Re: NetBSD/vax compiler bounty?

[emanuel stiebler <emu%e-bbes.com@localhost>]

On 2020-04-16 11:28, Martin Husemann wrote:
> On Thu, Apr 16, 2020 at 11:38:49AM +0100, David Brownlee wrote:
>> Just wondering  if people would be interested in contributing to a
>> bounty for fixing gcc/VAX issues?

I wouldn't mind at all ...

> What is needed is something like:
> 
>  - write a wiki page (or similar) how to easily install NetBSD 9.0 on simh
>    and test binaries there
What kind of performance would one get of the simh/va/netbsd on a decent
PC this days?

>  - file an upstream ticket for the (required) MODE_CC transition of
>    the vax backend
>  - test a native upstream gcc (w/o NetBSD or pkgsrc modifications) on that
>    setup, report any fallout in upstream bug reports (test = run the gcc
>    test suite?)
>  - try to make (native unpatched upstream) gcc to build itself and file
>    more tickets for failures

Any idea, how long it actually takes, to compile gcc on a vax?

> Then attach a bounty to each of the tickets (or the total of it), point
> to the wiki page and try to find volunteers to work on it.

Good idea to split it up ...

> The first item is simple, the second could be done mostly via cross compiling.
> But I think the expected end result should be that gcc can compile itself
> natively on VAX, which hasn't been true for upstream gcc for more than a
> decade (I guess), and for the patched version in our sources since gcc 4.8
> (at least).

Yes, exactly!

Cheers