This morning I upgraded our internal server to 2.5. Any pul requested executed against the server caused it throw an HTTP Error 500 at the end of each request. I had to roll back to 2.4.2 for today. I only have one web server running. Windows 2003 R2 x32 Apache 2.2.17 Python 2.7.2 mod_wsgi 3.3 ---------------- Server Running 2.5 ------------------------------------------ PS C:\users\xxxxxx\tc-extract\TcMig-Nx> hg in comparing with http://hgserver/hg/TcMig-Nx searching for changes changeset: 132:154c2525a3d5 user: xxxxxx xxxxxx<xxxxxx_xxxxxx@xxxxxx.com> date: Fri Feb 01 16:07:40 2013 -0500 summary: Add a flag for skipping the volume read when working remotely changeset: 133:1526fd7cf6e9 tag: tip user: xxxxxx xxxxxx<xxxxxx_xxxxxx@xxxxxx.com> date: Fri Feb 01 16:48:15 2013 -0500 summary: Error checking for null inputs from the initial relationship list PS C:\users\xxxxxx\tc-extract\TcMig-Nx> hg pull pulling from http://hgserver/hg/TcMig-Nx searching for changes adding changesets adding manifests adding file changes added 2 changesets with 3 changes to 2 files abort: HTTP Error 500: Internal Server Error PS C:\users\xxxxxx\tc-extract\TcMig-Nx> hg pull pulling from http://hgserver/hg/TcMig-Nx searching for changes no changes found abort: HTTP Error 500: Internal Server Error -----------------Roll Server Back to 2.4.2 ----------------------------------- PS C:\users\xxxxxx\tc-extract\TcMig-Nx> hg pul pulling from http://hgserver/hg/TcMig-Nx searching for changes no changes found PS C:\users\xxxxxx\tc-extract\TcMig-Nx>
Server Error Log just shows the following information Mon Feb 04 09:28:14 2013] [error] [client XXX.XXX.XXX.XXX] Premature end of script headers: hgweb.cgi [Mon Feb 04 09:28:22 2013] [error] [client XXX.XXX.XXX.XXX] Premature end of script headers: hgweb.cgi [Mon Feb 04 09:29:25 2013] [error] [client XXX.XXX.XXX.XXX] Premature end of script headers: hgweb.cgi [Mon Feb 04 09:32:32 2013] [error] [client XXX.XXX.XXX.XXX] Premature end of script headers: hgweb.cgi [Mon Feb 04 09:38:26 2013] [error] [client XXX.XXX.XXX.XXX] Premature end of script headers: hgweb.cgi [Mon Feb 04 09:38:57 2013] [error] [client XXX.XXX.XXX.XXX] Premature end of script headers: hgweb.cgi
Yes, there is a new bug there. You mention mod_wsgi 3.3 ... but are using hgweb.cgi? As a temporary workaround I suggest using wsgi.
Regression -> urgent
Can you add the output of pull with --debug. The 500 must be triggered by some the pushkey call (bookmarks, phase, obsolescence?). I would like to know which one.
nevermind, the issue is actually diagnose and a patch is in progress
Fixed by http://selenic.com/repo/hg/rev/e8efcc8ff5c0 Mads Kiilerich <madski@unity3d.com> hgweb.cgi: fix internal WSGI emulation (issue3804) The internal WSGI emulation in wsgicgi.py was not fully WSGI compliant and assumed that all responses sent a body. With 3fbdbeab38cc that caused a real bug when using hgweb.cgi. wsgicgi.py will now make sure headers always are sent, using the pattern from PEP 333 and similar to how it is done in c007e5c54b16. (please test the fix)
This appears to be fixed. We'll run with it all day and will report back at the COB today EST.
We've been running with this all day and have seen no problems.
*** Bug 3811 has been marked as a duplicate of this bug. ***
I am hitting this bug too. I have spend 45 minuted diagnosing this. I think this bug should be documented in the release noted *OR* better, publish 2.5.1 ASAP.
For what it's worth, the issue also shows up in Linux (my hosting provider's CloudLinux server). It's not limited to Windows, but seems consistent with version 2.5. ======================================================================== $ hg --verbose clone http://myhost/hg/cgi-bin/hgweb.cgi/myrepo/ destination directory: myrepo requesting all changes adding changesets adding manifests adding file changes added 1 changesets with 1 changes to 1 files abort: HTTP Error 500: Internal Server Error ======================================================================== Actually, the repository does get cloned, but the working dir doesn't get updated and I get the HTTP-500 error. I hope the fix gets released (as in 2.5.1) soon.
*** Bug 3807 has been marked as a duplicate of this bug. ***
*** Bug 3777 has been marked as a duplicate of this bug. ***
Reopened without explanation. reclosing.