hoodwink.d enhanced
RSS
2.0
XHTML
1.0

RedHanded

Why I Use Debian With Ruby #

by daigo in cult

Some debian developers discussed the split of Ruby standard libraries and gave a proposal to add two new packages and change the meaning of the current ‘ruby’ package at the debian-ruby list (not archived yet) (I need to modify dependency lines in my tDiary packages). I am sure that it will soon be available and I’d like to say “Wait a minute, it is too early to leave Debian.”

I like Debian’s Ruby because Ukai-san and Yamada-san are members of its maintainers. They are super hackers of Debian and Ruby. Ukai-san is the leader of Debian.jp and the president of Japan Linux Association. Yamada-san seems to maintain Vine’s Ruby also, however, make Debian’s package at first.

See the changelog.Debian.gz at /usr/share/doc/ruby1.8 in Debian. You can find Debian’s Ruby more stable than the original. I appreciate their contribution. I cite a recent entry. Who else can do that?

ruby1.8 (1.8.2-1) unstable; urgency=high

  * akira yamada <akira@debian.org>
  - new upstream version, 1.8.2:
      - removed debian/patches/{100_cvs_updates.patch,801_syck_segv.patch,
        802_syck_segv.patch}.  they ware included 1.8.2.
      - (urgency high) fixed segv bugs:
          - Process.groups= dumps core again.  [ruby-dev:25285]
          - YAML::Syck::Parser#load dumps core.  [ruby-core:03973]
          - IO.select dumps core.  [ruby-dev:25312]
      - fixed bugs:
          - prohibit to change access mode for special IO ports  (stdin,
            stdout and stderr).  [ruby-dev:25225]
  - added debian/patches/100_cvs_updates.patch:
      - (urgency high) fixed segv bugs:
          - String#center dumps core.  [ruby-dev:25341]
          - require "openssl" dumps core.  [ruby-dev:25325]
          - Zlib memory leak.  [ruby-list:39235][ruby-dev:25309]
          - String#ljust dumps core again.  [ruby-dev:25367]
          - bmcall() causes core dump.  [ruby-dev:25366]
          - Module#autoload? dumps core.  [ruby-dev:25373]
          - OpenSSL::Netscape::SPKI#challenge= dumps core.  [ruby-dev:25359]
          - Segfault in Thread#initialize / caller  [ruby-core:04067]
      - fixed bugs:
          - webrick/httpauth bugs.
            [ruby-list:40467][ruby-list:40482][ruby-dev:25336]
          - rand uniformity  [ruby-dev:25396]

 -- akira yamada <akira@debian.org>  Fri,  7 Jan 2005 11:17:10 +0900
27 Jan 2005 at 09:37 | 6 comments

Paul van Tilburg

said on 27 Jan 2005 at 10:36

I wanted say that I too really appreciate the work of Yamada and Ukai. The packages are always in a really good state and bugs are resolved swiftly. Our proposol is not to counter their work at all, but I feel since Ruby popularity seems to be growing, we need more involvement and communication. A Debian Ruby team won’t be a bad idea, IMO .

why

said on 27 Jan 2005 at 11:04

“Wait a minute, it is too early to leave Debian.”

Good stuff, daigo.

I’m so glad you posted this. I need to merge that 801_syck_segv.patch pronto.

bogbog

said on 27 Jan 2005 at 12:03

So many people bashing Debian lately, and insulting the maintainers. This is so sad. And as usual, some of the people protesting the most are the most clueless ones.

This is the fate of the packager; if it works, nobody will know. If there’s something wrong, there will be no end to complaints. In many ways, life is much easier… upstream. Some jobs are so thankless.

Austin

said on 28 Jan 2005 at 10:03

Clueless? No, bogbog, I’m not clueless on this. I personally think that the packagers for Ruby on Debian have made life harder for themselves, for Ruby, and for users by making the decisions that they’ve made. As I said on ruby-talk, the problem is that there are real users who are experiencing real problems because Ruby is so badly subdivided; there are certain expectations of the “Ruby environment”, and any breaking of those expectations by others is a bad thing.

The packager who repackages my work had best be prepared to catch hell if they do something that makes me, the developer, look bad. This is precisely what has happened to the developers of RubyGems and Rails recently because of the bad packaging decisions made by the Debian packagers. Great for making it available; now make it usable.

lypie

said on 31 Jan 2005 at 04:59

i’ve mailed the deb ruby maintainer without response in several days. the ubuntu guys otoh gave me a decent reply on irc in minutes even though their official policy is python usage. oh well. and calling some of the more respected members of the ruby community clueless. well that tells me a lot about debian users. i’m glad to no longer be running it =) otoh maybe the guy is just busy ;) but lets ignore that for a second, i wanna rant!

slumos

said on 29 Jul 2005 at 16:12

“I cite a recent entry. Who else can do that?”

I don’t mean to sound harsh, but did it occur to you to actually research the question before you wrote?

For example, knu-san is a FreeBSD committer and ports collection maintainer for Ruby and about a gob-zillion modules.

Is anyone actually surprised that Ruby luminaries would end up maintaining Ruby on whatever their prefered platform?

Incidentally, knu is also responsible for some truly excellent FreeBSD utilities which have caused Ruby to show up on a lot more FreeBSD systems than it would have otherwise. Heck, if it was a little more difficult to install Ruby on FreeBSD, portupgrade(1) alone would probably cause it to get into the base system.

(Yes, I’m replying late. What you get for showing up on a google search for “ruby autoload”. :-)

Comments are closed for this entry.