[cvsnt] commit of -kB file gives error message in cvsdelta.cpp

[Emma Forrester]

Further investigation on the part of my colleague has found a corner
case relating to this error message, which was discovered and fixed back
in Nov 2006.

The comments for the two changes to cvsdelta.cpp are:
rev 1.5.2.22 Make BOM invisible to diff as it just confuses people.
rev 1.5.2.23 cvsdelta corner case

We took that change, built it into our cvsnt and tested it with the file 
that showed the error. It appears to have resolved the problem.

Thanks very much for your help,
Emma

Emma Forrester wrote:
> Arthur,
> Thank you very much for your response to this.
> 
> Arthur Barrett wrote:
>> Emma,
>>
>> Can you please send the results of:
>> cvs rlog <repository path>/High0.tmf,v  
> 
> The following is the result:
> RCS file: <repository path>High0.tmf,v
> head: 1.1
> branch:
> locks: strict
> access list:
> symbolic names:
> keyword substitution: B
> total revisions: 1;     selected revisions: 1
> description:
> ----------------------------
> revision 1.1
> date: 2007/10/11 16:16:54;  author: <user name>;  state: Exp;  kopt: B; 
>   commitid: 16df470e4c32d46d;  filename: High0.tmf;
> import as Release 0.1
> =============================================================================
> 
>> Can you also describe what this file is (size, etc) and as much of its
>> history as possible (eg: started out in cvsnt 2.0.41, we change/checkin
>> about once a week, each file/revision is about 1Kb, last change was a
>> month/week/day ago, this, that, the other).
>>
> 
> The file is a 86K in size and as you can see from the log it has very 
> little history. It merely was imported to cvs (we've not changed 
> versions) with the -kB option and now we come to add a new revision of 
> it and have hit this error message.
> 
>>>  From my searching of this mailing list I can only find one 
>>> mention of a similar problem under the heading 
>> There is only one other mention I can find: on Aug 31, 2007 " bug: -kB
>> binary delta with branch cause error" reported on solaris - we've never
>> seen this in the lab despite trying a fair bit...  A copy of your RCS
>> file may help - can you send that to support at march-hare.com (do not send
>> RCS files to the newsgroup - there is a strict file-size limit).
>>
>> Regards,
>>
>>
>> Arthur Barrett
> 
> The suggestion has been made by a colleague that if we just changed the 
> file's binary type (to b) we would not get this issue so we will be 
> trying this rather than waiting on any other solution.
> 
> Regards,
> Emma
> _______________________________________________
> cvsnt mailing list
> cvsnt at cvsnt.org
> http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt