Maybe hg verify could also give some information on how to fix a damaged
repo (export your changes and get a new repo?). Or at least on what the
errors mean. 

PS: Thanks for the great work on Mercurial!

Yeah we should ask for hg verify in those cases

I run into a similar error, pulling from 
http://bitbucket.org/ArneBab/1w6 
with Mercurial 1.5

I don't get an error on the other machine with 1.6.1

Error: 

abort: index data/sonstiges/Interessante Charaktere darstellen Teil 2.txt.i
unknown format 25701

The reason for this error seems to be the integrity of the repo I want to
pull to, because I get integrity errors on hg verify, and one of them named
unknown format 25...

So this might be an uncaught exception provising misguided information. 

possible fix: catch the exception and ask the user to do a hg verif. 

Assumption why I got this: my system had some power failaurs and it runs off
a USB-stick. 

I also have a tarball, but it's 50MiB...=.

Not sure it's our fault here, setting to need-eg

> I'm also running tortosehg in parallels for windows and the Mac OSX for 
Leopard 10.5 on the same shared directory

I guess that's the reason why. It's probably not safe to do that.

I should also mention I got rid of the corrupted file (with my xcdatamodel) 
issue by removing the file and creating one with a different name. If I added 
a newly created file with the same name, then I'd get a corrupted message. 
Change the name, and the corruption seemed to go away letting me checkin.

Sure,

I'm running Hudson on my repository for continuous integration. I noticed 
the problem started as soon as I checked in a iWork Numbers document into 
the repository. I started getting this error when hudson tried to clone the 
repository:
/Users/Shared/hudson/build/TruckLoadOf
requesting all changes
adding changesets
adding manifests
adding file changes
transaction abort!
rollback completed
abort: index data/Data/Game Database.numbers.i unknown format 3492!


I figured that the Spreadsheet, in this case, was represented more like a 
binary file...and probably not great for storing into the repository, so I 
made another change to tell Mercurial to forget it. Then I'd get this error 
relating to the Doxygen file I have checked in, which is plain text. 

requesting all changes
adding changesets
transaction abort!
rollback completed
abort: index data/Doxyfile.i unknown format 24750!

The filename is "Doxyfile" and has no extension. It's plaintext generated 
from the Doxygen GUI. 

At some point, I also had the repository telling me that in XCode, my 
xcdatamodel Element file was corrupted, even though XCode itself had no 
problem reading and updating the file. This prevented me from doing any kind 
of checkin to the repository. The Element file it was referring to inside 
the model didn't seem to be plain text or XML either. 

I was running 1.43, but then upgraded to 1.5 after coming across some 
comments in tortoisehg that seemed like might have fixed it. I'm struggling 
to find the page that referenced a similar issue. Below is my before and 
after version number outputs:

hg version
Mercurial Distributed SCM (version 1.4.3)

Copyright (C) 2005-2010 Matt Mackall <mpm@selenic.com> and others
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Garys-iMac:TruckLoadOf garyyund$ hg version
Mercurial Distributed SCM (version 1.5+20100307)

This probably doesn't matter so much, but it's probably good to take the 
whole picture into account and not leave a variable out:
I'm also running tortosehg in parallels for windows and the Mac OSX for 
Leopard 10.5 on the same shared directory, both with the same version of 
Mercurial at all times, so when I update the windows flavor, I also update 
my Leopard version.

@g01:

Your problem look different, can you describe more precisely what you did?

I'm also receiving this problem as well. I tried recreating my repository, 
which exists locally, by simply copying all the files and re-adding them to a 
new repository. If I do an hg verify, I get things like:
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
Doxyfile@0: broken revlog! (index data/Doxyfile.i unknown format 24750)
 MyProject.xcodeproj/project.pbxproj@0: broken revlog! (index 
data/MyProject.xcodeproj/project.pbxproj.i unknown format 27918)

... etc.

mpm is on vacation until the 3rd, so it's probably better to just wait. If 
you think otherwise you'll need to provide the tarball or other information 
to another dev.

Is there any further information you need from me?

You can mail me the tarball at mpm at selenic.com.

At home on the Mac, I tried to push again to a fresh-init repo and got
another failed result.

On IRC, mpm suggested that there is an error coming from in the byte
patterns & the SSH connection.

This repo *can* clone from a local filesystem and push to HTTP.

I can bzip up the whole repository and put it on my web server for
further forensics by the Hg team.

$ hg push ssh://user@server/masters-hg/
remote:
pushing to ssh://user@server/masters-hg/
searching for changes
remote:    ****************************************************************
remote:    * WARNING: To protect the system from unauthorized use and to  *
remote:    * ensure that the system is functioning properly, activities   *
remote:    * on this system are monitored recorded and subject to audit.  *
remote:    * Use of this system is expressed consent to such monitoring   *
remote:    * and recording. Any unauthorized access or use of this system *
remote:    * is prohibited and subject to criminal and civil penalties.   *
remote:    ****************************************************************
remote:
remote: adding changesets
remote: adding manifests
remote: transaction abort!
remote: rollback completed
remote: abort: unknown compression type 'M'!
abort: unexpected response: empty string
$ hg --version
Mercurial Distributed SCM (version 1.4.1+20091201)

Copyright (C) 2005-2009 Matt Mackall <mpm@selenic.com> and others
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Finarfin:~/Documents/masters/repo vlion$ uname -a
Darwin Finarfin.local 10.2.0 Darwin Kernel Version 10.2.0: Tue Nov  3
10:37:10 PST 2009; root:xnu-1486.2.11~1/RELEASE_I386 i386
$

Further information:

I have 2 Windows 'client' computers, 1 Mac 'home server' computer, and
a remote server that someone else admins(the one I'm trying to ssh
to). My 2 local computers can all push/pull to the Mac quite happily.
I was able to clone on one of my Windows machines without an issue.

I updated the server version of hg as per tonfa's information and
created a brand-new repo. My client version of hg for this computer
was 1.4.2 on windows 7.

I then got this message when I tried to push:

running ""C:\Program Files\TortoiseHg\TortoisePlink.exe" -ssh -2
username@redactedserver.edu "hg -R masters-hg serve
 --stdio""
pushing to ssh://username@redactedserver.edu/masters-hg
searching for changes
178 changesets found
remote: adding changesets
remote: adding manifests
remote: transaction abort!
remote: rollback completed
remote: abort: unknown compression type 'M'!
abort: unexpected response: empty string

Mercurial Distributed SCM (version 1.4.2)

Copyright (C) 2005-2009 Matt Mackall <mpm@selenic.com> and others
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
D:\Documents\thesis\hg
>

Rummaging around the store/data/* tree, it appears that the 'demoUI' folder 
is simply not there. 

Checking through the revlog, revision 166 is where I added the demoUI 
folder. 

Looking at my other up-to-date clones of the repo, I can successfully verify 
up to rev 176, which is the correct version. 

(I don't have data loss, but when the SCM burps and dies, I figure it's a 
bug that should be reported. :-) )

~/masters/masters-repo-hg]$ ~/temp/mercurial-1.4.3/hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
 trunk/sandbox/demoUI/demoUI/Form1.Designer.cs@166: broken revlog! (index 
data/trunk/sandbox/demoUI/demoUI/Form1.Designer.cs.i is corrupted)
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: unpacking 17b94768e3e8: long int 
too large to convert to int
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: duplicate revision 1 (0)
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: 17b94768e3e8 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: unpacking 17b94768e3e8: long int 
too large to convert to int
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: unpacking d2bb910b2ea3: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: duplicate revision 3 (1)
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: 17b94768e3e8 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: unpacking 17b94768e3e8: long int 
too large to convert to int
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: duplicate revision 4 (2)
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: d2bb910b2ea3 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: unpacking d2bb910b2ea3: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: unpacking f7cd99ec3424: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: duplicate revision 6 (3)
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: 17b94768e3e8 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: unpacking 17b94768e3e8: long int 
too large to convert to int
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: duplicate revision 7 (4)
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: d2bb910b2ea3 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: unpacking d2bb910b2ea3: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: duplicate revision 8 (5)
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: f7cd99ec3424 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: unpacking f7cd99ec3424: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@169: unpacking c1111d6f8592: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: duplicate revision 10 (6)
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: 17b94768e3e8 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@166: unpacking 17b94768e3e8: long int 
too large to convert to int
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: duplicate revision 11 (7)
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: d2bb910b2ea3 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@167: unpacking d2bb910b2ea3: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: duplicate revision 12 (8)
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: f7cd99ec3424 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@168: unpacking f7cd99ec3424: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@169: duplicate revision 13 (9)
 trunk/sandbox/demoUI/demoUI/Form1.cs@169: c1111d6f8592 not in manifests
 trunk/sandbox/demoUI/demoUI/Form1.cs@169: unpacking c1111d6f8592: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/Form1.cs@170: unpacking 927d362309f7: unknown 
compression type '|'
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: unpacking dc1758bab638: long 
int too large to convert to int
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: duplicate revision 1 (0)
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: dc1758bab638 not in 
manifests
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: unpacking dc1758bab638: long 
int too large to convert to int
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@168: unpacking 419a11421972: 
unknown compression type '\xc8'
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: duplicate revision 3 (1)
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: dc1758bab638 not in 
manifests
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@166: unpacking dc1758bab638: long 
int too large to convert to int
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@168: duplicate revision 4 (2)
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@168: 419a11421972 not in 
manifests
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@168: unpacking 419a11421972: 
unknown compression type '\xc8'
 trunk/sandbox/demoUI/demoUI/demoUI.csproj@169: unpacking 2d90957aed6a: 
unknown compression type '\xc8'
 trunk/sandbox/demoUI/demoUI/program_trace.cs@168: unpacking ded05e5095ef: 
long int too large to convert to int
 trunk/sandbox/demoUI/demoUI/program_trace.cs@168: duplicate revision 1 (0)
 trunk/sandbox/demoUI/demoUI/program_trace.cs@168: ded05e5095ef not in 
manifests
 trunk/sandbox/demoUI/demoUI/program_trace.cs@168: unpacking ded05e5095ef: 
long int too large to convert to int
 trunk/sandbox/demoUI/demoUI/program_trace.cs@169: unpacking b8bdfe0eb3fe: 
unknown compression type '`'
 trunk/sandbox/demoUI/demoUI/reconstructer.cs@168: unpacking de125c27888f: 
long int too large to convert to int
 trunk/sandbox/demoUI/demoUI/reconstructer.cs@168: duplicate revision 1 (0)
 trunk/sandbox/demoUI/demoUI/reconstructer.cs@168: de125c27888f not in 
manifests
 trunk/sandbox/demoUI/demoUI/reconstructer.cs@168: unpacking de125c27888f: 
long int too large to convert to int
 trunk/sandbox/demoUI/demoUI/reconstructer.cs@169: unpacking 8d93b1bac5a8: 
unknown compression type '\x97'
 trunk/sandbox/demoUI/demoUI/xmos_cpu.cs@168: unpacking 2ddb18886dbe: long 
int too large to convert to int
 trunk/sandbox/demoUI/demoUI/xmos_cpu.cs@168: duplicate revision 1 (0)
 trunk/sandbox/demoUI/demoUI/xmos_cpu.cs@168: 2ddb18886dbe not in manifests
 trunk/sandbox/demoUI/demoUI/xmos_cpu.cs@168: unpacking 2ddb18886dbe: long 
int too large to convert to int
 trunk/sandbox/demoUI/demoUI/xmos_cpu.cs@169: unpacking 1e9210d621e2: 
unknown compression type 'U'
warning: orphan revlog 
'data/trunk/sandbox/demoUI/demoUI/Form1.Designer.cs.i'
1070 files, 172 changesets, 1437 total revisions
1 warnings encountered!
63 integrity errors encountered!
(first damaged changeset appears to be 166)
~/masters/masters-repo-hg]$

Use a recent hg, it will correctly report the file.

(just to make local inside the tarball, and use /path/to/hg verify)

I have 161M of data in the store/* directory.

Is there a way to figure out which file is fried?

On Tue, Feb 9, 2010 at 10:01 PM, Matt Mackall
<bugs@mercurial.selenic.com> wrote:
>
> Matt Mackall <mpm@selenic.com> added the comment:
>
> Alright, newer hg verify will do better than that, but not much.
>
> What's happened is that some portion of the .i file in .hg/store/data
> associated with a particular file has gotten damaged in such a way that
> Mercurial can do no operations involving the history of that file. If
> there's no proprietary data in that file, we'd like to look at it.
>
> If your local repo is simply a superset of changes on your server, you can
> back up your server repo, init an empty repo, and push your local repo back
> onto the server.
>
> You also may be able to take that .i file and compare it bytewise with the
> working copy on your local machine to spot the problem.
>
> Generally this sort of damage requires kernel bugs, memory corruption, or
> the like to create.
>
> ----------
> status: need-eg -> in-progress
>
> ____________________________________________________
> Mercurial issue tracker <bugs@mercurial.selenic.com>
> <http://mercurial.selenic.com/bts/issue2031>
> ____________________________________________________
>
>

Alright, newer hg verify will do better than that, but not much.

What's happened is that some portion of the .i file in .hg/store/data
associated with a particular file has gotten damaged in such a way that
Mercurial can do no operations involving the history of that file. If
there's no proprietary data in that file, we'd like to look at it.

If your local repo is simply a superset of changes on your server, you can
back up your server repo, init an empty repo, and push your local repo back
onto the server.

You also may be able to take that .i file and compare it bytewise with the
working copy on your local machine to spot the problem.

Generally this sort of damage requires kernel bugs, memory corruption, or
the like to create.

Here you go:

$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
** unknown exception encountered, details follow
** report bug details to http://www.selenic.com/mercurial/bts
** or mercurial@selenic.com
** Mercurial Distributed SCM (version 1.2.1)
** Extensions loaded: convert, graphlog
Traceback (most recent call last):
  File "/home/nath5573/bin/hg", line 27, in ?
    mercurial.dispatch.run()
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 16, in run
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
25, in dispatch
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
41, in _runcatch
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
372, in _dispatch
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
247, in runcommand
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
417, in _runcommand
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
377, in checkargs
  File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line
371, in <lambda>
  File "/net/grads/nath5573/lib/python/mercurial/util.py", line 718, in check
  File "/usr/lib/python2.4/commands.py", line 2944, in verify

  File "/net/grads/nath5573/lib/python/mercurial/hg.py", line 289, in verify
  File "/net/grads/nath5573/lib/python/mercurial/verify.py", line 15, in verify
  File "/net/grads/nath5573/lib/python/mercurial/verify.py", line 177,
in _verify
  File "/net/grads/nath5573/lib/python/mercurial/localrepo.py", line
512, in file
  File "/net/grads/nath5573/lib/python/mercurial/filelog.py", line 13,
in __init__
  File "/net/grads/nath5573/lib/python/mercurial/revlog.py", line 453,
in __init__
  File "/net/grads/nath5573/lib/python/mercurial/revlog.py", line 369,
in parseindex
ValueError: corrupt index file
$

On Tue, Feb 9, 2010 at 6:06 PM, Matt Mackall <bugs@mercurial.selenic.com> wrote:
>
> Matt Mackall <mpm@selenic.com> added the comment:
>
> Please run hg verify on the affected repos.
>
> ----------
> nosy: +mpm
>
> ____________________________________________________
> Mercurial issue tracker <bugs@mercurial.selenic.com>
> <http://mercurial.selenic.com/bts/issue2031>
> ____________________________________________________
>
>

Please run hg verify on the affected repos.

Some additional data:
Logged onto the server and tried to do a revert to a previous good version.
It went through the repository for a while and then died on a particular file 
I had modified today. I repeated this several times, walking back through the 
history, then branched once I was able to revert successfully. Unfortunately, 
merging in from the corrupted rev still gave me a corrupt index file 
exception.

I'm not sure right now why this is corrupted. The only thing I can think of is 
I have a local machine serving out hg with apache and occasionally i've done 
work on that repo directly.

Hi,

Somehow I fried my index file. I haven't done any oddball stuff. 
Commit/push/pull/merge is all I do.

Email me if you want specific details that can help you out.

remote: adding manifests
remote: adding file changes
remote: transaction abort!
remote: rollback completed
remote: ** unknown exception encountered, details follow
remote: ** report bug details to http://www.selenic.com/mercurial/bts
remote: ** or mercurial@selenic.com
remote: ** Mercurial Distributed SCM (version 1.2.1)
remote: ** Extensions loaded: convert, graphlog
remote: Traceback (most recent call last):
remote:   File "/home/nath5573/bin/hg", line 27, in ?
remote:     mercurial.dispatch.run()
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
16, in run
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
25, in dispatch
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
41, in _runcatch
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
372, in _dispatch
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
247, in runcommand
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
417, in _runcommand
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
377, in checkargs
remote:   File "/net/grads/nath5573/lib/python/mercurial/dispatch.py", line 
371, in <lambda>
remote:   File "/net/grads/nath5573/lib/python/mercurial/util.py", line 718, 
in check
remote:   File "/usr/lib/python2.4/commands.py", line 2641, in serve
remote:     
remote:   File "/net/grads/nath5573/lib/python/mercurial/sshserver.py", line 
40, in serve_forever
remote:   File "/net/grads/nath5573/lib/python/mercurial/sshserver.py", line 
47, in serve_one
remote:   File "/net/grads/nath5573/lib/python/mercurial/sshserver.py", line 
194, in do_unbundle
remote:   File "/net/grads/nath5573/lib/python/mercurial/localrepo.py", line 
2040, in addchangegroup
remote:   File "/net/grads/nath5573/lib/python/mercurial/localrepo.py", line 
512, in file
remote:   File "/net/grads/nath5573/lib/python/mercurial/filelog.py", line 
13, in __init__
remote:   File "/net/grads/nath5573/lib/python/mercurial/revlog.py", line 
453, in __init__
remote:   File "/net/grads/nath5573/lib/python/mercurial/revlog.py", line 
369, in parseindex
remote: ValueError: corrupt index file
('unexpected response:', '')

[command interrupted]