tcltk: proposal to replace with 1_8 branch head

Mamoru Tasaka mtasaka at ioa.s.u-tokyo.ac.jp
Tue May 11 17:33:15 UTC 2010


Hello, all:

Currently there are at least 2 bug reports against ruby tcltk module which
say that even some simple ruby scripts using ruby tcltk cause ruby crash:
- https://bugzilla.redhat.com/show_bug.cgi?id=560053
- https://bugzilla.redhat.com/show_bug.cgi?id=590503
Also I tried /usr/share/doc/ruby-tcltk-1.8.6.399/tk/sample/demos-en/widget
but this also causes segfault. So it seem that currently Fedora's tk support
on ruby is completely broken.

We now use ruby 1.8.6.x branch, however looking at
http://svn.ruby-lang.org/cgi-bin/viewvc.cgi/branches/ruby_1_8_6/ext/tk/
ext/tk codes for 1.8.6.x branch are generally 2 years old and it looks that
these codes are not so much maintained. So instead of using 1.8.6.x branch
ext/tk, I propose to replace ext/tk tree with that of ruby_1_8 branch, i.e.
http://svn.ruby-lang.org/cgi-bin/viewvc.cgi/branches/ruby_1_8/ext/tk/
as suggested by one of the ext/tk tree authors:
http://www.dumbo.ai.kyutech.ac.jp/nagai/RubyTk/?%BA%C7%BF%B7%C8%C7%A4%CERuby%2FTk%A4%F2%BB%C8%A4%A6%A4%CB%A4%CF
(note: the above url is in Japanese).

For now I tried to use 1.8.6p399 tarball + ext/tk directory of ruby_1_8 branch head
and the trial rpms are:
For F-14:
http://koji.fedoraproject.org/koji/taskinfo?taskID=2180611
http://koji.fedoraproject.org/scratch/mtasaka/task_2180611/
For F-13:
http://koji.fedoraproject.org/koji/taskinfo?taskID=2180618
http://koji.fedoraproject.org/scratch/mtasaka/task_2180618/
For F-12:
http://koji.fedoraproject.org/koji/taskinfo?taskID=2180631
http://koji.fedoraproject.org/scratch/mtasaka/task_2180631/

It seems that with these rpms the above 2 bugs are fixed and 
/usr/share/doc/ruby-tcltk-1.8.6.399/tk/sample/demos-en/widget works
( /usr/share/doc/ruby-tcltk-1.8.6.399/tk/sample/demos-jp/widget has
  some easy iconv related mistakes and need fixing ).
I want to push these rpms if there are no terrific errors on these
rpms. Any suggestions or opinions are welcomed.

Regards,
Mamoru



More information about the ruby-sig mailing list