On Wed, Mar 05, 2025 at 11:24:38AM +0000, Jonathan Perkin wrote:
* On 2025-03-04 at 17:30 GMT, nia wrote:
> On Tue, Mar 04, 2025 at 05:22:52PM +0000, David Holland wrote:
>
> > ...also, if you have a long-term plan for the compiler stuff, can you
> > write it up? If nothing else it would help me and probably a bunch of
> > other people understand better.
>
> I'm leaving most of this work to jperkin, but it's unlikely
> it'll make it back to the main pkgsrc branch. Likely this
> is my last "infrastrucutre" change specific to compilers
> for a while - it's just meant to improve maintainability.
The work I'll be doing in dreckly will be based on the cleanup I did a while
back that was posted to the lists. Due to the git repository rehash it's no
longer possible to compare, but the commits are still here:
https://github.com/TritonDataCenter/pkgsrc/commits/feature/gcc-selection/trunk/
if someone wants to do something similar for pkgsrc.
What are the issues to simply applying your patch set to pkgsrc?
Up until "remove USE_GCC_RUNTIME", it seems to me your patches
result in something correct and clean, rather than something
different. (And removing it seems like a good idea, but I don't
quite see how GCC_BOOTSTRAP_PKGS works.)