[#5737] returning strings from methods/instance_methods — TRANS <transfire@...>

I was just wondering why with #methods and #instance_methods, it was

11 messages 2005/09/08

[#5796] proposed attr writer patch — Daniel Berger <Daniel.Berger@...>

Hi all,

18 messages 2005/09/16

[#5798] Makefile error in OpenSLL extension (on Windows) — noreply@...

Bugs item #2472, was opened at 2005-09-16 18:56

11 messages 2005/09/17
[#5800] Re: [ ruby-Bugs-2472 ] Makefile error in OpenSLL extension (on Windows) — nobu.nokada@... 2005/09/17

Hi,

[#5851] Re: RubyGems in Ruby HEAD — Paul van Tilburg <paul@...>

Hi all,

34 messages 2005/09/21
[#5867] Re: RubyGems in Ruby HEAD — mathew <meta@...> 2005/09/21

Paul van Tilburg wrote:

[#5870] Re: RubyGems in Ruby HEAD — Marc Dequènes (Duck) <Duck@...> 2005/09/21

[#5920] Re: RubyGems in Ruby HEAD — mathew <meta@...> 2005/09/22

Marc Dequ竪nes (Duck) wrote:

[#5926] Re: RubyGems in Ruby HEAD — Pascal Terjan <pterjan@...> 2005/09/23

On 9/22/05, mathew <meta@pobox.com> wrote:

[#5931] Re: RubyGems in Ruby HEAD — Austin Ziegler <halostatue@...> 2005/09/23

On 9/23/05, Pascal Terjan <pterjan@gmail.com> wrote:

[#5898] Delegate and Forwardable Documentation — James Edward Gray II <james@...>

I've tried to send these files through a couple of times now with

17 messages 2005/09/22
[#5911] Re: Delegate and Forwardable Documentation — James Edward Gray II <james@...> 2005/09/22

On Sep 22, 2005, at 9:02 AM, James Edward Gray II wrote:

[#5924] Re: Delegate and Forwardable Documentation — James Edward Gray II <james@...> 2005/09/23

On Sep 22, 2005, at 11:53 AM, James Edward Gray II wrote:

[#5941] Re: Delegate and Forwardable Documentation — Yukihiro Matsumoto <matz@...> 2005/09/23

Hi,

[#5942] Re: Delegate and Forwardable Documentation — James Edward Gray II <james@...> 2005/09/23

On Sep 23, 2005, at 10:54 AM, Yukihiro Matsumoto wrote:

[#5947] Re: Delegate and Forwardable Documentation — Yukihiro Matsumoto <matz@...> 2005/09/23

Hi,

[#5921] Mutually dependent libs double loading. — TRANS <transfire@...>

I'm on Ruby 1.8.2.

14 messages 2005/09/23
[#5923] Re: Mutually dependent libs double loading. — Florian Gro<florgro@...> 2005/09/23

TRANS wrote:

[#5985] Finally an answer to my RubyGems question and some small suggestions — TRANS <transfire@...>

I appreciate those that attempted to offer me some info on this issue.

9 messages 2005/09/26

[#6001] Require Namepaces and RubyGems' effect on LoadPath problem — TRANS <transfire@...>

I've added namespaces to require. Works like this:

94 messages 2005/09/26
[#6002] Re: Require Namepaces and RubyGems' effect on LoadPath problem — Austin Ziegler <halostatue@...> 2005/09/26

On 9/26/05, TRANS <transfire@gmail.com> wrote:

[#6003] Re: Require Namepaces and RubyGems' effect on LoadPath problem — TRANS <transfire@...> 2005/09/26

On 9/26/05, Austin Ziegler <halostatue@gmail.com> wrote:

[#6005] Re: Require Namepaces and RubyGems' effect on LoadPath problem — Austin Ziegler <halostatue@...> 2005/09/26

On 9/26/05, TRANS <transfire@gmail.com> wrote:

[#6007] gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — Sam Roberts <sroberts@...> 2005/09/26

Quoting halostatue@gmail.com, on Tue, Sep 27, 2005 at 06:02:07AM +0900:

[#6013] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — Austin Ziegler <halostatue@...> 2005/09/27

On 9/26/05, Sam Roberts <sroberts@uniserve.com> wrote:

[#6014] Re: gems is a language change, not a pkging system — Sam Roberts <sroberts@...> 2005/09/27

Quoting halostatue@gmail.com, on Tue, Sep 27, 2005 at 10:29:17AM +0900:

[#6015] Re: gems is a language change, not a pkging system — James Edward Gray II <james@...> 2005/09/27

On Sep 26, 2005, at 8:54 PM, Sam Roberts wrote:

[#6016] Re: gems is a language change, not a pkging system — Sam Roberts <sroberts@...> 2005/09/27

Quoting james@grayproductions.net, on Tue, Sep 27, 2005 at 11:06:01AM +0900:

[#6018] Re: gems is a language change, not a pkging system — Austin Ziegler <halostatue@...> 2005/09/27

On 9/26/05, Sam Roberts <sroberts@uniserve.com> wrote:

[#6019] Re: gems is a language change, not a pkging system — Sam Roberts <sroberts@...> 2005/09/27

Quoting halostatue@gmail.com, on Tue, Sep 27, 2005 at 11:49:14AM +0900:

[#6024] Re: gems is a language change, not a pkging system — Austin Ziegler <halostatue@...> 2005/09/27

On 9/27/05, Sam Roberts <sroberts@uniserve.com> wrote:

[#6025] Re: gems is a language change, not a pkging system — Ralph Amissah <ralph.amissah@...> 2005/09/27

> Right now, they're watching people who have pretty much sat on the side

[#6026] Re: gems is a language change, not a pkging system — Austin Ziegler <halostatue@...> 2005/09/27

On 9/27/05, Ralph Amissah <ralph.amissah@gmail.com> wrote:

[#6043] Re: gems is a language change, not a pkging system — Ralph Amissah <ralph.amissah@...> 2005/09/28

I'll greatly weaken my post, and give everyone the opportunity to head me

[#6044] Re: gems is a language change, not a pkging system — Hugh Sasse <hgs@...> 2005/09/28

On Wed, 28 Sep 2005, Ralph Amissah wrote:

[#6073] Re: gems is a language change, not a pkging system — Mauricio Fern疣dez <mfp@...> 2005/09/28

Hello,

[#6074] Re: gems is a language change, not a pkging system — Jim Weirich <jim@...> 2005/09/29

On Wednesday 28 September 2005 07:35 pm, Mauricio Fern疣dez wrote:

[#6017] Re: gems is a language change, not a pkging system — Austin Ziegler <halostatue@...> 2005/09/27

On 9/26/05, Sam Roberts <sroberts@uniserve.com> wrote:

[#6046] Re: gems is a language change, not a pkging system — "Sean E. Russell" <ser@...> 2005/09/28

On Monday 26 September 2005 22:41, Austin Ziegler wrote:

[#6050] Re: gems is a language change, not a pkging system — Hugh Sasse <hgs@...> 2005/09/28

On Wed, 28 Sep 2005, Sean E. Russell wrote:

[#6207] Re: gems is a language change, not a pkging system — "Sean E. Russell" <ser@...> 2005/10/10

On Wednesday 28 September 2005 08:54, Hugh Sasse wrote:

[#6045] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — "Sean E. Russell" <ser@...> 2005/09/28

On Monday 26 September 2005 21:29, Austin Ziegler wrote:

[#6048] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — Austin Ziegler <halostatue@...> 2005/09/28

On 9/28/05, Sean E. Russell <ser@germane-software.com> wrote:

[#6059] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — Dominique Brezinski <dominique.brezinski@...> 2005/09/28

On 9/28/05, Austin Ziegler <halostatue@gmail.com> wrote:

[#6061] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — Austin Ziegler <halostatue@...> 2005/09/28

On 9/28/05, Dominique Brezinski <dominique.brezinski@gmail.com> wrote:

[#6062] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — Dominique Brezinski <dominique.brezinski@...> 2005/09/28

For what it is worth, I live life behind an authenticated proxy, so I

[#6099] Re: gems is a language change, not a pkging system (Re: Require Namepaces and RubyGems' effect on LoadPath problem) — "Sean E. Russell" <ser@...> 2005/09/30

On Wednesday 28 September 2005 08:43, Austin Ziegler wrote:

[#6009] Re: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault — Ralph Amissah <ralph.amissah@...>

(i) correction, segfault is with official ruby 1.8.3 (2005-09-21), not

21 messages 2005/09/27
[#6010] Fwd: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault — Ralph Amissah <ralph.amissah@...> 2005/09/27

[sorry for duplicate post]

[#6079] Re: Fwd: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault — ts <decoux@...> 2005/09/29

>>>>> "R" == Ralph Amissah <ralph.amissah@gmail.com> writes:

[#6081] Re: Fwd: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault — ts <decoux@...> 2005/09/29

>>>>> "t" == ts <decoux@moulon.inra.fr> writes:

[#6082] Re: Fwd: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault — Tanaka Akira <akr@...17n.org> 2005/09/29

In article <200509291419.j8TEJYid015419@moulon.inra.fr>,

Re: RubyGems in Ruby HEAD

From: Aredridel <aredridel@...>
Date: 2005-09-20 23:05:17 UTC
List: ruby-core #5840
> Right now, RubyGems represents a step backwards relative to Minero
> Aoki's setup.rb in many regards as far as repackagers are concerned.
> To state it simply, it is often more difficult to package something
> released in .gem format than in an equivalent tarball + setup.rb. I know
> this from my personal experience when repackaging a large number of
> libraries and applications [2] and most importantly from what several
> developers working on established repackaging efforts (Debian, FreeBSD,
> PLD, Suse) have told me. This is due to RubyGems breaking source
> compatibility with non-RubyGems system in several areas, including, but
> not limited to the following: 
> * lack of support for DATADIR 
> * obviously, require_gem 
> * the auto_require feature
> * in general, problems due to the new directory layout

Having Gems become standard as they are now is problematic for linux
distributors, because of the amount of effort to package them to play
nice with other software is high.

It's a break with backward compatibility to go changing require's
semantics if rubygems is automatically loaded, and if it isn't, we run
into more and more libraries and apps that won't work with non-gem
libraries.

Why not just use rubygems you might ask?

I maintain a small cluster ... I've got package management
infrastructure to roll out updates, with complete validation checks to
all the machines -- some seven hundred packages per machine, some fifty
of which are Ruby-related. Using two tools with vastly different mode of
operation would make the task of rolling out Ruby apps to those servers
an onerous one, and the existing tools work perfectly.

It's mostly the breaking API that bothers me: changing require's
semantics, and adding an ever-growing list of libraries that only work
with gems, thereby requiring all prerequisites to /also/ be gems, is
making packaging ruby libraries up efficiently problematic. Packaging
escapes the realm of ruby, and has tentacles into the rest of the
system.

With RMagick as a gem, and ImageMagick installed with RPM, if I upgrade
ImageMagick, RMagick breaks. With RMagick as an RPM as well, RPM throws
a dependency warning, and I can go fix the problem, instead of
discovering that my image thumbnailing application has been broken for
days because of the broken dependencies. Fixing that is out of scope for
rubygems -- it just won't happen, it cannot happen.

Rubygems works alright for pure-ruby packages, most of the time. The
edge cases, however, are not rare, and have some grave problems. I would
like to see those problems addressed, or at least make it so that those
who maintain projects whose scope does include non-ruby libraries can
make that happen.

Aredridel
Ruby packager for PLD-Linux (http://pld-linux.org)



In This Thread

Prev Next