AW: Re: Largefiles Issue on Update

Wed Dec 12 10:27:35 UTC 2012

Matt, 

Thanks for your reply. 

Using a clean update was my first try. The second approach I tried was to make a clean update of an older rev in which the .ppt had not been added yet. This worked, but I was not able to (clean) update to any rev with the largefile.

I can confirm that the .hglf/doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC Training.ppt.orig was there when I reported this issue.

Maybe hg was not able to move it out of the way?

My colleague made a new clone from the whole repo, but has a copy of the old defective subrepo. I just checked and realized that after some more trials he did both files are present as you suggested.

So if it would be any benefit I could try some new ideas on the old subrepo.

Thanks for your help!

Best regards

SMS Siemag Aktiengesellschaft

Dipl.-Inf. Andreas Kämpf
Electrical and Automation Systems
Process Control Systems Hot Rolling Mills

-
Wiesenstraße 30, 57271 Hilchenbach-Dahlbruch, Germany
Tel.:    +49 (0) 2733 29-3015
Fax:    +49 (0) 2733 29-773015
E-Mail: andreas.kaempf at sms-siemag.com

Date: Tue, 11 Dec 2012 01:34:09 +0000 (UTC)
From: Matt Harbison <matt_harbison at yahoo.com>
To: mercurial at selenic.com
Subject: Re: Largefiles Issue on Update
Message-ID: <ka62ig$pl6$1 at ger.gmane.org>
Content-Type: text/plain; charset=UTF-8

On Mon, 10 Dec 2012 15:29:05 +0000, K?mpf, Andreas (SMS Siemag AG) wrote:

> Hi!
> 
> When trying to perform an update on one special computer, I experience 
> a problem I cannot comprehend:
> 
> 
> C:\Users\u928>hg update --repository X:\SMSSiemag\MSC --config 
> ui.merge=internal:merge --rev 956 --check --traceback --debug --verbose
>   searching for copies back to rev 956 unmatched files in other:
>    .hglf/doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC
>    Training.ppt
> resolving manifests
>  overwrite: False, partial: False ancestor: 842471620b1b, local:
>  842471620b1b+, remote: 353293c1e2e4 .hglf/doc/04.MSC-Mill Level  
> 2/04.07.Training/04.07.01.PCFC Training.ppt: remote created -> g
> updating: .hglf/doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC 
> Training.ppt 1/1 files (100.00%) getting .hglf/doc/04.MSC-Mill Level 
> 2/04.07.Training/04.07.01.PCFC Training.ppt 1 files updated, 0 files 
> merged, 0 files removed, 0 files unresolved getting changed largefiles 
> found
> fd2294c048817939301996b557c03b85f6a1a378 in store Traceback (most 
> recent call last):
>   File "mercurial\dispatch.pyo", line 88, in _runcatch File
>   "mercurial\dispatch.pyo", line 740, in _dispatch File
>   "mercurial\dispatch.pyo", line 514, in runcommand File
>   "mercurial\dispatch.pyo", line 830, in _runcommand File
>   "mercurial\dispatch.pyo", line 801, in checkargs File
>   "mercurial\dispatch.pyo", line 737, in <lambda>
>   File "mercurial\util.pyo", line 471, in check File
>   "mercurial\extensions.pyo", line 144, in wrap File
>   "mercurial\util.pyo", line 471, in check File
>   "hgext\largefiles\overrides.pyo", line 289, in overrideupdate File
>   "mercurial\util.pyo", line 471, in check File
>   "mercurial\commands.pyo", line 5858, in update File
>   "mercurial\extensions.pyo", line 189, in wrap File
>   "hgext\largefiles\overrides.pyo", line 646, in hgupdate File
>   "hgext\largefiles\lfcommands.pyo", line 447, in updatelfiles File
>   "hgext\largefiles\lfcommands.pyo", line 447, in <lambda> File
>   "hgext\largefiles\lfcommands.pyo", line 475, in _updatelfile File
>   "shutil.pyo", line 82, in copyfile
> IOError: [Errno 2] No such file or directory:
> 'X:\\SMSSiemag\\MSC\\doc/04.MSC-Mill Level 
> 2/04.07.Training/04.07.01.PCFC Training.ppt'
> abort: No such file or directory: X:\SMSSiemag\MSC\doc/04.MSC-Mill 
> Level 2/04.07.Training/04.07.01.PCFC Training.ppt
> 
> Does this make sense to anyone? Where is the problem on this very 
> computer? On other computers it works fine, all running TortoiseHg 
> 2.5.1 with Mercurial 2.3.2. The team server is running Mercurial 2.2.2 
> via Lenny backports served by SCM Manager.
> 
> My next approach would be to delete the repo and clone it from our 
> team repo once again.
> 
> Thanks for any help!

Do you remember what you did to get the repo in this state?  I've never had this issue before (though I generally update with -C to get a clean slate).

According to the code, it appears the filesystem contains these:

  .hglf/doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC Training.ppt
  .hglf/doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC Training.ppt.orig

which makes it want to try to copy this:

  doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC Training.ppt

to this:

  doc/04.MSC-Mill Level 2/04.07.Training/04.07.01.PCFC Training.ppt.orig

But the source ppt file doesn't exist.  It seems like a bad assumption that if the standin file (under .hglf/) exists, the actual largefile must exist.

Moving this repo somewhere else and recloning should fix this if you are in a pinch (and I suspect so would 'update -C' or deleting the *.orig files and updating again), but I'm interested in getting more info about this before this particular state gets destroyed.

--Matt

Vorsitzender des Aufsichtsrates: Dr. Heinrich Weiss    
Vorstand: Burkhard Dahmen, Vorsitzender, Dr. Guido Kleinschmidt, Dieter Rosenthal, Eckhard Schulte, Ruediger Zerbe
Sitz der Gesellschaft: Duesseldorf
Registergericht: Amtsgericht Duesseldorf, HRB 25779
USt-IdNr: DE811136276