IRC channel logs
2026-10-30.log
back to list of logs
<aleksi>+1, I'm very excited to try out 0.28 on my end as well! <aleksi>On that topic, are there plans to cut a new version of stage0-posix? 1.9.1 comes with m2-planet 1.13.0 which is too old for the newest Mes <aleksi>Or maybe something else is funny with my build setup (I'm trying to update nixpkgs's bootstrap.) I'm hitting a segfault when trying to run hello world in mes compiled with m2. <aleksi>My bad, I've just found my issue. In nixpkgs we had patched mes to use MES_STACK of 6 million. This caused an assert fail, and __assert_fail segfaulted before it could print the message <ekaitz>aleksi: that's an old thing. We don't do that anymore. <ekaitz>Maybe that should be on the Changelog <ekaitz>you should tell the maintainers they should have said it <aleksi>Ahh no worries, so I suppose it was even more widespread <ekaitz>did you manage to run it and test it? <aleksi>Yes, mes was able to run hello world and mescc. I got a mescc error when compiling crt, so in the morning I'll be debugging that <stikonas>well, I'll see regarding stage0-posix, I could try but not sure when I'll have time, might be today, might be much later (3 weeks or so) <ekaitz>aleksi: hmmm what kind of error? <matrix_bridge><Jeremiah Orians> stabbles: it is fine to move untar to its own page; I am probably the only person who is going to care about those results (as I will be looking to fix all of the strictness bits when I get some free time) <matrix_bridge><Jeremiah Orians> And the M2-Planet/mes.c gap has long been the most sensitive step in the whole chain. As subtle changes in compiled output can easily have big knock on changes in the interpretation. <matrix_bridge><Jeremiah Orians> If nothing else when we see anyone repeat the 357 number, we know that they didn’t actually look at the details.