Bug#677330: rred crashes in apt-get update
On Mon, Jul 2, 2012 at 8:56 PM, Courtney Bane
<debian-bugs-6752@cbane.org> wrote:
> I'm not positive that this is the exact same problem, but I'm still
> seeing segfaults in rred with apt 0.9.6. I rebuilt rred with debugging
> information before my latest crash, and was able to get a full
> backtrace. I've attached a file with the short and full backtraces. I've
[…]
> Here's the short backtrace:
>
>> Core was generated by `/usr/lib/apt/methods/rred'.
>> Program terminated with signal 11, Segmentation fault.
>> #0 *__GI___libc_free (mem=0x6d6c10) at malloc.c:3709
>> (gdb) bt
>> #0 *__GI___libc_free (mem=0x6d6c10) at malloc.c:3709
>> #1 0x00007f794c8312da in inflateEnd (strm=0x6b6dc8) at inflate.c:1260
>> #2 0x00007f794c835c79 in gzclose_r (file=0x6b6d50) at gzread.c:579
>> #3 0x00007f794d9bec06 in FileFd::Close (this=0x7fffccaeda00) at contrib/fileutl.cc:1675
>> #4 0x0000000000404278 in RredMethod::Fetch (this=0x7fffccaedf90, Itm=0x6b1c20) at rred.cc:503
>> #5 0x00007f794d9fc62d in pkgAcqMethod::Run (this=0x7fffccaedf90, Single=false) at acquire-method.cc:371
>> #6 0x0000000000402822 in main (argc=<optimized out>, argv=<optimized out>) at rred.cc:563
Are you sure about the version numbers involved?
The backtrace doesn't match with the code as FileFd::Close doesn't have
a gzclose() call itself anymore, but offloads it to FileFdPrivate::CloseDown.
This looks more like 0.9.3 which really had a gzclose call at line 1675.
Best regards
David Kalnischkies
Reply to: