On Tue, Oct 19, 2021 at 07:50:40PM -0400, Brad Smith via SeaBIOS wrote:
On 8/14/2021 1:02 AM, Brad Smith wrote:
As an OpenBSD developer I'm very interested in these patches. SeaBIOS is the only package left in our ports tree still using lld to link and one of very few ports not building with Clang. I just happened to come across your diffs looking to see if anything had been done in this area when searching via Google.
It's been pretty quiet here on the SeaBIOS mailing list recently. There hasn't been that much new code in a while - the main emphasis has been on incremental changes.
If someone wants to send in a series of incremental patches that convert the build to work with both clang and binutils then I'll take a look at them. I don't see an issue with a more robust build. However, the main emphasis is going to be on not introducing a regression. That is, I don't think simplifying the build requirements is a sufficient reason to risk introducing a regression. In particular, one of the main goals of SeaBIOS is to support old software (30+ years) - it's really difficult to identify and track down regressions when they do occur.
There was a series of patches from Fāng-ruì Sòng in April of 2020:
There was a review of those patches and, if I recall correctly, it was determined that more testing would be needed before we were confident in merging. Alas, it seems no one wanted to do that testing.
The same author of those patches also went on to submit patches to binutils that knowingly broke SeaBIOS and did not inform either project about it until after the patches were accepted and deployed.
That caused build breakages in "linux beta distrubtions" and required both projects to deploy workarounds to mitigate the issue. So, we are obviously concerned given our goal of not introducing regressions.