Re: Bountysource campaign for gcc-vax

["Maciej W. Rozycki" <macro%orcam.me.uk@localhost>]

On Thu, 22 Apr 2021, John Paul Adrian Glaubitz wrote:

> > I tried doing a build of current as of a couple of days ago.
> > Cross compilation worked fine, and seems to give a pretty usable system.
> > However native building on a VAX fails when trying to compile gcc itself.
> 
> Yep, that's the state of the VAX backend before the MODE_CC conversion.

 I have fixed some issues in the course, but that's of course not a 
guarantee GCC will bootstrap itself natively on a VAX now.  There are 
still many failures in the testsuite which may or may not affect building 
the compiler itself, and it would be great if people at least tried to 
triage them, so that we know what's urgent to fix and what's an obscure 
corner case only, or perhaps a faulty test (e.g. assuming IEEE 754 FP).

 Myself I haven't yet come to look into the libstdc++ tests that run away 
and make regression testing crawl, but I'll see what I can do about them.

> > So, should I just ignore this, and wait for the much improved later gcc,
> > or did something actually get into gcc 10?
> 
> Since backporting all the changes would probably a huge effort, you have no
> other choice but to wait, I'm afraid. Maybe the VAX port can be switched
> over to gcc-11 the moment it has been uploaded to the NetBSD archives.

 You may try cherry-picking some changes though, especially those that 
addressed internal compiler errors.  Of course giving GCC 11 a try anyway 
would I think be a good idea anyway, even if not making it *the* compiler 
to use.

 Also if NetBSD happens to carry any VAX-specific local changes that have 
not been upstreamed, it would be good if they were.  Now it's a good time 
to have them merged early so there's plenty of time to verify them before 
GCC 12.  Any such changes necessarily have to be regression-tested though 
and if that turns out problematic, then I can pass them through my system. 
Just let me know (and send the patches of course).

  Maciej