Hi Mark,
I actually liked (most of) these new additions to the .gitignore, since it
actually makes git clean even more powerful. The ones that I would like to
see retained are
*.o
_*.f
*.lo
*.mod
*.a
*.la
I agree that git clean is quite useful (I discovered it fairly recently),
but it plays very nicely with .gitignore. In my opinion, the tree builder
you're talking about should use
git clean -x -d -f
The -x forces git to ignore the .gitignore. What I really like, though,
(and why I want to keep all of those extensions in the .gitignore), is that
git clean -X -d -f
removes ONLY those files covered by .gitignore (capital-X as opposed to
little-x above). Thus, the .gitignore can be effectively used to clean out
all traces of compiled files (if those are ignored) without even touching
new files you may have made yet not committed. In the end, I think it's
actually the point of the .gitignore to have most of these files there.
I think we should take out the file extensions that are actually used in the
tree (i.e. .save and .html), but leave the rest.
Others thoughts?
Thanks!
Jason
- Log -----------------------------------------------------------------
commit fc6849fcf609f4311fd601c276c31748b17611bd
Author: Mark Williamson <mw529.cam.ac.uk>
Date: Sun Mar 27 17:54:48 2011 +0100
Recent .gitignore rules broke the automated tests.
The cruisecontrol automated tree builder (
http://git.ambermd.org:8080/dashboard/tab/dashboard) makes use of the very
useful command "git clean -f -d" to bring the tree back to a fresh state
before building. Essentially, it acts as an "infallible" make clean. The
recent set of .gitignore rules broke this and manifest itself as a
*completely* unrelated build error:
ar:
/home/cc/cruisecontrol-bin-2.8.4/projects/amber/AmberTools/src/netcdf/src/fortran/../f90/.libs/libnetcdff90.a:
No such file or directory
make[2]: *** [libnetcdff.la] Error 9
make[1]: *** [install-recursive] Error 1
make: *** [/home/cc/cruisecontrol-bin-2.8.4/projects/amber/lib/libnetcdf.a]
Error 2
Hence, cruisecontrol has been spewing false positives. Tracking this down
has taken ~2 days of debugging :(
-----------------------------------------------------------------------
Summary of changes:
.gitignore | 8 --------
1 files changed, 0 insertions(+), 8 deletions(-)
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Sun Mar 27 2011 - 10:30:02 PDT
