[cvsnt] Re: RCS integrity check

[Randy McCharles]

Thanks Bo
I gave this a try. It did provide extensive information. From what I can tell most of it can be used to debug the installation/configuration of the viewcvs/cvs server. There was little additional information about the file being parsed or where python was getting hung up. i.e this file has 5800 lines and the python debug output does not indicate the source file line number it had trouble with. I did a more extensive visual inspection of the file and found the corruption and fixed it. I will send a request to viewcv, however, about somehow indicating where syntax errors in source files are.

Randy McCharles

SMART Technologies Inc.

Senior Software Developer

Tel. 403.802.3347  Fax 403.229.2531
randymccharles at smarttech.com
http://www.smarttech.com

Bringing people and ideas together.(tm)



-----Original Message-----
From: cvsnt-bounces at cvsnt.org [mailto:cvsnt-bounces at cvsnt.org]On Behalf
Of Bo Berglund
Sent: Saturday, February 04, 2006 1:29 AM
To: cvsnt at cvsnt.org
Subject: [cvsnt] Re: RCS integrity check


On Sat, 04 Feb 2006 02:01:29 +0000, Tony Hoyle
<tony.hoyle at march-hare.com> wrote:

>Randy McCharles wrote:
>> Can anyone recommend a tool to verify RCS integrity of CVS ,v files?
>> 
>> We have a few "unexpected EOF" errors reported by viewcvs, but the python backtrack doesn't tell us where in the file the problem is.
>> Since cvs checks out the files no problem, I have to think this is a minor syntax error in the history information somewhere that can be corrected if I know the line #. I hate to lose all the history but deleting/re-adding the files as rev 1.1
>> 
>> That said, it would be handy to periodically run an integrity-check of entire repositories from time to time.
>> 
>If cvs is checking the files out OK then the problem probably lies in 
>viewcvs... viewcvs hasn't accessed the rcs files directly for a while 
>now (doesn't work very well with the updated formats.. that's why cvsnt 
>has rcs functions).
>
>I'm sure there is a way of getting a better backtrace but that's more 
>python territory and I'm no expert.. Bo?
>

To get a better error output you can edit the file
viewcvs/lib/debug.py and change one line as follows:
SHOW_CHILD_PROCESSES = 0
to
SHOW_CHILD_PROCESSES = 1

If you now access the file with a problem through ViewCvs there will
be a rather lengthy error and debug report. This can be either
analyzed yourself or sent to the vievc list for help. Viewcvs changed
its name to viewvc a while back and it also changed its server.
The homepage is now here:
http://www.viewvc.org/
You can find links to its mail list subscription here:
http://viewvc.tigris.org/servlets/ProjectMailingListList



/Bo
(Bo Berglund, developer in Sweden)
_______________________________________________
cvsnt mailing list
cvsnt at cvsnt.org
http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt