like the negotiated channel, that is why you are not getting the results displayed. – Michael Apr 15 '13 at 19:13
  • @robinst 2.3.1 indeed, thanks - I corrected it and added the stack trace. – Klemens Zleptnig Apr 16 '13 at 13:19
  • Could you [file a bug against JGit](https://bugs.eclipse.org/bugs/enter_bug.cgi?product=JGit&rep_platform=All&op_sys=All) with the stack trace? It would be nice if you could also provide a link to the repository (don't know if it's public or not). Also please include a link to the following, which seems related: https://github.com/jelmer/dulwich/pull/46 – robinst Apr 16 '13 at 13:43
  • Just tested - I get the same error with Egit 2.1.0 and Eclipse Helios 3.6.2. – Klemens Zleptnig Apr 16 '13 at 14:06
  • I have the same exact issue with Assembla using Egit 2.3.1; org.eclipse.jgit.errors.PackProtocolException: Invalid channel 117 at org.eclipse.jgit.transport.SideBandInputStream.needDataPacket(SideBandInputStream.java:175) – John M Apr 16 '13 at 16:12
  • I filed a bug. The repository is not public, sorry. But it's a standard git Repository on Assembla. – Klemens Zleptnig Apr 16 '13 at 17:34
  • 4
    Link to bug if anyone else wants to follow its progress: https://bugs.eclipse.org/bugs/show_bug.cgi?id=405809 – robinst Apr 17 '13 at 13:04
  • Today the error message disappeared (with one project) and it seems to work now, I haven't updated anything. – Klemens Zleptnig May 31 '13 at 21:09

1 Answers1

2

I work for Assembla, this was a problem in the way we were handling sideband communication channels - it should be fixed. Oddly only jgit/egit complained about this and not the c version of git.

Please take a look again.

Michael
  • 10,124
  • 1
  • 34
  • 49