> -----Original Message----- > From: Emmanuel Venisse [mailto:[EMAIL PROTECTED] > Sent: Saturday, November 19, 2005 4:24 AM > To: scm-dev@maven.apache.org > Subject: Re: no such provider 'clearcase' > > Thanks Jeff. > > Do you have access to Perforce and Perforce environment?
Yes I do, a few in fact. > Can you help us to write providers for them? Yes - I can test things and contribute to discussions. I hope to help code too. Time is always our enemy, and my pace may not be fast enough. First step is "ramp up" - I havent looked yet for a "how to write a provider guide", etc., but that should be the first step. BTW, I am glad you are replacing "checkout" with "edit" for these SCMs that use those commands differently. > Jeff Jensen a écrit : > > Some more command/term comparisons: > > > > Edit works well for Perforce (edit and checkout are same for > > Perforce). VSS calls it checkout. I think edit works well for all. > > > > Perforce uses the term "revert" for "unedit". > > VSS uses the term "undo checkout" for "unedit", as ClearCase does. > > > > Lock/unlock exists in Perforce, but not VSS. > > VSS checkouts are either exclusive or not exclusive, as set by a > > server config setting. It is not per checkout like > Perforce and ClearCase. > > > > "cvs edit" is the closest thing to a lock in CVS - it is only > > notifying server/users of "edit in progress" for "cvs watch"-ers. > > not only. If a file is in edit mode, other users can't commit > modification on it. CVS does not have reserved/locked checkouts. Edit/watch is meant to track and notify users editing files, not do locked checkouts. CVS allows multiple "cvs edit"s on one file. The only time I can think of that a commit fails in this scenario is when one user uses the "advisory locks" feature (cvs edit -c) and another user attempting a commit did not start with a cvs edit too. It has been awhile since I last used CVS, so I could be off a little (or a lot!)... :-) > > Quoting Emmanuel Venisse <[EMAIL PROTECTED]>: > > > > > >>do we have lock/unlock implementations in other providers? > I don't think. > >>edit/unedit seems to be the correct name of those commands, those > >>terms are use at least for clearcase and cvs. > >> > >>Emmanuel > >> > >>dan tran a écrit : > >> > >>>edit/unedit seems to be a sensible names. Emmanuel, so we need to > >>>move the implementations of lock/unlock to other provider as well? > >>> > >>>Wim, yes ... unedit is undo checkout for clearcase. > >>> > >>>-Dan > >>> > >>> > >>> > >>>On 11/18/05, *Wim Deblauwe* <[EMAIL PROTECTED] > >>><mailto:[EMAIL PROTECTED]>> wrote: > >>> > >>> I will help with the checkout command once we > understand most > >>> common usecases. Agree now > >>> that you can focus on getting release plugin to > work. It is your > >>> call to throw except in > >>> checkout command. > >>> > >>> > >>> ok, thank you for wanting to help me out. > >>> > >>> > >>> One of my experience learning from > >>> maven scm folks is write junit test cases will > help a lot during > >>> integration with other maven plugin > >>> and continuum. Starteam was a big surprise for me since it > >>> integrates seemlessly. > >>> > >>> Special thanks you for leading this effort with clearcase > >>> implementation, i am sure other users will > >>> jump in once this is started. > >>> > >>> > >>> thank you for your kind words! > >>> > >>> > >>> > >>> One other thing: I have now implemented unedit() as > "undo checkout" > >>> in clearcase. I suppose you agree on that? > >>> > >>> regards, > >>> > >>> Wim > >>> > >>> > >> > >> > > > > > > > > > > > > >
