On Tue, Aug 11, 2026 at 10:49:58PM -0700, Jeremy Evans wrote:
> On 08/12 06:15, Theo Buehler wrote:
> > On Tue, Aug 11, 2026 at 06:04:38PM -0700, Jeremy Evans wrote:
> > > This updates to the latest commonmarker release.
> > > 
> > > In addition to the huge number of crate version changes, this also
> > > bumps MODCLANG_VERSION so it matches the version in base.
> > > 
> > > Tested on amd64. Will be committing in a couple days unless I hear
> > > objections.
> > 
> > If it works in real-life testing, I have no objection. One reason I did
> > not proceed with the update to 2.8.2 that I had started when I bumped
> > MODCLANG_VERSION to 21 was that I couldn't manage to get the tests
> > working at all. 2.9.0 still has the same problem. (I also tried bumping
> > rb_sys to the then latest version, which was a bit fiddly.)
> 
> For testing this, I was using it to run the tests for devel/ruby-tilt,
> which pass fine. That looks like it could use it's own update (or be
> removed, as it is a pure-ruby gem leaf port).

I really can't judge if this is sufficient. If you think so, just go
ahead with the update.

ruby-tilt is yours, so if you think it's time to remove this, feel free
to go ahead with my ok.

> I didn't try bumping rb_sys, as it wasn't necessary for this update.

Well, passing or at least somewhat working tests seemed desirable since
I have neither a use nor a test case for this. I just tried to eliminate
one potential source of the linking trouble.