On Sun, Mar 21, 2004 at 09:17:04AM +0900, GOTO Masanori wrote: > > I'm providing a gdb backtrace below on the same sparc32 box. > > Is this gdb result on auric? You wrote your backtrace was executed on > "auric.debian.org" in the previous mail. But look at > http://db.debian.org/machines.cgi. "auric.debian.org" is woody. No -- it was my SPARCstation 10. > > Program received signal SIGBUS, Bus error. > > 0x5004dd68 in __gconv_get_alias_db () from /lib/libc.so.6 > > (gdb) bt > > #0 0x5004dd68 in __gconv_get_alias_db () from /lib/libc.so.6 > > #1 0x501bdca8 in gconv () from /usr/lib/gconv/ISO8859-1.so > > #2 0x5004729c in iconv_close () from /lib/libc.so.6 > > Is this gdb result on woody machine, not sid? It was sid. > > en_US.ISO-8859-1 > > $ echo $LC_ALL > > en_US.ISO-8859-1 > > > > I'm not sure what you mean by "which strings did you test". > > What strings did you give to iconv(3)? What procedures do we > reproduce this bug? Beats me... but see below for how I encountered the bug. > I would like to know how to reproduce this bug. Please provide us the > sample program if you can. If ISO-8859-1 gconv has bug, then I guess > many users also get the problem. OK. Well it's a bit convoluted, but here is exactly what I did to encounter the bug. $ cvs -d :pserver:anonymous@cvs.alioth.debian.org:/cvsroot/d-i co debian-installer/build [...] $ cd debian-installer/build $ dpkg-checkbuilddeps [install build-deps] $ fakeroot make build_miniiso It will get to running mkisofs on a tree directory at which point it will SIGBUS. If you would like a shell on my SPARCstation talk to me on IRC. -- Joshua Kwan
Attachment:
signature.asc
Description: Digital signature