[cfe-commits] [PATCH] Add PNaCl ABIInfo

[David Meyer]

Chandler,

Branching is exactly what we're trying to avoid. When the project
started, LLVM was branched, and now we are stuck maintaining a branch
that differs from upstream by a diff of ~14,000 lines.

We'd like to avoid the same mistake with Clang, if possible, which is
why I don't mind delaying to have these discussions, as long as they
can be resolved in a reasonable amount of time.

- pdox





On Fri, Sep 23, 2011 at 12:41 AM, Chandler Carruth <chandlerc at google.com> wrote:
> On Thu, Sep 22, 2011 at 11:59 PM, John McCall <rjmccall at apple.com> wrote:
>>
>> > While our ABI is not 100% finalized, until this patch is committed, we
>> > can't do any external testing with Clang. (unless we create our own
>> > branch, which we'd rather not)
>>
>> I'm sorry, but we do expect code to go through code review, which
>> includes a thorough review of the underlying motivations and technical
>> design.
>
> I also don't see why this would block testing. There are lots of ways to
> patch something together if you just want to get feedback on what you're
> doing from users... That might actually provide good data for the very
> codereview and design discussion taking place. It's not like external
> parties can start using what you have in production environments or regular
> development processes if the very ABI is still in flux, but maybe I don't
> understand something.
> If you just want a way to share what you're working on, toss it up on github
> with the patch applied and point people at it. There are lots of ways of
> sharing code during the review process.