Re: treelang patch parts n of 6

[Geoff Keating <geoffk at geoffk dot org>]

> CC: geoffk@geoffk.org, gcc-patches@gcc.gnu.org,
>    "Joseph S. Myers" <jsm28@cam.ac.uk>
> Date: Mon, 06 May 2002 06:44:03 +1000
> From: Tim Josling <tej@melbpc.org.au>

> General
> -----------
> 
> I will address the issues raised.
> 
> I realise that not everyone believes that treelang is a good idea. It
> may well be valid that if you need a sample language to implement a
> front end to gcc, you probably should not implement a front end to gcc.
> Even for those who do not *need* treelang though, I think it will help
> and I even have some evidence for that statement.

I don't mind having an example frontend, but only if it's a good
example.  At present, the treelang frontend is more likely to lead
people into making mistakes.

> Regressions
> ----------------
> 
> >    Please correct these and re-submit the patch.
> >
> > The code was checked in already.... Hrm...
> 
> IMHO it is most unlikely that I have caused any regressions, given that
> I did not change any of the core gcc code and treelang does not even
> build by default.

That's not true.  We now have a file in the tree with an unclear
copyright status.  That is a regression, and a serious one.

> The faults pointed out today were present in the previous versions of
> treelang that have been posted/advised to the list over a period of time
> and none of these issues were pointed out then.

This is why we require that patches are reviewed before they can be
contributed and given positive approval.  It's quite possible that
no-one ever looked at those parts of the patches you posted.

-- 
- Geoffrey Keating <geoffk@geoffk.org> <geoffk@redhat.com>