[jdev] Re: Use of $Id$ in source files

[Thomas Charron]

On Wednesday 22 December 2004 10:11 am, Neil Stevens wrote:
> On Wednesday 22 December 2004 07:12 am, Thomas Charron wrote:
> > On Wednesday 22 December 2004 03:05 am, Neil Stevens wrote:
> > > The idea makes merging harder because it adds unnecessary conflicts
> > > (as any two branches will have conflicts in the Id line), and isn't
> > > necessary to acheive your goal of knowing what version was checked
> > > out.  The sandbox
> >   Umm..  $Id$ does NOT add any additional conflicts.  It's a special
> > text string that is actually stored and considered to be, $Id$, and NOT
> > the string that the version represents.  The string presented is the one
> > the sandbox considers it to be, and it allows the version information to
> > be present in CVS exports for release builds and the such..
> That only applies when you're using CVS as your only tool, which rules out
> such nifty things as graphical three-way mergers.

  Ahh, but if your using CVS as the END repository, then it doesn't matter 
WHAT you change, as long as the field is still in tact.  Once you check it 
in, it will magically change to the proper versioned information.  So if your 
doing a fancy 3 way merger (besides that your a glutton for punishment), it 
works perfectly fine.

  And most version control systems have similar tags that can be used..