On Tuesday, November 16, 2010 10:24:04 Dave Korn wrote:
> On 16/11/2010 10:54, Mike Frysinger wrote:
> > On Tuesday, November 16, 2010 02:56:58 Tristan Gingold wrote:
> >> On Nov 15, 2010, at 11:44 PM, Mike Frysinger wrote:
> >>> any qualms with starting a top level .gitignore ?  i'd have this be in
> >>> the top level .gitignore, and do a few more subdir specific ones.
> >> 
> >> Not sure about *.texi
> > 
> > you're right ... that one is wrong.  there are a lot of generated .texi
> > files, but there are many that are kept in the tree.  so i'll have to
> > use more specific matches.
> > 
> > but the general idea is OK with people ?
> 
>   Don't forget to add ".gitignore" to .cvsignore!  But yeah, sounds like a
> no-brainer to me.

i looked through .cvsignore and picked out a few more entries.  here is the 
top level file i'll commit if there's no other suggestions.

note: i know that one or two subdirs actually commit files named "config.h" to 
their tree, but that's a lot more of an exception than a rule.  so for those 
subprojects (looks like just newlib and winsup), i'd propose adding a 
.gitignore entry of "!config.h".
-mike

*.diff
*.patch
*.orig
*.rej

*~
*.a
*.flt
*.gdb
*.gmo
*.info
*.la
*.lo
*.o
*.tmp

.deps
.libs

autom4te.cache
config.cache
config.h
config.intl
config.log
config.status
dir.info
libtool
Makefile
stamp-*
POTFILES
*-POTFILES
*/po/Makefile.in

.gdbinit
.gdb_history
core

lost+found

*.log
*.sum