This is the mail archive of the kawa@sources.redhat.com mailing list for the Kawa project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Some CVS suggestions


pax! wrote:
- I think that Makefile.in files shouldn't be in CVS, since AFAIK they
are always generated from Makefile.am files. In fact, after compiling
the system, all Makefile.in appear as modified.
One could argue both ways.  Leaving out Makefile.in (and configure and
some other generated files) would be more logical, but it would require
that anyone building from CVS would have to have a compatible autoconf
(and automake and libtool etc).

GNU projects tend to include some of these generated files in CVS.
For example Gcc does.  The GNU standard is certainly to require that
they be in a source (.tar.gz) file.

Since I'm currently the only one checking things into CVS, and it is no
harder to check in Makefile.am *and* Makefile.in than just check in
Makefile.am, and having Makefile.in in the repository makes it easier
for people to use it, my inclination is to keep things as they are.

Also, changing things means we would have to modify the documentation,
in that building from CVS and building from .tar.gz would be different.

- There should be .cvsignore files in all directories, it makes CVS
easier to look at (at least in my XEmacs CVS client). I could
contribute my own .cvsignore files if needed.
I ahve no strong feelings either way.  I tend to build in a separate
directly from the source directory, so most generated files don't
show up in 'cvs update', but I can check in .cvsignore files, as
long as they won't require much maintainance.
--
	--Per Bothner
per@bothner.com   http://www.bothner.com/per/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]