Bug 145630 - aspell --mode=sgml loops endlessly
Summary: aspell --mode=sgml loops endlessly
Alias: None
Product: Red Hat Enterprise Linux 3
Classification: Red Hat
Component: aspell
Version: 3.0
Hardware: i686
OS: Linux
Target Milestone: ---
Assignee: Ivana Varekova
QA Contact: Ben Levenson
Depends On:
Blocks: 187539
TreeView+ depends on / blocked
Reported: 2005-01-20 09:16 UTC by Peter Collinson
Modified: 2007-11-30 22:07 UTC (History)
4 users (show)

Clone Of:
Last Closed: 2006-05-03 21:12:37 UTC

Attachments (Terms of Use)

External Trackers
Tracker ID Priority Status Summary Last Updated
Red Hat Product Errata RHBA-2006:0310 normal SHIPPED_LIVE aspell bug fix update 2006-07-19 18:25:00 UTC

Description Peter Collinson 2005-01-20 09:16:35 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4.3)
Gecko/20050104 Red Hat/1.4.3-3.0.7

Description of problem:
aspell check index.html
does not to terminate. The memory footprint grows and grows. Running
strace shows that it's in some loop trying to allocate memory.

aspell --mode=none check index.html
will operate correctly.

Trying LANG=C aspell check index.html
to remove possible locale problems has the same effect.

(I don't see an easy way of reverting to a previous aspell version,
too many dependencies).

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1. aspell check anyfile.html

Actual Results:  Nothing. It sits and waits.

Expected Results:  A spell check screen should appear.

Additional info:

Comment 1 Scott O'Neil 2005-04-18 21:29:30 UTC
I'm also able to recreate this problem using aspell- I use the
following command:

 /usr/bin/aspell -a -d american  -H

This is actually a pretty serious problem. If anyone uses Horde/IMP webmail on
my server, and asks Horde/IMP to do Spell Check, it uses these options and this
causes my server to crash! This seems like a simple way to crash a server remotely.


Comment 2 Scott Russell 2005-06-28 15:56:00 UTC
Also confirmed in RHEL 3 i386 AS U5 with aspell- installed.

Comment 3 Matthias Schroder 2005-10-04 16:43:21 UTC
In the strace logfile we found the suspicious line 

open("/usr/share/aspell/SGML&iso8859-1.map", O_RDONLY) = -1 ENOENT (No such file
or directory)

I assume it meant to open /usr/share/aspell/SGML.map ?

Still strange that it does not handle this gracefully.


Comment 4 Dmitry S. Makovey 2005-12-09 17:10:14 UTC
this is a serious bug disabling quite a bit of software (moodle is one example)
from working properly because it operates with HTML documents and "-H" (or
--mode=sgml) is essential... 
It's quite surprising that it haven't been adressed for longer time

Comment 5 Ivana Varekova 2005-12-15 15:06:19 UTC
Thank you for your comments. 
The test version (aspell-, which should fix this bug is on
http://people.redhat.com/varekova/aspell/aspell- (on
http://people.redhat.com/varekova/work.html there are i386 sources too). Could
you please test this version?
Ivana Varekova

Comment 6 Matthias Schroder 2005-12-15 15:54:19 UTC
The test aspell works for me. Funny thing is that I had tried the same fix, but
my version failed misserably.

Comment 7 Ivana Varekova 2005-12-16 09:26:09 UTC
Thank you for testing Matthias. Your suggestion in comment 3 is very close to
the patch in attached test version. 

Comment 8 Scott Russell 2005-12-16 13:41:17 UTC
Another confirmation that the test package solves the problem on RHEL3 AS U6
i386. I thought I saw this issue on RHEL4 AS U3 i386 but I couldn't replicate it
so I didn't test the new package there.

Thanks for getting this fixed!

Comment 9 Peter Collinson 2005-12-17 11:45:05 UTC
This fixes it for me too.. sorry I didn't respond more quickly.. 

Comment 10 Ivana Varekova 2006-01-02 15:36:31 UTC
Thank you for your tests.

Comment 20 Red Hat Bugzilla 2006-05-03 21:12:38 UTC
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.


Note You need to log in before you can comment on or make changes to this bug.