libera/#sbcl - IRC Chatlog
Search
17:49:02
yitzi
We've been having a problem with SBCL 2.2.4 in Clasp's metabuilder koga. I have not been able to come up with a simple reproducer, but the error does not appear to on v2.2.3. If you move some code around sometimes it will go away. Here is a rather long output from the builder with the backtrace at the end. https://plaster.tymoon.eu/view/3145
17:50:23
yitzi
Does anybody have any idea about what changes there were in v2.2.4 that might have affected SB-C::%DO-FORMS-FROM-INFO or source line tracking? Or any ideas at all.
19:26:26
Krystof
ok, thanks. That's worth a mail report to sbcl-devel if you can, even if you don't have a clear reproducer
19:26:52
Krystof
we release monthly, so 2.2.5 will be in ~10 days (but obviously we need to fix something if it's broken)
19:27:17
Krystof
that is, we'll release 2.2.5 whether or not we have a fix for this. issue, so the closer you can get to a simple reproducer the better
19:36:07
stassats
yitzi: https://github.com/clasp-developers/clasp/blob/88492167e68af29e73040174441b86392590ed9a/src/koga/source.lisp#L83 not passing recursive-p would be a guess
22:15:56
stassats
so the finalizer for run-program has (sb-unix:unix-close fd), but if the fd is recycled it closes the wrong thing
22:16:20
stassats
for the whole make-fd-stream, but run-program has started finalizing its streams recently
22:17:08
drmeister
Keep talking like I follow what you are saying. yitzi will be back later - I'll point him to this discussion.