cvs commit: src/gnu/usr.bin/binutils/libbfd/i386 bfd.h

John Polstra jdp at polstra.com
Fri Apr 23 09:07:45 PDT 2004


On 19-Apr-2004 Kris Kennaway wrote:
> On Mon, Apr 19, 2004 at 09:31:51AM +0300, Ruslan Ermilov wrote:
>> On Sun, Apr 18, 2004 at 04:13:14PM -0700, Kris Kennaway wrote:
>> > On Sun, Apr 18, 2004 at 12:13:10PM -0600, M. Warner Losh wrote:
>> > 
>> > > In the past, we've done this sort of import as a process to get things
>> > > going, and then deleted the files from CVS repo to not bloat it too
>> > > badly.  After things were complete, the files were re-imported on top
>> > > of the traditional place.
>> > 
>> > Note that the files from the last time this happened never were deleted:
>> > 
>> > /home/ncvs/src/contrib/ still contains gcc.295 libstdc++.295 gdb.291
>> > libio.295 libobjc.295
>> > 
>> Running ``find -type f'' in these directories shows no files.
> 
> Interesting, I certainly have files in my copy of the repo updated via
> cvsup.

You must have made your first CVSup run on a pre-existing local copy
of the repository.  If CVSup finds "extra" files that it knows nothing
about (because they don't exist in the master repository), it won't
delete them or otherwise molest them.  The basic rule is, if CVSup
didn't create the file then it won't delete it.  The next time CVSup
doesn't delete your kernel configuration file, you can thank it for
that. :-)

BTW, there's a script in the "contrib" subdirectory of the CVSup
source distribution that can scan your file tree and warn you about
this sort of thing.

> Perhaps this is another reason why removal of stuff from the
> cvs repository is evil :-)

Agreed.  Evil.  Eeeeeviiill.

John


More information about the cvs-src mailing list