On Tue 2013-01-08 22:16:30 -0500, Elliott Mitchell wrote: > First item, are bugs #650241 and #694827 really distinct from each other? > They read suspiciously similar to each other...
they do seem similar, i agree. and i wrote them both, nearly exactly one year apart :P The main distinction is that 650241 refers to problems running from a non-multiarch host to within a multiarch fakechroot, while in 694827, both host and fakechroot are multiarch. It's certainly possible that they could both be resolved by the same fix, though. I would have expected the latter bug to be simpler to fix than the former (since the latter bug deals with the same libc6 on both systems. > Next, this reads like a problem of differing behavior between versions of > ld-linux.so.2. As such, both of these might in fact be more facetes of > #650234. I think #650234 is a different bug from the other two. > Could you try the command: > > strace -s 4096 -f fakeroot fakechroot chroot root /lib/ld-linux.so.2 > /usr/bin/wget -q -O/dev/null https://localhost:1346/ 2>&1 | grep gnutls.so i'm testing this on an amd64 machine now so /lib/ld-linux.so.2 is /lib64/ld-linux-x86-64.so.2, so i've adjusted your command. But here's the output: 127 dkg@alice:~/tmp$ fakeroot fakechroot chroot root /lib64/ld-linux-x86-64.so.2 /usr/bin/wget -q -O/dev/null https://localhost:9999/ 4 dkg@alice:~/tmp$ strace -s 4096 -f fakeroot fakechroot chroot root /lib64/ld-linux-x86-64.so.2 /usr/bin/wget -q -O/dev/null https://localhost:9999/ 2>&1 | grep gnutls.so [pid 29587] open("/usr/lib/x86_64-linux-gnu/libfakeroot/libgnutls.so.26", O_RDONLY) = -1 ENOENT (No such file or directory) [pid 29587] open("/home/dkg/tmp/root/usr/lib/libgnutls.so.26", O_RDONLY) = -1 ENOENT (No such file or directory) [pid 29587] open("/home/dkg/tmp/root/lib/libgnutls.so.26", O_RDONLY) = -1 ENOENT (No such file or directory) [pid 29587] open("/usr/lib/x86_64-linux-gnu/libgnutls.so.26", O_RDONLY) = 3 0 dkg@alice:~/tmp$ find root -name libgnutls.so.26 root/usr/lib/x86_64-linux-gnu/libgnutls.so.26 0 dkg@alice:~/tmp$ > If this includes the extra paths you were expecting from the command in > #694827, then this literally is the same bug as #650234. I'm wondering > how many other fakechroot bugs would disappear with the solution to > #650234; #650234 is actually a rather severe issue... If i understand your suggestion correctly, this appears to demonstrate that #650234 is a distinct bug from the other two. --dkg
pgpwcL0r9rJKR.pgp
Description: PGP signature
