I'm testing swig-1.3.27 on an Athlon T-bird (though I think the arch is irrelevant). I've tried to emerge it three times, and each time it locked up X-windows and I had to reboot. I'm going to see if I can capture more information by trying it in a console window. I did that once, and "top" showed the "configure" step well over 1 GB virtual! There's something really bizarre going on here. 1.3.25 compiles fine. Check back tomorrow; I'm going to try to capture the "top" output so I can see what's happening. Unless this is a known problem, of course.
Created attachment 76153 [details] Log of the failed emerge This is the log of the emerge that fails ("lua" USE flag set)
Created attachment 76154 [details] Log of "emerge info" after the crash
Created attachment 76155 [details] Log of "top" during the failed emerge I ran "top" to a log file to capture the process sizes during the emerge. This should show which process has a memory leak.
I know approximately what is going on. The crash only happens when the "lua" USE flag is set. Somewhere in the "configure" step for lua, some process develops a memory leak and grows to the point where the kernel kills it. I think it's the "configure" process, but for some reason it didn't show up in the log file from "top". I'll try to capture the process details with some other tools tonight if necessary. Meanwhile, turning off the "lua" option is a viable workaround.
By the way ... there's another issue with the ebuild: QA Notice: USE Flag 'ocaml' not in IUSE for dev-lang/swig-1.3.27
From "Log of failed emerge": !!! Please attach the config.log to your bug report: !!! /var/tmp/portage/swig-1.3.27/work/swig-1.3.27/config.log Also, you appear to be testing swig in a full stable system, and it's not working at lua. I'd try putting lua with testing keywords (echo "dev-lang/lua >> /etc/portage/package.keywords) and seeing if it makes any difference. Otherwise I'd probably mark the bug as LATER, seeing as how 1.3.27 is going to be awhile before it's even considered for stable (Heck, I'm still working out the 1.3.25 issues to get it into stable).
*** Bug 119387 has been marked as a duplicate of this bug. ***