[cvsnt] Re: [ViewCVS] Re: [ViewCVS/cvsnt] ViewCvs problem with CVSNT build68

[Bo Berglund]

Terry,
thanks for your help. The fix you suggested to the two ViewCvs files
seem to have cured my problem.
I edited both files according to your instruction and now it all works
again.

/Bo

On Thu, 23 Jan 2003 10:43:51 +0000, Terry.Ninnis at Pgen.com wrote:

>
>
>Paul,
>I have only just started using Build 65 and have also had the problem.
>
>It appears that the STDERR output from rlog is the cause of the problem and I
>have fixed my installation by doing the following:
>
>In the file popen.py:
>Change the line:
>def popen(cmd, args, mode, capture_err=1):
>to read:
>def popen(cmd, args, mode, capture_err):
>
>In the file viewcvs.py:
>Change all the calls to popen.popen to have an additional parameter set to 1
>except for the rlog call which is set to 0.
>e.g
>Change the line:
>  rlog =
>popen.popen(os.path.normpath(os.path.join(cfg.general.rcs_path,'rlog')), chunk,
>'rt')
>to read:
>  rlog =
>popen.popen(os.path.normpath(os.path.join(cfg.general.rcs_path,'rlog')), chunk,
>'rt', 0)
>
>But change:
>  fp = popen.popen(os.path.join(cfg.general.rcs_path, 'co'), (rev_flag,
>full_name), 'rb')
>to read:
>  fp = popen.popen(os.path.join(cfg.general.rcs_path, 'co'), (rev_flag,
>full_name), 'rb',1)
>
>This then ignores the STDERR stuff from rlog and everything then works. However,
> if rlog returns anything else on STDERR then you wont know about it.
>
>In the end, the code needs to cater for the errors and not just ignore it but
>this solution is OK for me.
>
>Regards
>
>Terry Ninnis
>


/Bo
(Bo Berglund, developer in Sweden)