On Tue, 2022-12-27 at 16:58 +0100, Bruno Haible wrote: > But there are other new failures, in features/output-sync and > variables/MAKEFLAGS.
Ugh, the MAKEFLAGS issue is because Solaris "false" insists on exiting with an exit code 255 rather than exit code 1 (I realize the exact exit code is not standardized but pretty much everyone else uses 1). I've run into this before but forgot to check Dmitry's changes. This is easy to fix. The output-sync issue I thought I already fixed but apparently my changes were not sufficient. I'm going out of town for a few days so I'll look into it this weekend.
