> It seems to me that Python Gump is NOT acting on the "tag" attribute,
> but is only getting the code from HEAD. (does a manual cvs co need to be
> done?) Traditional gump however is handling this correctly.

Gumpy seem to be aware of tag, and uses " -r 'tag' " on the CVS command line
for both checkout and update. I wondered if the issue was a change in the
'location' (like Stefan recently mentioned) but I don't think that is it
also.

Looking at this I see no -r -- so I have to assume an internal wiring
problem.

http://lsd.student.utwente.nl/gump/jakarta-poi/work/update_jakarta-poi.html

My initial suspicion (from looking at the code) is it'd take tag on <cvs but
not on <module. I'll fix that to look at both.

    ------------------------

BTW: Any thoughts on a way to ask Gump to start from fresh on a module,
delete the local copies? Gump doesn't know if metadata changes (at least
today). If we added a flag into the metadata we'd (quite likely) leave it in
be accident, and have it over used. We need something "one time", but
distributed...

We could add a version number into the metadata, both for format -- but also
for module and project 'metadata version'. We could do a clean checkout each
time the module version incremented (c.f. the value stored in the DB.)

    ------------------------

> Feel free to hit me with a cluebat if required :)

No you don't, but if you happen to have a  "CVS cluebat" handy, mind giving
me a whack w/ it? I'm pure monkey see, monkey do on CVS stuff., sorry.

regards

Adam


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to