[DRE-maint] Bug#443409: libxml-ruby1.8: Segfault when parsing file

[Ted Reed]

Package: libxml-ruby1.8
Version: 0.3.8-1
Severity: important

I get a consistent segfault when attempting to parse a file.

Code to reproduce:

require 'xml/libxml'
doc = XML::Document.file('jbo-en.xml')

It should be noted that this file is over a megabyte in size. I can make
it available via the web upon request. I cannot get a segfault when testing with
other XML files, so there's something particular about this file that's
causing the segfault. I attempted an strace, the most relevant part I
reproduce here:

read(3, "ord=\"caution\" valsi=\"o\'i\" />\n\n<n"..., 4096) = 4096
brk(0xf72000)                           = 0xf72000
read(3, " word=\"robe\" valsi=\"pastu\" />\n\n<"..., 4096) = 4096
read(3, "lword word=\"sea lion\" valsi=\"pin"..., 4096) = 4096
brk(0xf93000)                           = 0xf93000
read(3, "re and during\" valsi=\"pujeca\" />"..., 4096) = 4096
brk(0xfb4000)                           = 0xfb4000
read(3, "\" sense=\"cause to begin again\" v"..., 4096) = 4096
close(3)                                = 0
munmap(0x2acfa33af000, 4096)            = 0
--- SIGSEGV (Segmentation fault) @ 0 (0) ---

It also seems to be particular to the Ruby bindings, as the Python
bindings to libxml2 have no such problem. (I can open the file just
fine, but the Python bindings are... ugly.)

-- System Information:
Debian Release: 4.0
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'stable')
Architecture: amd64 (x86_64)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)

Versions of packages libxml-ruby1.8 depends on:
ii  libc6                      2.3.6.ds1-13  GNU C Library: Shared libraries
ii  libruby1.8                 1.8.5-4       Libraries necessary to run Ruby 1.
ii  libxml2                    2.6.27.dfsg-1 GNOME XML library
ii  zlib1g                     1:1.2.3-13    compression library - runtime

libxml-ruby1.8 recommends no packages.

-- no debconf information