This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [web] Revised instructions for testing using a simulator
On 29-Sep-2002, Kai Henningsen <kaih@khms.westfalen.de> wrote:
> phil@jaj.com (Phil Edwards) wrote on 26.09.02 in <20020926090119.A4457@disaster.jaj.com>:
>
> > Ah. That makes sense; CVS currently can't handle trees checked out from
> > multiple repositories. (Although it would seem obvious to do so, given
> > that the information is replicated all the way down the tree. They plan
> > to fix that, I'm told.)
>
> Well, it can handle subdirectories from different repositories just fine -
> I'm actually using that in a nightly cronjob.
In my experience, CVS does not handle subdirectories from different
repositories nicely. There are two problems. Firstly, it gives some
spurious warning messages. That's easy enough to ignore. But secondly,
there are difficulties because `cvs update -r<tag>' applies the update
even to subdirectories from a different repository in which the specified
<tag> does not exist, thus attempting to delete all the files in such
subdirectories. This makes it difficult to e.g. update the gcc
distribution to the gcc-3_4-basic-improvements branch without updating
the gcc/mercury subdirectory.
To avoid these problems, I check out the different repositories into
different directories and then use symlinks to link them together.
Fortunately `cvs update -r<tag>' doesn't recurse through symlinks.
--
Fergus Henderson <fjh@cs.mu.oz.au> | "I have always known that the pursuit
The University of Melbourne | of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh> | -- the last words of T. S. Garp.