|
Boost-Build : |
From: David Abrahams (david.abrahams_at_[hidden])
Date: 2002-02-11 09:21:57
----- Original Message -----
From: "Vladimir Prus" <ghost_at_[hidden]>
>
> Hmm... I was trying to state that we can't have decent crawl-up behaviour
> without glob. Do you think it's possible?
No, I agree with you.
> > > I'll probably try to merge Glob changes from jam 2.4. BTW,
> > > perforce_jam_2_4 branch cannot be used for merge at all -- CVS just
gives
> > > a
> > > long list of complains. Shall we try to create a branch once again, or
> > > use diff3 locally?
> >
> > OUCH!
> >
> > You're right; it's completely useless the way it is because CVS doesn't
> > like the fact that the source was initially added on a branch. The only
way
> > to recover, AFAIK, is to rebuild the repository history :-(
>
> Or to create yet another branch.
>
> > My sense is that we're hosed and must resort to a local diff3 (or switch
to
> > using Perforce...but I don't want to chance it right now). Emacs can
help
> > with ediff.
>
> OK, will try this way.
The worst part of this, IMO, is that I can't find a 3-way diff program that
can ignore whitespace differences. Because Perforce Jam was full of
inconsistent Tab usage, many lines have had tabs replaced by spaces, which
complicates the merge. We could, I guess, write our own diff utility :-(.
Maybe the thing to do is to get the GNU diff3 source and hack it.
-Dave
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk