David Crossley wrote: > Sina K. Heshmati wrote: >> >> I think we should stop using the svncommit and svncheckout tasks since: >> >> - There's almost no documentation available for them. >> - They don't remove the need for executing the svn command. >> >> Maybe we could consider using [1], which is BTW released under ASL 1.1 [2]. >> >> Kind regards, >> Sina >> >> [1] http://subclipse.tigris.org/svnant.html >> [2] http://subclipse.tigris.org/licenses/svnAnt.html > > Thanks Sina. However, if i read correctly then the version > of SvnAnt is "bound against subversion" to a particular > version of Subversion. > > So i suppose that we would need to provide different versions > of SvnAnt to match their client version. Sounds messy.
True. I had not realized the implications. > I like Tim's suggestion better. Executing the SVN command directly from Ant? Sure. We already have them and they have posed less problems than the alternative methods. Besides, targets like publish are most likely to be used on the server. Kind regards, SinDoc
