> I would like an explanation of the purpose of restoring the 2013 > behavior. Did anyone request that?
No, nobody requested it. I just noticed that older GCC versions (e.g. 4.7.4) left the working directory clean after running the script, while current ones leave the archives behind, and it struck me as a bit sloppy - not a real problem, just untidy and a small waste of disk space. That was really the whole motivation. I hadn't seen the review thread you linked, so my framing of the 2013 change as accidental was wrong - thanks for the correction. Given Paolo's comment, I don't think there's a case for changing the default. I can rework the patch to keep the current default (archives kept) and just expose --save-archives/--no-save-archives as an explicit opt-in, for anyone who wants a supported way to clean up instead of manually doing `rm -f *.tar.*` in the directory. The directory-already-populated check (skip download/verify/extract if the target dir has content) would stay as-is either way, since that part doesn't touch the archive-deletion question. Separately, the partial-extraction issue I mentioned earlier is independent of all this - it exists in the script today regardless of what happens with archive cleanup - so it's probably worth fixing on its own merits either way. Let me know if that's worth sending as v2, or if this isn't wanted at all. Daniil
