On Thu, 22 Jan 2026, Maciej W. Rozycki wrote: > In any case I've built a fresh version of the compiler with no issue and > regression-testing is under way. As I recall it took ~72h to complete on > the last occasion, so it'll take a while yet and then another run will be > required with your change applied.
Regression-testing has completed, taking ~64h on this occasion with each of the two iterations. There are 3 new failures and a couple of failures removed, all of which I suspect are intermittent, but I'll double-check yet. Also I have realised the target's limited storage was being clogged with core dumps, the production of some of which also hogged the system due to their excessive size. I have disabled core file generation for the second run (with your patch applied), as they're not needed by default and just waste the limited target's resources, so this run went a little more smoothly. I have posted test results from the original run at: <https://inbox.sourceware.org/gcc-testresults/[email protected]/>. Sadly there are a number of regressions across the VAX target-specific tests, including a bunch of ICEs. Overall the results are so-so to say the least, but that's no news and not a big change from GCC 15. The big issue is of course the non-functional exception unwinder. I'll have look into your testcase in the coming days and try to address the target-specific regressions before the release as well. Maciej
