Re: fossil build failure on riscv64
> Granted, I have no clue what the package does, so I don't know what the consequences could be.
The fossil system basically combines git, apache, a wiki, a ticket
system, a mailing list server, and a threaded discussion server, in a
single executable. So it has a substantial network-facing attack
surface. And it's written in C.
Anyway, upstream generated a fix, which I just tested (takes hours
because riscv64 takes two orders of magnitude longer to compile) and
... the fix did not work. So this issue is not resolved. But upstream
is on it! Assuming a fix is forthcoming, and the issue is in fossil
and not GCC, I'll remove the riscv64 clang hot-wiring when I upload
the next upstream release. Or, if this issue pops up on other
platforms as they upgrade GCC, I'll cherry-pick the fix and remove the
clang hot-wire then. If the issue *is* in GCC, I'll generate a minimal
example and report it to the GCC project.
Reply to: