Red Hat Bugzilla – Bug 126314
"s", "m", or "u" in ja_JP causes segfault
Last modified: 2007-11-30 17:10:44 EST
Description of problem: When not running in the "C" locale (especially a multibyte locale like ja_JP.UTF-8) Version-Release number of selected component (if applicable): info-4.6-3 How reproducible: 50%. Keep doing an "s" and you'll eventually segfault it... Steps to Reproduce: 1. run FC2 in a Japanese locale. (confirm with "locale(1)" command) 2. find a page with lots of text/menus on it (libc's index is good) 3. keep using "s" command until it segfaults. Additional info: Have not tested with other multibyte locales, but I suspect that any multibyte locale is vulnerable to this bug.
actually, not just the "s" command... decending into sub-trees and ascending enough times will also cause segfaults.
Haven't been able to reproduce this. I built 4.7 earlier on -- does that have the same problem? Can you show me a stack trace?
4.7 seems to work ok (rebuilt from fc3 src rpm) fc2 with UTF-8 japanese. Nice work.
I wonder if this warrants an update then -- another incentive is that mainline gcc doesn't build with texinfo-4.6 any more. Thanks for testing.
considering that info-4.6 in its current state is not very usable with Japanese (and I'm willing to bet Chinese and Korean too), yes, I'd recommend an update to 4.7.