implement va_start and va_arg

[folkertdev]

depends on

• Expose va_arg and va_start gcc#84
• add new_va_arg and VaList gccjit.rs#70

[antoyo]

Thanks for your work on this!
One question:

View changes since this review

[antoyo]

I published a new version of both gccjit and gccjit_sys.

You will also need to update the commit in this file. This file contains the hash from GCC, not gccjit.rs.

[antoyo]

I'll rebuild the m68k GCC.
I'll also try to check if there's something wrong with the GCC version without int128.

[folkertdev]

I've removed the test file, so I'll leave it to you to make CI pass.

[antoyo]

The m68k CI is fixed, but the stdarch tests are still failing.
I did a rerun from the master and it passes, so there might be something in this PR that breaks the stdarch tests.

I opened a test PR to run the stdarch tests another time: let's see if the tests pass there.

Edit: The stdarch CI uses the pinned GCC, so maybe this is caused by the GCC update.

[antoyo]

I finished the sync and the stdarch tests now pass.
Could you please rebase this PR to get this fix?

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[folkertdev]

All green now!

[antoyo]

Thanks a lot for this contribution!

Did you want me to send the GCC patch upstream (to the forge) or do you want to do it yourself?

[folkertdev]

No if you can handle that that would be great

[antoyo]

Ok. I'll send this patch in the following weeks.