This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC 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: Please don't commit changes to gcc/go/gofrontend


On Tue, Nov 19, 2013 at 7:08 AM, Diego Novillo <dnovillo@google.com> wrote:
> On Tue, Nov 19, 2013 at 9:48 AM, Ian Lance Taylor <iant@google.com> wrote:
>> Hi, as noted in gcc/go/README.gcc, the files in gcc/go/gofrontend are
>> actually mirrored from a different repository.  Please do not directly
>> commit changes to those files.  Instead, send the changes to me.  I
>> will commit them upstream.  Thanks.
>
> Ugh, sorry.  This is really counter-intuitive.  To properly test
> changes, one needs the patch in the tree.  So when I do the final
> commit, it is not easy to remember that I need to take it out (and
> taking it out means undoing a local commit, which is yet another
> operation).
>
> Sorry, but I would expect these problems to continue.  Is it possible
> for you to cope in some other way?  Like your merging script noticing
> changes and incorporating them into your tree?

I have no problem coping.

Because the Go frontend is under a BSD-style license, the steering
committee asked me to keep the master copy out of the tree, which I am
doing.  The frontend is then under a different CLA regime;
contributors sign the Google CLA rather than going through the FSF
copyright assignment process.  This is why the procedure is as it is.
(I'm not worrying about the CLA for these trivial mechanical changes.)

There aren't many people who commit tree-wide changes like this.  If
you get it wrong, I will cope.  But it's easier if you remember.

It's also fine to not test Go.  That was the agreement I made when the
Go frontend was added under this peculiar regimen.  Of course I
appreciate your testing Go, but you don't have to.


> Alternately, would it be possible to install some kind of svn hook
> that only allows certain users to commit?

It seems to me that that would be more frustrating, as your commit
would fail.



On Tue, Nov 19, 2013 at 7:08 AM, Richard Sandiford
<rsandiford@googlemail.com> wrote:

> +1 FWIW.  I think anyone making an interface change is going to assume
> that they should update all in-tree users.  It's unusual to say that
> people should knowingly break the build

Yes, it is unusual.

I don't see any good immediate solutions here.

Longer term, we (Chris Manghane and I) are actively working toward
removing all GCC header files from the gofrontend sources.  That is
the meaning of the various "use the backend interface" patches.
However, it is a long slog.  When that is done, these sorts of
mechanical changes will never need to touch the gofrontend sources.

Ian


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