BTW: this issue seems to be reappearance of issue 26444 (but on
another dataset).

I'm assuming this is an issue about the termination of the cvs
operation and not the cvs locks, given the reference to issue 26444. 
As such, investigating the history.

timeframe: update in 1 hour.

Brian,

I've found isolated command to reproduce:

cvs co -A -D "2003-10-06 15:34 +0000" core/registry/test/unit/src/META-INF

Opened an internal issue to investigate for corrupt files in one of
the repositories listed.  Please provide more detail when the
termination message happens, specifically the files right before the
termination.

timeframe: update in 4 hours.

Rudolf, I came across that as well and have provided it to our
internal engineers.  We're looking for a file in that vicinity which
is corrupt in some way or other.

Here's the output I got:
cvs server: Updating core/registry/test/unit/src/META-INF
U core/registry/test/unit/src/META-INF/MANIFEST.MF
Terminated with fatal signal 11
Core dumped; preserving /tmp/cvs-serv11434 on server.
CVS locks may need cleaning up.

Found what looks like the corrupt file.  Having it removed right now
and will test the checkout when it has been relocated.

We have removed an offending file.  It was the mainfest.mf file in the
attic for this project:

/cvs/core/registry/test/unit/src/META-INF/Attic:
-r--r--r--   1 tigrisc  tigris        79 Oct  6 06:44 manifest.mf,v

The checkout of the META-INF directory in the core project worked for
me after the file's relocation.  Please let me know if this resolves
the issue for you as well.

Brian,

it looks like it's fixed for now. We'll ask you again for support when
we'll decide how to handle history of these two files.

Would you for now please put the manifest.mf file back, but renamed?
Say "badmanifest.mf".

Jesse, you have committed the capital case filename version of the
manifest. Is that intentional? We need to handle history of these
files and I don't know core/registry module enough to tell which file
to keep (upper/lower case?).

Rudolf, I'm not positive the name is the issue (as the file is
corrupt) but I will ask for it to be restored as you have asked.

The file has been put back with the modified name.

MANIFEST.MF (all uppercase) is the correct filename. It should exist
in the trunk (*not* in Attic/) in one revision, 1.1, with contents:

---%<---
Manifest-Version: 1.0

Name: org/netbeans/core/registry/convertors/TestBeanConvertor.class
NetBeans-Convertor:
{http://core.netbeans.org/registry/test/ns}test-bean,
org.netbeans.core.registry.convertors.TestBean

---%<---

(Note that this is five lines of text: M-V, blank, Name, NB-C, blank.
The blank lines and positions of line breaks are significant.)

The current CVS repository contents looks correct to me, does the
problem still exist? I see no files in Attic, and the following
(apparently correct) result from cvs log:

---%<---
RCS file: /cvs/core/registry/test/unit/src/META-INF/MANIFEST.MF,v
Working file: MANIFEST.MF
head: 1.1
branch:
locks: strict
access list:
symbolic names:
	BLD200310060100: 1.1
	BLD200310030100: 1.1
	BLD200310020100: 1.1
	QBE200310010100: 1.1.0.2
	BLD200310010100: 1.1
	BLD200309300100: 1.1
keyword substitution: kv
total revisions: 1;	selected revisions: 1
description:
----------------------------
revision 1.1
date: 2003/09/29 23:09:12;  author: jglick;  state: Exp;
#36297: basic implementation of Convertors API within default Registry
API implementation. Good enough to permit clients to begin migration
to Convertors API.
---%<---

Whatever was wrong appears fixed now, reopen if you see any problems.

Brian,

I've forgot to ask you to test the cvs command again.

There's a regression and the file

  core/registry/test/unit/src/META-INF/manifest.mf
resp.
  core/registry/test/unit/src/META-INF/Attic/manifest.mf

should be removed from server's filesystem.

Yesterday I was commiting to core/registry to my private branch
registry_35050 file manifest.mf and got strange error:

More commits to come...
RCS file: /cvs/core/registry/test/unit/src/META-INF/Attic/manifest.mf,v
done
commit.c:2046: failed assertion `*rcsnode == NULL'
cvs [server aborted]: received abort signal
cvs commit: saving log message in C:\DOCUME~1\dk125166\LOCALS~1\Temp\19

I tried it again and got:

cvs server: cannot lock
`/cvs/core/registry/test/unit/src/META-INF/Attic/manifes
t.mf,v'.

My branch was created before the MANIFEST.MF was added to trunk. Was
the problem in different case?

David,

actually your file "manifest.mf" seems to be the corrupted file.


Reassigning to support to remove that file.

Updated the internal issue 22976 for removeing this file. Will update 
again once they rmoved the file.

Thanks.
Priya

Priya,

please can you more escalate this issue, because we can not make
continual NetBeans build until Collabnet remove corrupted file.

Thank you for your help
-Marek

Paged the Ops engineer to remove the file and escalted the internal 
issue. Will update soon here as soon as he removes it now.

Thanks
Priya

This file has been removed.

Thanks .
Priya

So, can I try to add file to my branch again? With capital letters
now? I do not need it urgently, I'm just asking.

Hi all,

the update which was done at 07:28:20 PT failed too. So it looks like 
that the issue is still not fixed.

BTW: This issue is urgent because it prevent us to create any build.

I am looking into this now.  I'm not sure how the MANIFEST.MF file was
removed, yesterday we removed the file manifest.mf,v from the attic
not the live repository.

timeframe: update today.

Can someone at Netbeans confirm/deny whether the checkout of the
'core' project (specifically
/cvs/core/registry/test/unit/src/META-INF/MAINFEST.MF) is working now?

I should specify that my testing was successful of a checkout of the
MANIFEST.MF file, but only after I deleted it from my working copy.

Looks OK to me so far...

Hi all, I can confirm that the file can be checked out, but when you 
try the command:
cvs co -A -D "2003-10-06 15:34 +0000" 
core/registry/test/unit/src/META-INF
than you'll get this message:
Terminated with fatal signal 11
Core dumped; preserving /tmp/cvs-serv1001 on server.
CVS locks may need cleaning up.
----
So I would like to ask you to fix it. This is still P1 for us.

So, current state is NOT FIXED. 

To Jesse: for all tests in this issue use timestamp, please

it is possible to checkout file
core/registry/test/unit/src/META-INF/MANIFEST.MF, but not the rest of
the directory.

$ rm -r core
$ cvs co -D "2003-10-07 01:00 +0000" core/registry/test/unit/src/META-INF
cvs server: Updating core/registry/test/unit/src/META-INF
U core/registry/test/unit/src/META-INF/MANIFEST.MF
Terminated with fatal signal 11
Core dumped; preserving /tmp/cvs-serv1530 on server.
CVS locks may need cleaning up.
$



Would someone at CollabNet finaly remove the file manifest.mf file as
I requested nearly 24 hours ago?

Look & remove these possible filenames
  core/registry/test/unit/src/META-INF/manifest.mf,v
resp.
  core/registry/test/unit/src/META-INF/Attic/manifest.mf,v

bash-2.05a$ ls -al /cvs/core/registry/test/unit/src/META-INF
total 32
drwxrwxr-x   3 tigrisc  tigris      4096 Oct  7 18:19 .
drwxrwxr-x   4 tigrisc  tigris      4096 Sep 29 11:08 ..
drwxrwxr-x   2 tigrisc  tigris      4096 Oct  6 11:23 Attic
-r--r--r--   1 tigrisc  tigris       731 Oct  7 18:19 MANIFEST.MF,v
bash-2.05a$ ls -al /cvs/core/registry/test/unit/src/META-INF/Attic
total 24
drwxrwxr-x   2 tigrisc  tigris      4096 Oct  6 11:23 .
drwxrwxr-x   3 tigrisc  tigris      4096 Oct  7 18:19 ..
-r--r--r--   1 tigrisc  tigris        79 Oct  6 11:23 badmanifest.mf,v

Engineer moved the badmanifest.mf,v out of the repository. Please 
check now again the problem is fixed.

Thanks,
Priya.

The core module seems to be fine now, the given checkout command
passed ok.
We need to wait 1 cycle of continuous build for final confirmation.
Will update in 1-2 hours.

We verified the fix. 

Verified

We recently moved out from Collabnet's infrastructure