On Mon, Jun 07, 2010 at 10:25:51PM +0200, Dominique Dhumieres wrote:
> > I won't disable gcj in fink's libtool yet, I'd like to see if I get more
> > reports of failures.
> 
> What is working is to use gcc44.

Is this the same issue of gcj in gcc 4.5.0 not compiling java under
darwin9? If so, gcc 4.4.x works under darwin9 because it is using the unwinder
in its own libgcc whereas in gcc 4.5.0 the system unwinder is now always
used. We fixed the problem with darwin10 (in that case an unwinder call was
noop'ed in the libSystem unwinder) but haven't debugged the failure
under darwin9's libgcc unwinder. If I understand this properly, the
darwin9 unwinder is a modified FSF libgcc unwinder (derived from 4.2.1)
whereas the darwin10 unwinder uses a compatibility unwinder for non-compact
unwind info. So they are completely different beasts.
              Jack
> 
> > Can you check that it also hangs without -fno-common, then run the same
> > thing under gdb, wait a while, interrupt it and get a backtrace and file
> > a bug.
> 
> Without any flag the backtrace depends on how long I wait:
> (1) medium time 
> 
> . done
> ^C
> Program received signal SIGINT, Interrupt.
> 0x95174d4c in __register_frame ()
> (gdb) bt
> #0  0x95174d4c in __register_frame ()
> #1  0x9517531c in _Unwind_Find_FDE ()
> #2  0x951720c8 in _Unwind_GetIPInfo ()
> #3  0x95173754 in _Unwind_Backtrace ()
> #4  0x0373ccb4 in _Jv_StackTrace::GetStackTrace ()
> #5  0x03782094 in java::lang::VMThrowable::fillInStackTrace ()
> #6  0x03b334f4 in java::lang::Throwable::fillInStackTrace ()
> #7  0x03b3362c in java::lang::Throwable::Throwable ()
> #8  0x03b341f0 in java::lang::ClassNotFoundException::ClassNotFoundException 
> ()
> #9  0x03779764 in java::lang::Class::forName ()
> #10 0x03b7a1dc in java::net::URLClassLoader::addURLImpl ()
> #11 0x03b7a83c in java::net::URLClassLoader::addURL ()
> #12 0x037976f8 in gnu::gcj::runtime::SystemClassLoader::init ()
> #13 0x03730228 in _Jv_CreateJavaVM ()
> #14 0x03730354 in _Jv_RunMain ()
> #15 0x00001ee4 in start ()
> 
> (2) longer time (probably 1.5Gb eaten)
> 
> . done
> GC Warning: Repeated allocation of very large block (appr. size 1048576000):
>       May lead to memory leak and poor performance.
> ^C
> Program received signal SIGINT, Interrupt.
> 0xffff9374 in ?? ()
> (gdb) bt
> #0  0xffff9374 in ?? ()
> #1  0x0373caec in _Jv_StackTrace::UnwindTraceFn ()
> #2  0x9517377c in _Unwind_Backtrace ()
> #3  0x0373ccb4 in _Jv_StackTrace::GetStackTrace ()
> #4  0x03782094 in java::lang::VMThrowable::fillInStackTrace ()
> #5  0x03b334f4 in java::lang::Throwable::fillInStackTrace ()
> #6  0x03b3362c in java::lang::Throwable::Throwable ()
> #7  0x03b341f0 in java::lang::ClassNotFoundException::ClassNotFoundException 
> ()
> #8  0x03779764 in java::lang::Class::forName ()
> #9  0x03b7a1dc in java::net::URLClassLoader::addURLImpl ()
> #10 0x03b7a83c in java::net::URLClassLoader::addURL ()
> #11 0x037976f8 in gnu::gcj::runtime::SystemClassLoader::init ()
> #12 0x03730228 in _Jv_CreateJavaVM ()
> #13 0x03730354 in _Jv_RunMain ()
> #14 0x00001ee4 in start ()
> 
> Dominique

------------------------------------------------------------------------------
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
_______________________________________________
Fink-users mailing list
Fink-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fink-users

Reply via email to