No, the fix is broken. I have an uglier fix that I am testing. I also have a small unit test that is failing :-) Best, christos > On Jan 28, 2023, at 4:09 PM, Maciej W. Rozycki <macro%orcam.me.uk@localhost> wrote: > > Christos -- > > On Fri, 27 Jan 2023, Christos Zoulas wrote: > >>> On Jan 27, 2023, at 7:44 PM, Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost> wrote: >>> >>> christos@ wrote: >>> >>>> I think I found the problem.... >>> >>> Thanks, with the following your commit >>> https://mail-index.netbsd.org/source-changes/2023/01/27/msg143037.html >>> now vax bootloader builds. > > Do you have plans to submit your fix upstream? Preferably with a GAS > test case (I can help you with that) reduced from the original problematic > assembly source. The VAX port has so little target-specific coverage in > the testsuite and without that it's hard to track quality; we don't know > what the reference is even. > > Please note that it's not necessary anymore to assign the copyright of > the FSF to contribute to upstream binutils (cf. binutils/MAINTAINERS): > > -------------------------------------------------------------------------- > The alternative is to sign off the contribution by agreeing to the > Developer's Certificate of Origin (version 1.1 or later) and adding a > line to the end of the contribution that looks something like this: > > Signed-off-by: Random J Developer <random%developer.example.org@localhost> > > The details of the Developer's Certificate or Origin can be found here: > > https://developercertificate.org/ > -------------------------------------------------------------------------- > > With the requirement for copyright assignment lifted I do encourage you to > contribute any local changes you may have so as to save everyone duplicate > effort. > > Maciej
Attachment:
signature.asc
Description: Message signed with OpenPGP