Hi, I've finally managed to build a 2.6d bdb with full debug symbols, but when I run it, no code seem to be executed, the segfault occurs before _start, as is confirmed by "ldd -r":
$ ldd -r ./bin/bdb libbigloobdl_s-2.6d.so => /home/ydirson/bigloo-2.6d-32-O1-g/lib/2.6d/libbigloobdl_s-2.6d.so (0x4001d000) libbigloo_s-2.6d.so => /home/ydirson/bigloo-2.6d-32-O1-g/lib/2.6d/libbigloo_s-2.6d.so (0x40054000) libbigloogc-2.6d.so => /home/ydirson/bigloo-2.6d-32-O1-g/lib/2.6d/libbigloogc-2.6d.so (0x401ce000) libdl.so.2 => /lib/libdl.so.2 (0x401fe000) libm.so.6 => /lib/libm.so.6 (0x40208000) libc.so.6 => /lib/libc.so.6 (0x40282000) libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x403a7000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000) /usr/bin/ldd: line 1: 31549 Segmentation fault LD_TRACE_LOADED_OBJECTS=1 LD_WARN=yes LD_BIND_NOW=yes LD_VERBOSE= "$file" Could that be a ld.so issue, maybe triggered by unexpected things in ELF headers ? On Fri, Jun 04, 2004 at 12:04:56AM +0200, Yann Dirson wrote: > On Tue, Jun 01, 2004 at 11:34:36AM +0200, I wrote: > > I'll try to build at -O0, and then at -O1 if -O0 worked, and will let you > > know. I should also have a closer look at how gcc-3.2 performs, but that > > would be a pain to have to use it, since it's supposed to be phased out from > > Debian... > > Here is my status of the issue: > > - the problem still exists with "gcc-3.3 -O0" > - a diff of an ARM build tree against an i386 one shows no difference > (appart from comments and strings showing setup information) > > > The 1st point rules out a possible bug in a gcc optimisation; the 2nd one > rules out possible problems impacting the scheme->C part of the bigloo > compiler's job. > > Unless I miss something, that leaves us with the following possibilities: > > - problem in the bigloo executable after C generation (I don't know enough > here to rule out this possibility, but intuitively I don't give it much > weight) > > - generated C code has arm-specific portability issues - Philip mentionned > structure alignment as a common problem > > - gcc-3.[34] at least has a non-optim-related arm-specific bug > > > Next step: I've launched a build with --cflags=-g --coflags=-O1 and will try > to pinpoint the precise problem. Stay tuned. > > -- > Yann Dirson <[EMAIL PROTECTED]> | Why make M$-Bill richer & richer ? > Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: > Pro: <[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity > http://ydirson.free.fr/ | Check <http://www.debian.org/> -- Yann Dirson <[EMAIL PROTECTED]> | Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Freedom, Power, Stability, Gratis http://ydirson.free.fr/ | Check <http://www.debian.org/>