Le 5 déc. 04, à 23:25, Dan Nelson a écrit :
In the last episode (Dec 05), Xavier Beaudouin said:
Seems as usual that pike is *very* difficult to port to these platforms. Can some people have a look ?
... Or if you cannot then pike will be marked as broken on these.
The first two builds seem to just have a bad PLIST:
----- Forwarded message from User Ports-ia64 ports-ia64@FreeBSD.org ----- ===> Building package for pike72-7.2.570 tar: pike/7.2.570/lib/modules/Parser.pmod/_parser.so: Could not stat: No such file or directory
----- Forwarded message from User Ports-amd64 ports-amd64@FreeBSD.org ----- ===> Building package for pike72-7.2.570 tar: pike/7.2.570/lib/modules/Parser.pmod/_parser.so: Could not stat No such file or directory
, either that, or pike isn't dumping its modules for some reason.
Humm... I asking myself why _parser.so don't exist on amd64 / i64, but exist on sparc64 and i386 ???
----- Forwarded message from User Ports-alpha ports-alpha@FreeBSD.org ----- /tmp/a/ports/lang/pike72/work/pike-7.2.570/src/treeopt.in (method=QQ) Fatal error 'longjmp()ing between thread contexts is undefined by POSIX 1003.1' at line ? in file /a/portbuild/alpha/4/src/lib/libc_r/uthread/uthread_jmp.c (errno = ?) Abort trap (core dumped)
This is a problem; getting a thread dump from that corefile might help to determine which two threads are involved here. Too bad libc_r doesn't print the thread ids in the error message. It might also be caused by pike using a different stack, which I don't think is the case.
How can we fix that ?
/Xavier -- Xavier Beaudouin - Unix System Administrator & Projects Leader. President of Kazar Organization : http://www.kazar.net/ Please visit http://caudium.net/, home of Caudium & Camas projects