From thewoolleyman at gmail.com  Tue Jan  1 20:17:44 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Tue, 1 Jan 2008 18:17:44 -0700
Subject: [Rubygems-developers] Should binary platform gems be auto-selected
	before ruby platform on windows?
Message-ID: <e3af97510801011717r1fac11d6nf6850f2b47c0ed63@mail.gmail.com>

I have some test gems.  There is a dependent gem [1] which depends on
another dependency gem.  The dependency gem has two platforms, one
default ruby [2], and one with a platform of 'mswin32' [3].

When I install the dependent gem [1], it selects the ruby platform
dependency [2] instead of the mswin specific platform dependency [3].

Is this the desired behavior?  I sort of thought that a binary
platform gem should be preferred over a plain-ruby platform, if it
exists.  Don't some multi-platform gems sometimes require compilation
of a 'ruby' platform, but also offer a precompiled binary as well?

Thanks,
-- Chad

[1] http://geminstaller.rubyforge.org/svn/trunk/spec/fixture/gems/specifications/dependent-stubgem-depends-on-multiplatform-1.0.0.gemspec

[2] http://geminstaller.rubyforge.org/svn/trunk/spec/fixture/gems/specifications/stubgem-multiplatform-1.0.1.gemspec

[3] http://geminstaller.rubyforge.org/svn/trunk/spec/fixture/gems/specifications/stubgem-multiplatform-1.0.1-mswin32.gemspec

From luislavena at gmail.com  Tue Jan  1 21:01:33 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 2 Jan 2008 00:01:33 -0200
Subject: [Rubygems-developers] Should binary platform gems be
	auto-selected before ruby platform on windows?
In-Reply-To: <e3af97510801011717r1fac11d6nf6850f2b47c0ed63@mail.gmail.com>
References: <e3af97510801011717r1fac11d6nf6850f2b47c0ed63@mail.gmail.com>
Message-ID: <71166b3b0801011801h603cbf8fp490561e8859537f1@mail.gmail.com>

On Jan 1, 2008 11:17 PM, Chad Woolley <thewoolleyman at gmail.com> wrote:
> I have some test gems.  There is a dependent gem [1] which depends on
> another dependency gem.  The dependency gem has two platforms, one
> default ruby [2], and one with a platform of 'mswin32' [3].
>
> When I install the dependent gem [1], it selects the ruby platform
> dependency [2] instead of the mswin specific platform dependency [3].
>

Great example Chad, I'll check and debug the Rubygem installer process
tomorrow when return to the office.

> Is this the desired behavior?  I sort of thought that a binary
> platform gem should be preferred over a plain-ruby platform, if it
> exists.  Don't some multi-platform gems sometimes require compilation
> of a 'ruby' platform, but also offer a precompiled binary as well?
>

No, it shouldn't be the desired behavior, since the "platform
automatism" should avoid the raise of issues if there is a gem
properly build for the current platform.

'ruby' [2] gem should be used as fallback if no other gem matching the
platform was found.

> Thanks,
> -- Chad

Thanks for your testing,

>
> [1] http://geminstaller.rubyforge.org/svn/trunk/spec/fixture/gems/specifications/dependent-stubgem-depends-on-multiplatform-1.0.0.gemspec
>
> [2] http://geminstaller.rubyforge.org/svn/trunk/spec/fixture/gems/specifications/stubgem-multiplatform-1.0.1.gemspec
>
> [3] http://geminstaller.rubyforge.org/svn/trunk/spec/fixture/gems/specifications/stubgem-multiplatform-1.0.1-mswin32.gemspec


-- 
Luis Lavena
Multimedia systems
-
A common mistake that people make when trying to design
something completely foolproof is to underestimate
the ingenuity of complete fools.
Douglas Adams

From thewoolleyman at gmail.com  Tue Jan  1 21:48:18 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Tue, 1 Jan 2008 19:48:18 -0700
Subject: [Rubygems-developers] Should binary platform gems be
	auto-selected before ruby platform on windows?
In-Reply-To: <71166b3b0801011801h603cbf8fp490561e8859537f1@mail.gmail.com>
References: <e3af97510801011717r1fac11d6nf6850f2b47c0ed63@mail.gmail.com>
	<71166b3b0801011801h603cbf8fp490561e8859537f1@mail.gmail.com>
Message-ID: <e3af97510801011848h784afb72peb351607a98820a1@mail.gmail.com>

On Jan 1, 2008 7:01 PM, Luis Lavena <luislavena at gmail.com> wrote:
> Thanks for your testing,

Thank my test suite for GemInstaller :)  I've spent quite a while
making it all pass against the latest rubygems, and this was the last
faiiing test.

The failing test is "should handle a multiplatform dependency chain"
in this spec: http://geminstaller.rubyforge.org/svn/trunk/spec/functional/geminstaller_spec.rb

I've made it pass by disabling the binary platform check on windows +
RubyGems >= 0.9.4, so it will be easy to see if you fix it by removing
that condition.

Thanks,
Chad

From thewoolleyman at gmail.com  Thu Jan  3 16:07:31 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Thu, 3 Jan 2008 14:07:31 -0700
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
Message-ID: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>

Is there any way to do a gem update --system, but only to a specified
version?  I'd like to automate my gem system updates, but only up to a
version that I have tested, like = 1.0.1.

Thx,
-- Chad

From luislavena at gmail.com  Thu Jan  3 16:20:25 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Thu, 3 Jan 2008 19:20:25 -0200
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
Message-ID: <71166b3b0801031320r7308ea1an9829796b1bc65a19@mail.gmail.com>

On Jan 3, 2008 7:07 PM, Chad Woolley <thewoolleyman at gmail.com> wrote:
> Is there any way to do a gem update --system, but only to a specified
> version?  I'd like to automate my gem system updates, but only up to a
> version that I have tested, like = 1.0.1.
>

AFAIK, no.

But you can specify the source of update:

gem update --system --source http://my-gem-server

HTH,
-- 
Luis Lavena
Multimedia systems
-
A common mistake that people make when trying to design
something completely foolproof is to underestimate
the ingenuity of complete fools.
Douglas Adams

From thewoolleyman at gmail.com  Thu Jan  3 16:33:29 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Thu, 3 Jan 2008 14:33:29 -0700
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <71166b3b0801031320r7308ea1an9829796b1bc65a19@mail.gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
	<71166b3b0801031320r7308ea1an9829796b1bc65a19@mail.gmail.com>
Message-ID: <e3af97510801031333o69bfb872h6642364e3445dfa0@mail.gmail.com>

On Jan 3, 2008 2:20 PM, Luis Lavena <luislavena at gmail.com> wrote:
> But you can specify the source of update:
>
> gem update --system --source http://my-gem-server

hmm, I wonder how this works (he says lazily without looking at the
source).  Does it install the same version of rubygems that is running
the server?

From jim.weirich at gmail.com  Thu Jan  3 17:31:56 2008
From: jim.weirich at gmail.com (Jim Weirich)
Date: Thu, 3 Jan 2008 17:31:56 -0500
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
Message-ID: <7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>


On Jan 3, 2008, at 4:07 PM, Chad Woolley wrote:

> Is there any way to do a gem update --system, but only to a specified
> version?  I'd like to automate my gem system updates, but only up to a
> version that I have tested, like = 1.0.1.

You should be able to do a

    gem install rubygems-update -v1.0.1

and run the update_rubygems script manually (warning: from memory and  
untested)

It's two steps, but it is under your control then.

-- 
-- Jim Weirich
-- jim.weirich at gmail.com





From thewoolleyman at gmail.com  Fri Jan  4 01:52:31 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Thu, 3 Jan 2008 23:52:31 -0700
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
	<7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
Message-ID: <e3af97510801032252j49f59d5dhc8bd761d2b0b5403@mail.gmail.com>

On Jan 3, 2008 3:31 PM, Jim Weirich <jim.weirich at gmail.com> wrote:
> You should be able to do a
>
>     gem install rubygems-update -v1.0.1
>
> and run the update_rubygems script manually (warning: from memory and
> untested)
>
> It's two steps, but it is under your control then.

Cool, thanks.  I was hoping for something that could be invoked
completely programatically, though.

-- Chad

From jim.weirich at gmail.com  Fri Jan  4 08:54:00 2008
From: jim.weirich at gmail.com (Jim Weirich)
Date: Fri, 4 Jan 2008 08:54:00 -0500
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <e3af97510801032252j49f59d5dhc8bd761d2b0b5403@mail.gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
	<7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
	<e3af97510801032252j49f59d5dhc8bd761d2b0b5403@mail.gmail.com>
Message-ID: <2C918594-A83B-4C39-A9EF-283BFF46328B@gmail.com>


On Jan 4, 2008, at 1:52 AM, Chad Woolley wrote:
>>
>> It's two steps, but it is under your control then.
>
> Cool, thanks.  I was hoping for something that could be invoked
> completely programatically, though.

Ah, sorry.  I was unclear.  When I said manually, I meant separately  
from the gem install command.  Both steps can be done programmatically.

-- 
-- Jim Weirich
-- jim.weirich at gmail.com





From ruby at j-davis.com  Fri Jan  4 18:19:07 2008
From: ruby at j-davis.com (Jeff Davis)
Date: Fri, 04 Jan 2008 15:19:07 -0800
Subject: [Rubygems-developers] Gem build process for C extension
Message-ID: <1199488747.10057.250.camel@dogma.ljc.laika.com>


I am writing a C extension and I was using mkmf, but I had some problems
with that (described below), and so looked into mkrf which appears to
work better.

However:

(a) How do I tell rubygems to invoke "rake" rather than "make" during
the build process? I assume I just need to put something in
the .gemspec, but I don't know what.

(b) How do I tell rubygems that my gem now has a dependency on the
"mkrf" gem?

The problem that I had with mkmf was that it would not properly quote
pathnames, so my extension couldn't be built when people had spaces in
the include paths or lib paths. It was basically impossible to work
around that in mkmf, but I was able to work around it more easily in
mkrf.

Are there documents or examples showing how to properly use gems in
conjunction with a C extension and its build process? How do other
people make sure path names are quoted properly? Should I use mkmf or
mkrf?

Is there comprehensive documentation for a .gemspec file?

Regards,
	Jeff Davis


From ruby at j-davis.com  Fri Jan  4 19:53:23 2008
From: ruby at j-davis.com (Jeff Davis)
Date: Fri, 04 Jan 2008 16:53:23 -0800
Subject: [Rubygems-developers] gem says it builds RDoc, but it's not there
Message-ID: <1199494403.10057.261.camel@dogma.ljc.laika.com>


I can easily build RDoc for my C extension. When I install the gem, it
says it's installing ri and RDoc documentation, but it simply isn't
installed. What can cause that? Do I need to pre-build it and bundle it
with the gem?

One possibly complicating factor is that my gem consists of two modules
by different names that provide some classes by the same names. One is a
backwards-compatibility legacy version, and the two modules aren't
designed to be loaded at the same time. Would that cause this failure?

Regards,
	Jeff Davis


From kevin.clark at gmail.com  Fri Jan  4 21:54:15 2008
From: kevin.clark at gmail.com (Kevin Clark)
Date: Fri, 4 Jan 2008 18:54:15 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <1199488747.10057.250.camel@dogma.ljc.laika.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
Message-ID: <f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>

> (a) How do I tell rubygems to invoke "rake" rather than "make" during
> the build process? I assume I just need to put something in
> the .gemspec, but I don't know what.

If you have a Rakefile or a mkrf_config in the ext directory, it
should use those instead of an extconf.

> (b) How do I tell rubygems that my gem now has a dependency on the
> "mkrf" gem?

Just add mkrf as a dependency like you would with other gems.

> Are there documents or examples showing how to properly use gems in
> conjunction with a C extension and its build process? How do other
> people make sure path names are quoted properly? Should I use mkmf or
> mkrf?

There's an example in the Pickaxe, and one here
http://www.tarkblog.org/rubyextension

I can't speak to quoting pathnames. I'd say mkrf is probably a better
starting point if _any_ hacking is to be done. I've quarantined the
part of my brain that understood mkmf for fear the disease would
spread. It isn't as widely used though, so there may be things
missing. Please _do_ file bug reports if that's the case.

> Is there comprehensive documentation for a .gemspec file?

Probably, but honestly at this point I end up going to the source or
to my other gems.


-- 
Kevin Clark
http://glu.ttono.us

From ruby at j-davis.com  Fri Jan  4 22:43:38 2008
From: ruby at j-davis.com (Jeff Davis)
Date: Fri, 04 Jan 2008 19:43:38 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
Message-ID: <1199504618.10057.282.camel@dogma.ljc.laika.com>

On Fri, 2008-01-04 at 18:54 -0800, Kevin Clark wrote:
> > (a) How do I tell rubygems to invoke "rake" rather than "make" during
> > the build process? I assume I just need to put something in
> > the .gemspec, but I don't know what.
> 
> If you have a Rakefile or a mkrf_config in the ext directory, it
> should use those instead of an extconf.

Doesn't extconf.rb (using mkrf) generate the Rakefile? I don't bundle a
Makefile or a Rakefile. After extconf.rb is run, the Rakefile is
perfect, but I don't think "gem install" is running "rake", because it
fails.

> > (b) How do I tell rubygems that my gem now has a dependency on the
> > "mkrf" gem?
> 
> Just add mkrf as a dependency like you would with other gems.

Is there some documentation about the .gemspec? This is new territory
for me.

> > Are there documents or examples showing how to properly use gems in
> > conjunction with a C extension and its build process? How do other
> > people make sure path names are quoted properly? Should I use mkmf or
> > mkrf?
> 
> There's an example in the Pickaxe, and one here
> http://www.tarkblog.org/rubyextension
> 
> I can't speak to quoting pathnames. I'd say mkrf is probably a better
> starting point if _any_ hacking is to be done. I've quarantined the
> part of my brain that understood mkmf for fear the disease would
> spread. It isn't as widely used though, so there may be things
> missing. Please _do_ file bug reports if that's the case.
> 

I agree completely that mkrf is a better starting point. All I had to do
was quote the include paths (mkmf made this more difficult by supplying
quotes for some paths and not others, and changing between versions...).

I would suggest that user-supplied data be quoted appropriately from
mkrf (rather than the calling code), such that an include or lib dir
with spaces (or some other strangeness) be automatically made safe for
inclusion in the shell commands.

I would also suggest a mechanism to supply CFLAGS (perhaps I missed
it?). This is less important to me, however.

Making it easier to write extensions for Ruby helps everyone, and mkrf
looks like the right way to do that. Thanks for working on it.

Regards,
	Jeff Davis


From luislavena at gmail.com  Fri Jan  4 22:57:57 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Sat, 5 Jan 2008 01:57:57 -0200
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <1199504618.10057.282.camel@dogma.ljc.laika.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<1199504618.10057.282.camel@dogma.ljc.laika.com>
Message-ID: <71166b3b0801041957x6f4a5293x338922864f78268c@mail.gmail.com>

On Jan 5, 2008 1:43 AM, Jeff Davis <ruby at j-davis.com> wrote:
> On Fri, 2008-01-04 at 18:54 -0800, Kevin Clark wrote:
> > > (a) How do I tell rubygems to invoke "rake" rather than "make" during
> > > the build process? I assume I just need to put something in
> > > the .gemspec, but I don't know what.
> >
> > If you have a Rakefile or a mkrf_config in the ext directory, it
> > should use those instead of an extconf.
>
> Doesn't extconf.rb (using mkrf) generate the Rakefile? I don't bundle a
> Makefile or a Rakefile. After extconf.rb is run, the Rakefile is
> perfect, but I don't think "gem install" is running "rake", because it
> fails.
>

I actually don't know about mkrf, but I think you should use
mkrf_config.rb instead of extconf.rb to name the "builder" script.

Anyway, I don't haver personal experience with that or the capacity of
rubygems to handle mkrf.

> > > (b) How do I tell rubygems that my gem now has a dependency on the
> > > "mkrf" gem?
> >
> > Just add mkrf as a dependency like you would with other gems.
>
> Is there some documentation about the .gemspec? This is new territory
> for me.
>

http://docs.rubygems.org/read/chapter/20#page85

> > > Are there documents or examples showing how to properly use gems in
> > > conjunction with a C extension and its build process? How do other
> > > people make sure path names are quoted properly? Should I use mkmf or
> > > mkrf?
> >
> > There's an example in the Pickaxe, and one here
> > http://www.tarkblog.org/rubyextension
> >
> > I can't speak to quoting pathnames. I'd say mkrf is probably a better
> > starting point if _any_ hacking is to be done. I've quarantined the
> > part of my brain that understood mkmf for fear the disease would
> > spread. It isn't as widely used though, so there may be things
> > missing. Please _do_ file bug reports if that's the case.
> >
>
> I agree completely that mkrf is a better starting point. All I had to do
> was quote the include paths (mkmf made this more difficult by supplying
> quotes for some paths and not others, and changing between versions...).
>
> I would suggest that user-supplied data be quoted appropriately from
> mkrf (rather than the calling code), such that an include or lib dir
> with spaces (or some other strangeness) be automatically made safe for
> inclusion in the shell commands.

On what platform are you? since is not impossible but uncommon find
paths with spaces under *nix.

> Making it easier to write extensions for Ruby helps everyone, and mkrf
> looks like the right way to do that. Thanks for working on it.
>

Last time I tried to look at mkrf, couldn't get an updated repository
por point of contact for the project (code was dated 2006 and never
updated).

-- 
Luis Lavena
Multimedia systems
-
A common mistake that people make when trying to design
something completely foolproof is to underestimate
the ingenuity of complete fools.
Douglas Adams

From kevin.clark at gmail.com  Fri Jan  4 23:43:51 2008
From: kevin.clark at gmail.com (Kevin Clark)
Date: Fri, 4 Jan 2008 20:43:51 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <71166b3b0801041957x6f4a5293x338922864f78268c@mail.gmail.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<1199504618.10057.282.camel@dogma.ljc.laika.com>
	<71166b3b0801041957x6f4a5293x338922864f78268c@mail.gmail.com>
Message-ID: <f53e6950801042043o3c07f5ech6fa07faea8ab589b@mail.gmail.com>

> Last time I tried to look at mkrf, couldn't get an updated repository
> por point of contact for the project (code was dated 2006 and never
> updated).

Point of contact at this point is the mailing list:

http://rubyforge.org/mailman/listinfo/mkrf-users

And I think Alain pushed a point release late last year.

-- 
Kevin Clark
http://glu.ttono.us

From jeremy at hinegardner.org  Fri Jan  4 23:07:57 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Fri, 4 Jan 2008 21:07:57 -0700
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <1199504618.10057.282.camel@dogma.ljc.laika.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<1199504618.10057.282.camel@dogma.ljc.laika.com>
Message-ID: <20080105040757.GO4742@hinegardner.org>

On Fri, Jan 04, 2008 at 07:43:38PM -0800, Jeff Davis wrote:
> On Fri, 2008-01-04 at 18:54 -0800, Kevin Clark wrote:
> > > (a) How do I tell rubygems to invoke "rake" rather than "make" during
> > > the build process? I assume I just need to put something in
> > > the .gemspec, but I don't know what.
> > 
> > If you have a Rakefile or a mkrf_config in the ext directory, it
> > should use those instead of an extconf.
> 
> Doesn't extconf.rb (using mkrf) generate the Rakefile? I don't bundle a
> Makefile or a Rakefile. After extconf.rb is run, the Rakefile is
> perfect, but I don't think "gem install" is running "rake", because it
> fails.

Yup that is what is happening, it is running 'make' and there is no
Makefile present.

When rubygems looks to install your extension it has to decide how to do
it.  An extconf.rb file says to use the mkmf method which is:

    ruby extconf.rb
    make

and a mkrf_conf file says to use the mkrf method which is:

    ruby mkrf_conf.rb
    rake

enjoy,

-jeremy


-- 
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From jeremy at hinegardner.org  Fri Jan  4 22:31:52 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Fri, 4 Jan 2008 20:31:52 -0700
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
Message-ID: <20080105033152.GM4742@hinegardner.org>

On Fri, Jan 04, 2008 at 06:54:15PM -0800, Kevin Clark wrote:
> > (a) How do I tell rubygems to invoke "rake" rather than "make" during
> > the build process? I assume I just need to put something in
> > the .gemspec, but I don't know what.
> 
> If you have a Rakefile or a mkrf_config in the ext directory, it
> should use those instead of an extconf.

This only works with rubygems 0.9.5 or higher, before that rubygems did not
support mkrf extensions.

> > (b) How do I tell rubygems that my gem now has a dependency on the
> > "mkrf" gem?
> 
> Just add mkrf as a dependency like you would with other gems.
> 
> > Are there documents or examples showing how to properly use gems in
> > conjunction with a C extension and its build process? How do other
> > people make sure path names are quoted properly? Should I use mkmf or
> > mkrf?
> 
> There's an example in the Pickaxe, and one here
> http://www.tarkblog.org/rubyextension

You can also look at the ruby-libtommath gem.  Its a pretty simple
example of using gems and mkrf to package an extension. 

> I can't speak to quoting pathnames. I'd say mkrf is probably a better
> starting point if _any_ hacking is to be done. I've quarantined the
> part of my brain that understood mkmf for fear the disease would
> spread. It isn't as widely used though, so there may be things
> missing. Please _do_ file bug reports if that's the case.
> 
> > Is there comprehensive documentation for a .gemspec file?
> 
> Probably, but honestly at this point I end up going to the source or
> to my other gems.

The rubygems GemSpec references is online
http://rubygems.org/read/chapter/20#page85 It is not completely
comprehensive, but I've found it invaluable.

enjoy,

-jeremy

-- 
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From drbrain at segment7.net  Sat Jan  5 04:13:43 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:13:43 -0800
Subject: [Rubygems-developers] [ANN] RubyGems 1.0.0
In-Reply-To: <4b6f054f0712200553n1c3998fale59f73b2a6d57728@mail.gmail.com>
References: <8ADD0CF0-C6A0-4BA6-8D5D-32769B79B0FA@segment7.net>
	<4b6f054f0712200553n1c3998fale59f73b2a6d57728@mail.gmail.com>
Message-ID: <8FB2848D-14A7-4B69-A08E-54B9B9CE5E36@segment7.net>

On Dec 20, 2007, at 05:53 AM, Trans wrote:
> On Dec 20, 2007 3:33 AM, Eric Hodel <drbrain at segment7.net> wrote:
>> * Improved detection of RUBYOPT loading rubygems
>
> Could you explain this one more. What changed?

Some linux distro adds an extra file that loads rubygems, and add it  
to RUBYOPT.  To install correctly, RubyGems needs to load the  
installation-to-be-installed, not the currently-installed version.   
Clearing out RUBYOPT fixes this.

From drbrain at segment7.net  Sat Jan  5 04:15:06 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:15:06 -0800
Subject: [Rubygems-developers] [ANN] RubyGems 1.0.0
In-Reply-To: <deb2337a0712200738u5fb196cbxbc8f47fdd0c59e11@mail.gmail.com>
References: <8ADD0CF0-C6A0-4BA6-8D5D-32769B79B0FA@segment7.net>
	<deb2337a0712200738u5fb196cbxbc8f47fdd0c59e11@mail.gmail.com>
Message-ID: <61A9EEB5-90B3-40F6-A21A-EE1B74BC5815@segment7.net>

On Dec 20, 2007, at 07:38 AM, Rick DeNatale wrote:
> On 12/20/07, Eric Hodel <drbrain at segment7.net> wrote:
>> If you have a recent version of RubyGems (0.8.5 or later), then all
>> you need to do is:
>>
>>   $ gem update --system   (you might need to be admin/root)
>>
>> (Note: You may have to run the command twice if you have any  
>> previosly
>> installed rubygems-update gems).
>
> How would one know if they had to run the command twice, or should I
> just do it twice for grins?

I think very old versions of RubyGems have this problem.

From drbrain at segment7.net  Sat Jan  5 04:16:49 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:16:49 -0800
Subject: [Rubygems-developers] LOAD_PATH Ordering with Ruby 1.9 Gems
In-Reply-To: <55D0FA52-3CBF-4406-ADE2-90F11CEED97B@gmail.com>
References: <55D0FA52-3CBF-4406-ADE2-90F11CEED97B@gmail.com>
Message-ID: <8885A87A-D88A-4DC0-9410-D0BEFFF3DAAF@segment7.net>

On Dec 21, 2007, at 06:39 AM, Jim Weirich wrote:

> When I try to override the installed rake in Ruby 1.9 with a gem, I
> get the following error:
>
> /Users/jim/local/ruby19/bin/rake:19:in `load': Is a directory - /
> Users/jim/local/ruby19/lib/ruby/gems/1.9/gems/rake-0.8.0/lib/rake
> (Errno::EISDIR)
> 	from /Users/jim/local/ruby19/bin/rake:19:in `<main>'
>
> It seems that the 'lib' directory in the rake gem preceeds the 'bin'
> directory, so loading 'rake' will find 'lib/rake' and complain about
> that.  Pre-1.9 versions of gem put the bin directory first, so that
> load 'rake' would find that.
>
> Was this a deliberate change?

gem_prelude.rb was adding paths in the wrong order.  I fixed it  
independently of your discovery before 1.9.0.

> (Actually, I never liked the whole idea of including the bin
> directory in the load path in the first place.  Seems to me that we
> could have made the executable stub a wee bit smarter ... but I
> digress).

It's not optimal, but it works, so I'm hesitant to change it just yet.

From drbrain at segment7.net  Sat Jan  5 04:20:34 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:20:34 -0800
Subject: [Rubygems-developers] How best to change defaults for RubyGems
	in JRuby
In-Reply-To: <200712220022.35701.avatar@spellboundnet.com>
References: <4754EA6F.8090209@sun.com>
	<9C21C37A-98C5-466B-A83F-1A32D8AC8920@segment7.net>
	<476B2728.1040508@sun.com>
	<200712220022.35701.avatar@spellboundnet.com>
Message-ID: <E5F1987B-1DF0-4675-8270-CD9B989D7E51@segment7.net>

On Dec 21, 2007, at 21:22 PM, Donavan Pantke wrote:
> On Thursday 20 December 2007 09:38:32 pm Charles Oliver Nutter wrote:
>> Eric Hodel wrote:
>>> I added a patch that moved most of the defaults methods into  
>>> rubygems/
>>> defaults.rb.  Perhaps RubyGems could conditionally define things
>>> depending upon platform in that file?
>>
>> Well let me describe the changes I'm making. I'm updating JRuby to
>> RubyGems 1.0.0 right now.
>>
>> First change in install_update_options.rb:
>>
>>   # Default options for the gem install command.
>>   def install_update_defaults_str
>>     '--rdoc --no-force --no-test --wrappers --env-shebang'
>>   end
>>
>> I don't know if this is anything more than informational
>>
>> Second change is in installer.rb:
>>
>>     if true || @env_shebang then
>>       "#!/usr/bin/env jruby"
>>     else
>>
>> The "if true" needs some explaining. I can't find where to make
>> env_shebang true by default. The other place I tried, also in  
>> installer,
>> didn't seem to make a difference:
>>
>>    options = {
>>
>>       :force => false,
>>       :install_dir => Gem.dir,
>>       :exec_format => false,
>>       :env_shebang => true # this doesn't seem to work
>>
>>     }.merge options
>
> This is because :env_shebang  is being set by  
> install_update_options.rb. Since
> the option already exists, the merge is overriding your default. I  
> submitted
> patch 16508 which should make the above work properly.
>
>>
>> ...so the if true || @env_shebang is just to hardcode it to always  
>> use
>> env shebangs (since JRuby startup script is always just a shell  
>> script).
>> And of course the shebang string itself is changed to say "jruby"
>> instead of "ruby" since it's normally hardcoded to be "ruby". It  
>> would
>> probably be nice if that used some runtime-specific value.
>
> Honestly, it would be real nice if Ruby interpreters would define a  
> standard
> constant that contains the name of the interpreter executable. Would  
> be easy
> to define, and instead of playing patching games, gems could just  
> use that
> constant.

RbConfig::CONFIG['ruby_install_name']

Could somebody file a bug for me?  I won't get back to RubyGems fixups  
for a while yet.

>> So basically what would be nice is if I didn't have to remember to  
>> hunt
>> down and patch these locations each time we updated RubyGems...or  
>> simply
>> to have a single place I can tweak them to JRuby. That's probably  
>> going
>> to be useful in the future with other implementations getting close  
>> to
>> running RubyGems as well.

A /etc/gemrc would be nice.  Same file format as ~/.gemrc would work.   
I don't have time to write it right now, though.

From drbrain at segment7.net  Sat Jan  5 04:23:16 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:23:16 -0800
Subject: [Rubygems-developers] Questions on 1.0.0 on JRuby
In-Reply-To: <4772D634.7010400@sun.com>
References: <476AFBB5.1090604@sun.com> <4772D634.7010400@sun.com>
Message-ID: <9C67E1B6-F9F0-4739-AA32-64D0598ACDFB@segment7.net>

On Dec 26, 2007, at 14:31 PM, Charles Oliver Nutter wrote:
> Charles Oliver Nutter wrote:
>> 1. jgem?
>>
>> I saw this during the 1.0.0 install:
>>
>> install -c -m 0755 /tmp/gem /Users/headius/NetBeansProjects/jruby/ 
>> bin/jgem
>>
>> Why jgem? We've been shipping with just 'gem'.
>
> I'd still like to know why this is the case. We don't want to ship
> RubyGems as "jgem" in JRuby. If anything, we might call it "jem",  
> but I
> strongly prefer just leaving it as "gem".

I think it should be an option in setup.rb, but I may have forgotten  
to add it.  It has caused some confusion, unfortunately.

>> 2. Some version issue on JRuby
>
> Eric suggested, and was correct, that this was an issue with JRuby not
> calling yaml_initialize when reconstituting objects from YAML. I've
> fixed that on JRuby trunk and removed the to_ints hack. Thanks!

Cool.

From drbrain at segment7.net  Sat Jan  5 04:24:54 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:24:54 -0800
Subject: [Rubygems-developers] Problem with multiplatforms gems when
	using SourceIndex#search
In-Reply-To: <e3af97510712271435x35cc3e32m88b56f11e3ad93da@mail.gmail.com>
References: <e3af97510712271435x35cc3e32m88b56f11e3ad93da@mail.gmail.com>
Message-ID: <EA389F62-8FF7-4FAB-AD43-CDB939FA956C@segment7.net>

On Dec 27, 2007, at 14:35 PM, Chad Woolley wrote:
> I have a test scenario where I have two test gems installed - they
> have an identical platform and version, but different platforms (one
> 'ruby', one 'mswin32').  This is on a mac.
>
> The problem is that I can't see how to make SourceIndex#search select
> based on platform - it only seems to search based on name and version.
> Given the gems described above, there's no way to specifically ask if
> the ruby or mswin32 gem is installed.  In other words, there's no way
> to pass a Gem::Dependency that will select a specific platform.

I think currently the RubyGems code post-filters.  It was getting late  
in the release when I noticed the problem, so I didn't try to make a  
better API for it.  I'm open to suggestions.

From drbrain at segment7.net  Sat Jan  5 04:30:21 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:30:21 -0800
Subject: [Rubygems-developers] Can anyone shed a little light on
	executable gems for me?
In-Reply-To: <fe0998f40712280033t39797e3bh7e59c6baf01134a4@mail.gmail.com>
References: <fe0998f40712280033t39797e3bh7e59c6baf01134a4@mail.gmail.com>
Message-ID: <43F81F69-793F-443E-8DE9-CF4C04EC685B@segment7.net>

On Dec 28, 2007, at 24:33 AM, michael greenly wrote:

> It seems to me that something has changed quite a bit about how  
> executable gems are handled in the last few versions?
>
> In the past there was a $GEM_HOME/bin directory but it doesn't seem  
> to be there any more?  Instead it seems the executable files for  
> installed gems end up in the same place that the gem command exits?   
> Is this right?  Or am I off base on this?

When installing a gem, if the install dir is the same as the default  
gem dir, the stub goes where `gem` goes.  If the install dir is  
different, it goes in $GEM_HOME/bin.

> The reason I ask is I'm left wondering how best to separate a Ruby  
> 1.8 environment from a Ruby 1.9 environment so that they can each  
> have a for example a 'rails' gem installed and work in both  
> environments at the same time but with separate gems?

RubyGems 1.0.1 has the --format-executable option to `gem install`, so  
if you have ruby18 and ruby19, gems installed via ruby18 will have 18  
appended to their bin stubs, and gems installed via ruby19 will have  
19 appended to their bin stubs.

> There seemed to be a really straight forward solution prior to this  
> change?  If the executable files for gems were installed in  
> $GEM_HOME/bin then simply installing Ruby using ./configure -- 
> program-suffix=1.8 or --program-suffix= 1.9 like all the linux  
> distributions do anyway would keep everything from overlapping.   
> Then it would be a simple(maybe?) matter of hacking in a "gem run  
> <executablegem>" command.  This then would allow the end user to do  
> " gem1.8 run rails" or "gem1.9 run rails" etc...

Add these lines to ~/.gemrc:

install: --format-executable
update: --format-executable

then you will get rails1.8 and rails1.9.

From drbrain at segment7.net  Sat Jan  5 04:33:01 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:33:01 -0800
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
	<7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
Message-ID: <F9D4DD1E-4808-44F0-AA35-832C313ED520@segment7.net>

On Jan 3, 2008, at 14:31 PM, Jim Weirich wrote:
> On Jan 3, 2008, at 4:07 PM, Chad Woolley wrote:
>
>> Is there any way to do a gem update --system, but only to a specified
>> version?  I'd like to automate my gem system updates, but only up  
>> to a
>> version that I have tested, like = 1.0.1.
>
> You should be able to do a
>
>    gem install rubygems-update -v1.0.1
>
> and run the update_rubygems script manually (warning: from memory and
> untested)
>
> It's two steps, but it is under your control then.

If you have multiple rubygems updates installed, you can choose which  
one to install with:

update_rubygems _0.9.4_

From drbrain at segment7.net  Sat Jan  5 04:35:46 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:35:46 -0800
Subject: [Rubygems-developers] gem says it builds RDoc,
	but it's not there
In-Reply-To: <1199494403.10057.261.camel@dogma.ljc.laika.com>
References: <1199494403.10057.261.camel@dogma.ljc.laika.com>
Message-ID: <E8FBF3E1-6528-4DD4-AE88-B0DB8575DA6E@segment7.net>

On Jan 4, 2008, at 16:53 PM, Jeff Davis wrote:
> I can easily build RDoc for my C extension. When I install the gem, it
> says it's installing ri and RDoc documentation, but it simply isn't
> installed. What can cause that? Do I need to pre-build it and bundle  
> it
> with the gem?

Do you have all the paths set correctly in the gemspec for RDoc  
generation?  Sometimes this is non-intuitive.

From drbrain at segment7.net  Sat Jan  5 04:37:25 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 01:37:25 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <20080105033152.GM4742@hinegardner.org>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<20080105033152.GM4742@hinegardner.org>
Message-ID: <23EC18C0-CAF3-45BA-BB6B-721C485DE361@segment7.net>

On Jan 4, 2008, at 19:31 PM, Jeremy Hinegardner wrote:
> On Fri, Jan 04, 2008 at 06:54:15PM -0800, Kevin Clark wrote:
>>> Is there comprehensive documentation for a .gemspec file?
>>
>> Probably, but honestly at this point I end up going to the source or
>> to my other gems.
>
> The rubygems GemSpec references is online
> http://rubygems.org/read/chapter/20#page85 It is not completely
> comprehensive, but I've found it invaluable.

The RDoc in `gem server` is probably better, but only by a little bit.

From transfire at gmail.com  Sat Jan  5 07:53:02 2008
From: transfire at gmail.com (Trans)
Date: Sat, 5 Jan 2008 07:53:02 -0500
Subject: [Rubygems-developers] [ANN] RubyGems 1.0.0
In-Reply-To: <8FB2848D-14A7-4B69-A08E-54B9B9CE5E36@segment7.net>
References: <8ADD0CF0-C6A0-4BA6-8D5D-32769B79B0FA@segment7.net>
	<4b6f054f0712200553n1c3998fale59f73b2a6d57728@mail.gmail.com>
	<8FB2848D-14A7-4B69-A08E-54B9B9CE5E36@segment7.net>
Message-ID: <4b6f054f0801050453k1c86a68am378a3d9f653fd479@mail.gmail.com>

On Jan 5, 2008 4:13 AM, Eric Hodel <drbrain at segment7.net> wrote:
> On Dec 20, 2007, at 05:53 AM, Trans wrote:
> > On Dec 20, 2007 3:33 AM, Eric Hodel <drbrain at segment7.net> wrote:
> >> * Improved detection of RUBYOPT loading rubygems
> >
> > Could you explain this one more. What changed?
>
> Some linux distro adds an extra file that loads rubygems, and add it
> to RUBYOPT.  To install correctly, RubyGems needs to load the
> installation-to-be-installed, not the currently-installed version.
> Clearing out RUBYOPT fixes this.

Good enough.

Thanks,
T.

From thewoolleyman at gmail.com  Sat Jan  5 14:01:51 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Sat, 5 Jan 2008 12:01:51 -0700
Subject: [Rubygems-developers] Problem with multiplatforms gems when
	using SourceIndex#search
In-Reply-To: <EA389F62-8FF7-4FAB-AD43-CDB939FA956C@segment7.net>
References: <e3af97510712271435x35cc3e32m88b56f11e3ad93da@mail.gmail.com>
	<EA389F62-8FF7-4FAB-AD43-CDB939FA956C@segment7.net>
Message-ID: <e3af97510801051101t4c7b21d4saf7ec51324199db7@mail.gmail.com>

On Jan 5, 2008 2:24 AM, Eric Hodel <drbrain at segment7.net> wrote:
> I think currently the RubyGems code post-filters.  It was getting late
> in the release when I noticed the problem, so I didn't try to make a
> better API for it.  I'm open to suggestions.

What do you mean 'post-filters'?

I think that you should provide some way to pass a Platform in to
SourceIndex#search.  This could be a third optional parameter, or it
could even be a wrapper method if you are worried about the backward
compatibility of that method.

It seems like it should work the same way that install does when you
pass a platform, which I believe is handled in DependencyInstaller, on
about line 80 of 1.0.1: "Gem::Platform.match spec.platform"

Is this is difficult for some reason?

Thanks,
-- Chad

From thewoolleyman at gmail.com  Sat Jan  5 16:45:03 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Sat, 5 Jan 2008 14:45:03 -0700
Subject: [Rubygems-developers] Problem with multiplatforms gems when
	using SourceIndex#search
In-Reply-To: <e3af97510801051101t4c7b21d4saf7ec51324199db7@mail.gmail.com>
References: <e3af97510712271435x35cc3e32m88b56f11e3ad93da@mail.gmail.com>
	<EA389F62-8FF7-4FAB-AD43-CDB939FA956C@segment7.net>
	<e3af97510801051101t4c7b21d4saf7ec51324199db7@mail.gmail.com>
Message-ID: <e3af97510801051345o17260d1cs8db72ebeca7df503@mail.gmail.com>

On Jan 5, 2008 12:01 PM, Chad Woolley <thewoolleyman at gmail.com> wrote:
> I think that you should provide some way to pass a Platform in to
> SourceIndex#search.

After thinking about this some more, I think this would be the best
signature for SourceIndex.search going forward (backward-compatibility
aside):

search(gem_pattern, exact_platform_match = false, platform =
Gem::Platform.local)

This would have the behavior of:
1. returning all platforms by default if only gem_pattern is passed,
2. returning the local platform if gem_pattern and exact_platform_match = true
3. returning the specified platform only if gem_pattern,
exact_platform_match = true, and platform is passed
4. returning all platforms if gem_pattern, exact_platform_match =
false, and platform is passed

#4 is a bit strange, it's identical to #1 and the platform would
essentially be ignored, but I don't think that's a big deal.
Otherwise this approach would be sensible and flexible, and preserve
the existing API.

Thoughts?

-- Chad

From thewoolleyman at gmail.com  Sat Jan  5 22:08:42 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Sat, 5 Jan 2008 20:08:42 -0700
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <F9D4DD1E-4808-44F0-AA35-832C313ED520@segment7.net>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
	<7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
	<F9D4DD1E-4808-44F0-AA35-832C313ED520@segment7.net>
Message-ID: <e3af97510801051908n3d969f3ap5059d2936eeffeb@mail.gmail.com>

On Jan 5, 2008 2:33 AM, Eric Hodel <drbrain at segment7.net> wrote:
> If you have multiple rubygems updates installed, you can choose which
> one to install with:
>
> update_rubygems _0.9.4_

OK, this works for my purposes.  One problem, though.  It always
appears to install the latest installed update version, regardless of
the parameter that is passed.  Strangely, it seemed to work when I did
the same thing via irb (explicitly run "gem 'rubygems-update',
'1.0.0', then "load 'update_rubygems'").

I opened a bug for this:
http://rubyforge.org/tracker/index.php?func=detail&aid=16842&group_id=126&atid=575

Thanks,
Chad

From drbrain at segment7.net  Sat Jan  5 23:59:27 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 5 Jan 2008 20:59:27 -0800
Subject: [Rubygems-developers] gem update --system with a version
	requirement?
In-Reply-To: <e3af97510801051908n3d969f3ap5059d2936eeffeb@mail.gmail.com>
References: <e3af97510801031307j40878ef3r7b8497fa359bf995@mail.gmail.com>
	<7C335BA4-7E27-476E-A9AE-DD56435DFBCD@gmail.com>
	<F9D4DD1E-4808-44F0-AA35-832C313ED520@segment7.net>
	<e3af97510801051908n3d969f3ap5059d2936eeffeb@mail.gmail.com>
Message-ID: <6ED4B1F4-2492-423E-AB8C-7CCF84D10F40@segment7.net>

On Jan 5, 2008, at 19:08 PM, Chad Woolley wrote:
> On Jan 5, 2008 2:33 AM, Eric Hodel <drbrain at segment7.net> wrote:
>> If you have multiple rubygems updates installed, you can choose which
>> one to install with:
>>
>> update_rubygems _0.9.4_
>
> OK, this works for my purposes.  One problem, though.  It always
> appears to install the latest installed update version, regardless of
> the parameter that is passed.  Strangely, it seemed to work when I did
> the same thing via irb (explicitly run "gem 'rubygems-update',
> '1.0.0', then "load 'update_rubygems'").
>
> I opened a bug for this:
> http://rubyforge.org/tracker/index.php?func=detail&aid=16842&group_id=126&atid=575

Thanks.

From charles.nutter at sun.com  Sun Jan  6 22:46:30 2008
From: charles.nutter at sun.com (Charles Oliver Nutter)
Date: Sun, 06 Jan 2008 21:46:30 -0600
Subject: [Rubygems-developers] How best to change defaults for RubyGems
 in JRuby
In-Reply-To: <E5F1987B-1DF0-4675-8270-CD9B989D7E51@segment7.net>
References: <4754EA6F.8090209@sun.com>
	<9C21C37A-98C5-466B-A83F-1A32D8AC8920@segment7.net>
	<476B2728.1040508@sun.com>
	<200712220022.35701.avatar@spellboundnet.com>
	<E5F1987B-1DF0-4675-8270-CD9B989D7E51@segment7.net>
Message-ID: <4781A096.1030208@sun.com>

Eric Hodel wrote:
> RbConfig::CONFIG['ruby_install_name']
> 
> Could somebody file a bug for me?  I won't get back to RubyGems fixups  
> for a while yet.

Done.

http://rubyforge.org/tracker/index.php?func=detail&aid=16878&group_id=126&atid=575

- Charlie

From charles.nutter at sun.com  Sun Jan  6 22:47:58 2008
From: charles.nutter at sun.com (Charles Oliver Nutter)
Date: Sun, 06 Jan 2008 21:47:58 -0600
Subject: [Rubygems-developers] Questions on 1.0.0 on JRuby
In-Reply-To: <9C67E1B6-F9F0-4739-AA32-64D0598ACDFB@segment7.net>
References: <476AFBB5.1090604@sun.com> <4772D634.7010400@sun.com>
	<9C67E1B6-F9F0-4739-AA32-64D0598ACDFB@segment7.net>
Message-ID: <4781A0EE.9020601@sun.com>

Eric Hodel wrote:
> On Dec 26, 2007, at 14:31 PM, Charles Oliver Nutter wrote:
>> Charles Oliver Nutter wrote:
>>> 1. jgem?
>>>
>>> I saw this during the 1.0.0 install:
>>>
>>> install -c -m 0755 /tmp/gem /Users/headius/NetBeansProjects/jruby/ 
>>> bin/jgem
>>>
>>> Why jgem? We've been shipping with just 'gem'.
>> I'd still like to know why this is the case. We don't want to ship
>> RubyGems as "jgem" in JRuby. If anything, we might call it "jem",  
>> but I
>> strongly prefer just leaving it as "gem".
> 
> I think it should be an option in setup.rb, but I may have forgotten  
> to add it.  It has caused some confusion, unfortunately.

Yes, it should just default to 'gem'....most people I know either do 
their own aliasing or prefer to have it be named 'gem'.

I'll file a bug.

- Charlie

From charles.nutter at sun.com  Sun Jan  6 22:49:53 2008
From: charles.nutter at sun.com (Charles Oliver Nutter)
Date: Sun, 06 Jan 2008 21:49:53 -0600
Subject: [Rubygems-developers] Questions on 1.0.0 on JRuby
In-Reply-To: <9C67E1B6-F9F0-4739-AA32-64D0598ACDFB@segment7.net>
References: <476AFBB5.1090604@sun.com> <4772D634.7010400@sun.com>
	<9C67E1B6-F9F0-4739-AA32-64D0598ACDFB@segment7.net>
Message-ID: <4781A161.3030408@sun.com>

Eric Hodel wrote:
> On Dec 26, 2007, at 14:31 PM, Charles Oliver Nutter wrote:
>> Charles Oliver Nutter wrote:
>>> 1. jgem?
>>>
>>> I saw this during the 1.0.0 install:
>>>
>>> install -c -m 0755 /tmp/gem /Users/headius/NetBeansProjects/jruby/ 
>>> bin/jgem
>>>
>>> Why jgem? We've been shipping with just 'gem'.
>> I'd still like to know why this is the case. We don't want to ship
>> RubyGems as "jgem" in JRuby. If anything, we might call it "jem",  
>> but I
>> strongly prefer just leaving it as "gem".
> 
> I think it should be an option in setup.rb, but I may have forgotten  
> to add it.  It has caused some confusion, unfortunately.

http://rubyforge.org/tracker/index.php?func=detail&aid=16879&group_id=126&atid=575

- Charlie

From ruby at j-davis.com  Mon Jan  7 13:59:17 2008
From: ruby at j-davis.com (Jeff Davis)
Date: Mon, 07 Jan 2008 10:59:17 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <71166b3b0801041957x6f4a5293x338922864f78268c@mail.gmail.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<1199504618.10057.282.camel@dogma.ljc.laika.com>
	<71166b3b0801041957x6f4a5293x338922864f78268c@mail.gmail.com>
Message-ID: <1199732357.10057.321.camel@dogma.ljc.laika.com>

On Sat, 2008-01-05 at 01:57 -0200, Luis Lavena wrote:
> > > > (b) How do I tell rubygems that my gem now has a dependency on the
> > > > "mkrf" gem?
> > >
> > > Just add mkrf as a dependency like you would with other gems.
> >
> > Is there some documentation about the .gemspec? This is new territory
> > for me.
> >
> 
> http://docs.rubygems.org/read/chapter/20#page85

Thanks. For some reason I was unable to find this before.

> 
> On what platform are you? since is not impossible but uncommon find
> paths with spaces under *nix.
> 

My gem has users on many platforms, including windows. Regardless,
quoting paths appropriately should be _encouraged_ by the tools (by
making it easy) rather than discouraged.

Regards,
	Jeff Davis


From ruby at j-davis.com  Mon Jan  7 14:06:47 2008
From: ruby at j-davis.com (Jeff Davis)
Date: Mon, 07 Jan 2008 11:06:47 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <20080105033152.GM4742@hinegardner.org>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<20080105033152.GM4742@hinegardner.org>
Message-ID: <1199732807.10057.329.camel@dogma.ljc.laika.com>

On Fri, 2008-01-04 at 20:31 -0700, Jeremy Hinegardner wrote:
> On Fri, Jan 04, 2008 at 06:54:15PM -0800, Kevin Clark wrote:
> > > (a) How do I tell rubygems to invoke "rake" rather than "make" during
> > > the build process? I assume I just need to put something in
> > > the .gemspec, but I don't know what.
> > 
> > If you have a Rakefile or a mkrf_config in the ext directory, it
> > should use those instead of an extconf.
> 
> This only works with rubygems 0.9.5 or higher, before that rubygems did not
> support mkrf extensions.
> 

Thank you for the information.

This puts me in the unfortunate position of requiring recent versions of
rubygems to even build my extension, or dealing with the existing
problems in mkmf.

I don't have the option of an unportable build process, so I suppose
I'll figure something out.

I am seriously considering now using an extconf.rb that uses mkrf, and
then a static Makefile that just calls "rake"... I wonder if that would
work?

Regards,
	Jeff Davis


From drbrain at segment7.net  Mon Jan  7 16:32:48 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Mon, 7 Jan 2008 13:32:48 -0800
Subject: [Rubygems-developers] Gem build process for C extension
In-Reply-To: <1199732357.10057.321.camel@dogma.ljc.laika.com>
References: <1199488747.10057.250.camel@dogma.ljc.laika.com>
	<f53e6950801041854n5c68729bqe674ac6fec19c88@mail.gmail.com>
	<1199504618.10057.282.camel@dogma.ljc.laika.com>
	<71166b3b0801041957x6f4a5293x338922864f78268c@mail.gmail.com>
	<1199732357.10057.321.camel@dogma.ljc.laika.com>
Message-ID: <88CE4E66-C362-4012-A097-87583E473953@segment7.net>

On Jan 7, 2008, at 10:59 AM, Jeff Davis wrote:
> On Sat, 2008-01-05 at 01:57 -0200, Luis Lavena wrote:
>>
>> On what platform are you? since is not impossible but uncommon find
>> paths with spaces under *nix.
>
> My gem has users on many platforms, including windows. Regardless,
> quoting paths appropriately should be _encouraged_ by the tools (by
> making it easy) rather than discouraged.

Absent somebody writing patches, it's probably not going to get  
fixed.  It's a very low priority bug.  It's probably very easy to fix,  
if you have the problem and the time.

From transfire at gmail.com  Fri Jan 11 00:39:33 2008
From: transfire at gmail.com (Trans)
Date: Thu, 10 Jan 2008 21:39:33 -0800 (PST)
Subject: [Rubygems-developers] Google Groups Archive
Message-ID: <51563b2f-68fa-4a71-8c5c-fd67d3a95386@q39g2000hsf.googlegroups.com>

Just wanted to let you all know that I started a google groups archive
of this mailing list, as I prefer to use gg's web interface for
mailing lists, rather the clutter my inbox. you may have the same
preference, if so here you go...

  http://groups.google.com/group/rubygems-developers?hl=en

T.

From luislavena at gmail.com  Sat Jan 12 08:44:40 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Sat, 12 Jan 2008 11:44:40 -0200
Subject: [Rubygems-developers] Should binary platform gems be
	auto-selected before ruby platform on windows?
In-Reply-To: <e3af97510801011848h784afb72peb351607a98820a1@mail.gmail.com>
References: <e3af97510801011717r1fac11d6nf6850f2b47c0ed63@mail.gmail.com>
	<71166b3b0801011801h603cbf8fp490561e8859537f1@mail.gmail.com>
	<e3af97510801011848h784afb72peb351607a98820a1@mail.gmail.com>
Message-ID: <71166b3b0801120544u521bd0f0nb7e6b4e5664e9ecc@mail.gmail.com>

On Jan 2, 2008 12:48 AM, Chad Woolley <thewoolleyman at gmail.com> wrote:
> On Jan 1, 2008 7:01 PM, Luis Lavena <luislavena at gmail.com> wrote:
> > Thanks for your testing,
>
> Thank my test suite for GemInstaller :)  I've spent quite a while
> making it all pass against the latest rubygems, and this was the last
> faiiing test.
>
> The failing test is "should handle a multiplatform dependency chain"
> in this spec: http://geminstaller.rubyforge.org/svn/trunk/spec/functional/geminstaller_spec.rb
>
> I've made it pass by disabling the binary platform check on windows +
> RubyGems >= 0.9.4, so it will be easy to see if you fix it by removing
> that condition.
>

Hello Chad,

I couldn't get enough free time to check this, but done some test with
real life gems.

On a clean installation, I decided to install mongrel_service
(i386-mswin32 platform) which listed as dependency mongrel gem, all
using RubyGems 1.0.1:

*** LOCAL GEMS ***

fxri (0.3.6)
fxruby (1.6.12)
hpricot (0.6)
log4r (1.0.5)
rake (0.8.1)
sources (0.0.1)

>gem install mongrel_service --no-ri --no-rdoc
Bulk updating Gem source index for: http://gems.rubyforge.org
Successfully installed gem_plugin-0.2.3
Successfully installed cgi_multipart_eof_fix-2.5.0
Successfully installed mongrel-1.1.3-x86-mswin32
Successfully installed win32-service-0.5.2-x86-mswin32
Successfully installed mongrel_service-0.3.4-x86-mswin32
5 gems installed

It choosed mongrel i386-mswin32 gem over the ruby one.

Can you still reproduce the problem with 1.0.1 rubygems codebase?

-- 
Luis Lavena
Multimedia systems
-
A common mistake that people make when trying to design
something completely foolproof is to underestimate
the ingenuity of complete fools.
Douglas Adams

From thewoolleyman at gmail.com  Sun Jan 13 02:33:59 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Sun, 13 Jan 2008 00:33:59 -0700
Subject: [Rubygems-developers] Should binary platform gems be
	auto-selected before ruby platform on windows?
In-Reply-To: <71166b3b0801120544u521bd0f0nb7e6b4e5664e9ecc@mail.gmail.com>
References: <e3af97510801011717r1fac11d6nf6850f2b47c0ed63@mail.gmail.com>
	<71166b3b0801011801h603cbf8fp490561e8859537f1@mail.gmail.com>
	<e3af97510801011848h784afb72peb351607a98820a1@mail.gmail.com>
	<71166b3b0801120544u521bd0f0nb7e6b4e5664e9ecc@mail.gmail.com>
Message-ID: <e3af97510801122333t2b1e1a74h8d1c94801f17f9fb@mail.gmail.com>

On Jan 12, 2008 6:44 AM, Luis Lavena <luislavena at gmail.com> wrote:
> Can you still reproduce the problem with 1.0.1 rubygems codebase?

Thanks for looking into it.

It works now from the command line, but the test referenced above is
still failing when I assert that the binary platorm is installed on
windows with rubygems 1.0.1.

Not sure if it only happens when I invoke SourceIndex#search
programatically, or if there's a bug in my test.  I'll look into it
and get back to you...

-- Chad

From laurent.sansonetti at gmail.com  Tue Jan 15 10:48:52 2008
From: laurent.sansonetti at gmail.com (Laurent Sansonetti)
Date: Tue, 15 Jan 2008 16:48:52 +0100
Subject: [Rubygems-developers] [PATCH] gem unpack fails in gem isn't in the
	default repository
Message-ID: <1be7247c0801150748l2266ce3fy94857bf68bb89468@mail.gmail.com>

Hi,

Currently with trunk, gem unpack fails if you pass it a gem that isn't
in the default gems repository, but another repository.

It is the case in Mac OS X 10.5, where there are 2 repositories, but
all pre-installed gems being in the non-default one.

$ ruby -r rubygems -e "p Gem.dir, Gem.path"
"/Library/Ruby/Gems/1.8"
["/System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/gems/1.8",
"/Library/Ruby/Gems/1.8"]

$ gem unpack termios
ERROR:  While executing gem ... (Gem::Exception)
    Cannot load gem at
[/Library/Ruby/Gems/1.8/cache/termios-0.9.4.gem] in /Volumes/Data

The following patch fixes the problem.

$ gem unpack termios
Unpacked gem: '/Volumes/Data/termios-0.9.4'

Laurent

Index: lib/rubygems/commands/unpack_command.rb
===================================================================
--- lib/rubygems/commands/unpack_command.rb	(revision 1594)
+++ lib/rubygems/commands/unpack_command.rb	(working copy)
@@ -73,7 +73,7 @@
     # Furthermore, the name match must be exact (ignoring case).
     if gemname =~ /^#{selected.name}$/i
       filename = selected.full_name + '.gem'
-      return File.join(Gem.dir, 'cache', filename)
+      return Gem.path.map { |d| File.join(d, 'cache', filename)
}.find { |p| File.exist?(p) }
     else
       return nil
     end

From hgs at dmu.ac.uk  Tue Jan 15 12:11:46 2008
From: hgs at dmu.ac.uk (Hugh Sasse)
Date: Tue, 15 Jan 2008 17:11:46 +0000 (WET)
Subject: [Rubygems-developers] Docs found by google seem old.
Message-ID: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>

If I google for gemspec I get 
http://rubygems.org/read/chapter/20

which says:
<quote>

1.1 Specification Reference

This page is generated by hieraki rake task in the RubyGems CVS. Any changes to this page will be lost. Contact a member of the RubyGems team if you have suggestions.

Last generated: 2005-06-01 00:11:33 EDT (Wednesday)
</quote>

Whom should I email to get them to refresh this, or are they reading
the list?  Or is something else necessary to kill it's pagerank?

        Hugh


From jeremy at hinegardner.org  Wed Jan 16 15:00:31 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Wed, 16 Jan 2008 13:00:31 -0700
Subject: [Rubygems-developers] problem packaging rubybgems 1.0.1,
	install --prefix changed?
Message-ID: <20080116200031.GA7244@hinegardner.org>

Hi all,

I'm packaging rubygems 1.0.1 up as an rpm for our internal server builds
and I'm encountering a problem.  In 0.9.4 the setup.rb was the standard
setup.rb from Minero Aoki and to install into a fake root directory 
you would do 'ruby setup.rb install --prefix=$INSTALL_DIR' and
everything would be installed relative to $INSTALL_DIR.  That is, things
would be installed into:

    $INSTALL_DIR/usr/bin
    $INSTALL_DIR/usr/lib/ruby/site_lib/1.8/

Now in 1.0.1 a completely different setup.rb is used with a different
'install --prefix' behavoir.  Without --prefix everything will install
just fine, into a normal system, but not into a system for packaging.
Now if I use --prefix=$INSTALL_DIR I get:

    $INSTALL_DIR/bin
    $INSTALL_DIR/lib

This ends up installing rubygems.rb into $INSTALL_DIR/lib/rubygems.rb
which is definitately not where it needs to go.

So, my question is, what is the proper approach to tell setup.rb to
install relative to a root directory?  That is, what would be the best
way to now package up rubygems for distribution via RPM?

Is this a bug?  Or is this a plan for the future?  Would a patch for
setup.rb for say --installdir be appropriate?  I'll be happy to file
this in the tracker and put together a patch if necessary.

I guess in this case, for now I'll be falling back to 0.9.4 for now
until we can resolve this issue.

enjoy,

-jeremy

-- 
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From darix at web.de  Wed Jan 16 15:20:38 2008
From: darix at web.de (Marcus Rueckert)
Date: Wed, 16 Jan 2008 21:20:38 +0100
Subject: [Rubygems-developers] problem packaging rubybgems 1.0.1,
	install --prefix changed?
In-Reply-To: <20080116200031.GA7244@hinegardner.org>
References: <20080116200031.GA7244@hinegardner.org>
Message-ID: <20080116202038.GN16797@pixel.global-banlist.de>

On 2008-01-16 13:00:31 -0700, Jeremy Hinegardner wrote:
> I'm packaging rubygems 1.0.1 up as an rpm for our internal server builds
> and I'm encountering a problem.  In 0.9.4 the setup.rb was the standard
> setup.rb from Minero Aoki and to install into a fake root directory 
> you would do 'ruby setup.rb install --prefix=$INSTALL_DIR' and
> everything would be installed relative to $INSTALL_DIR.  That is, things
> would be installed into:
> 
>     $INSTALL_DIR/usr/bin
>     $INSTALL_DIR/usr/lib/ruby/site_lib/1.8/
> 
> Now in 1.0.1 a completely different setup.rb is used with a different
> 'install --prefix' behavoir.  Without --prefix everything will install
> just fine, into a normal system, but not into a system for packaging.
> Now if I use --prefix=$INSTALL_DIR I get:
> 
>     $INSTALL_DIR/bin
>     $INSTALL_DIR/lib
> 
> This ends up installing rubygems.rb into $INSTALL_DIR/lib/rubygems.rb
> which is definitately not where it needs to go.
> 
> So, my question is, what is the proper approach to tell setup.rb to
> install relative to a root directory?  That is, what would be the best
> way to now package up rubygems for distribution via RPM?
> 
> Is this a bug?  Or is this a plan for the future?  Would a patch for
> setup.rb for say --installdir be appropriate?  I'll be happy to file
> this in the tracker and put together a patch if necessary.
> 
> I guess in this case, for now I'll be falling back to 0.9.4 for now
> until we can resolve this issue.

i just solved that issue actually. :)

see the attached patch. my %install section looks like:
{{{
  %install
  GEM_HOME=%{buildroot}%{_libdir}/ruby/gems/%{rb_ver}/ \
      ruby -rvendor-specific setup.rb --buildroot=%{buildroot}
}}}

i am also working on a patch to allow
"gem install --buildroot=%{buildroot} somegem"

it works already with locally available gems but fails with remote gems.

hope this helps

    darix

-- 
           openSUSE - SUSE Linux is my linux
               openSUSE is good for you
                   www.opensuse.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rubygems-1.0.1_install_into_buildroot.patch
Type: text/x-patch
Size: 1468 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/rubygems-developers/attachments/20080116/95f95948/attachment.bin 

From lcl at tarazed.demon.co.uk  Wed Jan 16 17:44:36 2008
From: lcl at tarazed.demon.co.uk (Len Lawrence)
Date: Wed, 16 Jan 2008 22:44:36 +0000
Subject: [Rubygems-developers] Failure to find newly installed gem - fox16
Message-ID: <478E88D4.6070600@tarazed.demon.co.uk>

System: x86_64, Mandriva 2008 (supplies 0.9.0, a buggy version of gem)

ruby 1.8.6
RubyGems 1.0.1

I joined this list after trying to sort out a problem with the fxruby
gem.  Lyle Thomson put me on the right track, which allowed me to run a
Fox test program (derived from David Berube).  However there is a
problem with path searching which Lyle thought might be caused by some
kind of confusion in gem regarding the target architecture.  He said
"Take it to the RubyGems developers".

gem install fxruby-1.6.13 places the fox16 files in the /usr/lib64
branch, but:

[lcl at menkalinan fox]$ ruby -e 'puts $:'
/usr/lib/ruby/site_ruby/1.8
/usr/lib/ruby/site_ruby/1.8/x86_64-linux-gnu
/usr/lib/ruby/site_ruby
/usr/lib/ruby/1.8
/usr/lib/ruby/1.8/x86_64-linux-gnu
.

No sign of a gem load path there.  I even tried copying the installed
gem to the /usr/lib branch and again raised an error on the first line
of the program:
  require 'fox16'

"cannot find fox16"

If I use:

ruby -I/usr/lib/ruby/gems/1.8/gems/fxruby-1.6.13/ext/fox16 zeg.rb

the program runs.

That is a bit annoying, even with the use of RUBYOPT, but I can live with
it.  I can't help feeling I am missing something here.  The default load
paths bypass the gems branch altogether.

gem environment gemdir => /usr/lib64/ruby/gems/1.8

If gemdir is added to the search paths ruby still fails to find the
gem.  Only the full path to ext/fox16 will do.  I really do not know
what is going on here and hoped that somebody could shed some light on
a newbie's problems.

Len


From avatar at spellboundnet.com  Wed Jan 16 22:07:29 2008
From: avatar at spellboundnet.com (Donavan Pantke)
Date: Wed, 16 Jan 2008 22:07:29 -0500
Subject: [Rubygems-developers] problem packaging rubybgems 1.0.1,
	install --prefix changed?
In-Reply-To: <20080116202038.GN16797@pixel.global-banlist.de>
References: <20080116200031.GA7244@hinegardner.org>
	<20080116202038.GN16797@pixel.global-banlist.de>
Message-ID: <200801162207.29395.avatar@spellboundnet.com>

On Wednesday 16 January 2008 03:20:38 pm Marcus Rueckert wrote:
> On 2008-01-16 13:00:31 -0700, Jeremy Hinegardner wrote:
> > I'm packaging rubygems 1.0.1 up as an rpm for our internal server builds
> > and I'm encountering a problem.  In 0.9.4 the setup.rb was the standard
> > setup.rb from Minero Aoki and to install into a fake root directory
> > you would do 'ruby setup.rb install --prefix=$INSTALL_DIR' and
> > everything would be installed relative to $INSTALL_DIR.  That is, things
> > would be installed into:
> >
> >     $INSTALL_DIR/usr/bin
> >     $INSTALL_DIR/usr/lib/ruby/site_lib/1.8/
> >
> > Now in 1.0.1 a completely different setup.rb is used with a different
> > 'install --prefix' behavoir.  Without --prefix everything will install
> > just fine, into a normal system, but not into a system for packaging.
> > Now if I use --prefix=$INSTALL_DIR I get:
> >
> >     $INSTALL_DIR/bin
> >     $INSTALL_DIR/lib
> >
> > This ends up installing rubygems.rb into $INSTALL_DIR/lib/rubygems.rb
> > which is definitately not where it needs to go.
> >
> > So, my question is, what is the proper approach to tell setup.rb to
> > install relative to a root directory?  That is, what would be the best
> > way to now package up rubygems for distribution via RPM?
> >
> > Is this a bug?  Or is this a plan for the future?  Would a patch for
> > setup.rb for say --installdir be appropriate?  I'll be happy to file
> > this in the tracker and put together a patch if necessary.
> >
> > I guess in this case, for now I'll be falling back to 0.9.4 for now
> > until we can resolve this issue.
>
> i just solved that issue actually. :)
>
> see the attached patch. my %install section looks like:
> {{{
>   %install
>   GEM_HOME=%{buildroot}%{_libdir}/ruby/gems/%{rb_ver}/ \
>       ruby -rvendor-specific setup.rb --buildroot=%{buildroot}
> }}}
>
> i am also working on a patch to allow
> "gem install --buildroot=%{buildroot} somegem"
>
> it works already with locally available gems but fails with remote gems.

OK, I can understand wanting a buildroot-style prefix for installing RubyGems 
itself, but I don't understand the practical need for a buildroot for 
installing gems. The only difference I see with a buildroot install and 
using --install-dir is that the binaries are dropped in the bin directory of 
the install dir instead of the system bindir. Wouldn't the following be 
perfectly legit in in a spec file?

mkdir -p $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/
mkdir -p $RPM_BUILD_ROOT/%{bindir}
gem install -i $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/ some_gem
mv $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/some_gem_bin 
$RPM_BUILD_ROOT/%{bindir}

This is the kind of predicatable behavior that I've been trying to accomplish 
with my previous patches. Let me know what you think.

Thanks!
Donavan
>
> hope this helps
>
>     darix

From drbrain at segment7.net  Thu Jan 17 02:39:59 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Wed, 16 Jan 2008 23:39:59 -0800
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
Message-ID: <4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>

On Jan 15, 2008, at 09:11 AM, Hugh Sasse wrote:
> If I google for gemspec I get
> http://rubygems.org/read/chapter/20
>
> which says:
> <quote>
>
> 1.1 Specification Reference
>
> This page is generated by hieraki rake task in the RubyGems CVS. Any  
> changes to this page will be lost. Contact a member of the RubyGems  
> team if you have suggestions.
>
> Last generated: 2005-06-01 00:11:33 EDT (Wednesday)
> </quote>
>
> Whom should I email to get them to refresh this, or are they reading
> the list?  Or is something else necessary to kill it's pagerank?

IIRC, this is on Jim's server, and hasn't been updated since forever.

From drbrain at segment7.net  Thu Jan 17 02:41:56 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Wed, 16 Jan 2008 23:41:56 -0800
Subject: [Rubygems-developers] Failure to find newly installed gem -
	fox16
In-Reply-To: <478E88D4.6070600@tarazed.demon.co.uk>
References: <478E88D4.6070600@tarazed.demon.co.uk>
Message-ID: <E3DC179D-8007-4BB0-981A-5A5AA82E334F@segment7.net>

On Jan 16, 2008, at 14:44 PM, Len Lawrence wrote:

> System: x86_64, Mandriva 2008 (supplies 0.9.0, a buggy version of gem)
>
> [...]
>
> If gemdir is added to the search paths ruby still fails to find the
> gem.  Only the full path to ext/fox16 will do.  I really do not know
> what is going on here and hoped that somebody could shed some light on
> a newbie's problems.

Does this still happen with RubyGems 1.x?

0.9.0 is not only very old, but allows malicious gems to write to any  
path on your system.

From darix at web.de  Thu Jan 17 04:58:20 2008
From: darix at web.de (Marcus Rueckert)
Date: Thu, 17 Jan 2008 10:58:20 +0100
Subject: [Rubygems-developers] problem packaging rubybgems 1.0.1,
	install --prefix changed?
In-Reply-To: <200801162207.29395.avatar@spellboundnet.com>
References: <20080116200031.GA7244@hinegardner.org>
	<20080116202038.GN16797@pixel.global-banlist.de>
	<200801162207.29395.avatar@spellboundnet.com>
Message-ID: <20080117095820.GO16797@pixel.global-banlist.de>

On 2008-01-16 22:07:29 -0500, Donavan Pantke wrote:
> OK, I can understand wanting a buildroot-style prefix for installing RubyGems 
> itself, but I don't understand the practical need for a buildroot for 
> installing gems. The only difference I see with a buildroot install and 
> using --install-dir is that the binaries are dropped in the bin directory of 
> the install dir instead of the system bindir. Wouldn't the following be 
> perfectly legit in in a spec file?
> 
> mkdir -p $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/
> mkdir -p $RPM_BUILD_ROOT/%{bindir}
> gem install -i $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/ some_gem

if you go that road, you want:
gem install --install-dir=%{buildroot}$(gem environment gemdir) foo.gem

> mv $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/some_gem_bin 
> $RPM_BUILD_ROOT/%{bindir}
> 
> This is the kind of predicatable behavior that I've been trying to accomplish 
> with my previous patches. Let me know what you think.

oh it makes a difference.
if you use --install-dir it copies the bin files to a wrong place. uses
a different lib dir (and totally ignores stuff like biarch systems)

--build-root just prefixes everything with the temporary root dir and we
are done. no manually copying stuff around. no constructing paths
manually. and all that.

now that i upgraded my package to 1.0.1, i will checkout your patches.
from the previous mail i think i will use some kind of custom default.rb
during testing. (similar to my -rvendor-specific stuff for non-gem ruby
libs) so the user gets the normal gem dir and my stuff ends up in the
vendor_gem dir. but binaries should e.g. always be installed into the
ruby bin dir. i will report back.

    darix

-- 
           openSUSE - SUSE Linux is my linux
               openSUSE is good for you
                   www.opensuse.org

From lcl at tarazed.demon.co.uk  Thu Jan 17 08:37:50 2008
From: lcl at tarazed.demon.co.uk (Len Lawrence)
Date: Thu, 17 Jan 2008 13:37:50 +0000
Subject: [Rubygems-developers] Failure to find newly installed gem
	-	fox16
In-Reply-To: <E3DC179D-8007-4BB0-981A-5A5AA82E334F@segment7.net>
References: <478E88D4.6070600@tarazed.demon.co.uk>
	<E3DC179D-8007-4BB0-981A-5A5AA82E334F@segment7.net>
Message-ID: <478F5A2E.8020700@tarazed.demon.co.uk>

Eric Hodel wrote:
> On Jan 16, 2008, at 14:44 PM, Len Lawrence wrote:
>   
> Does this still happen with RubyGems 1.x?
>
> 0.9.0 is not only very old, but allows malicious gems to write to any  
> path on your system.
>
>   
Yes, but problem solved.  I feel like an idiot - something WAS missing:

  require 'rubygems'

The test example started with the require 'fox16' statement.  Once 
corrected it worked perfectly.

Retreats in embarassment...

Len


From hgs at dmu.ac.uk  Thu Jan 17 09:51:01 2008
From: hgs at dmu.ac.uk (Hugh Sasse)
Date: Thu, 17 Jan 2008 14:51:01 +0000 (WET)
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
Message-ID: <Pine.GSO.4.64.0801171447220.486@brains.eng.cse.dmu.ac.uk>

On Wed, 16 Jan 2008, Eric Hodel wrote:

> On Jan 15, 2008, at 09:11 AM, Hugh Sasse wrote:
> > If I google for gemspec I get
> > http://rubygems.org/read/chapter/20
> >
> > which says:
> > <quote>
        [...]
> > Last generated: 2005-06-01 00:11:33 EDT (Wednesday)
> > </quote>
> >
> > Whom should I email to get them to refresh this, or are they reading
> > the list?  Or is something else necessary to kill it's pagerank?
> 
> IIRC, this is on Jim's server, and hasn't been updated since forever.

Jim Weirich?   I was waiting for him to come in on this thread, but
maybe he is off-list or has set mail to off.   If he's the right Jim
I'll email him about it.  Thank you.

        Hugh

From jim.weirich at gmail.com  Thu Jan 17 09:59:28 2008
From: jim.weirich at gmail.com (Jim Weirich)
Date: Thu, 17 Jan 2008 09:59:28 -0500
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
Message-ID: <C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>


On Jan 17, 2008, at 2:39 AM, Eric Hodel wrote:

> IIRC, this is on Jim's server, and hasn't been updated since forever.

Actually, Chad's server.  But point taken.

The rake task to generate this page is no longer working (probably  
out of date w.r.t. the current code base).  I would recommend fixing  
the rake task and regenerating the page.  Once that is done, we can  
get the page updated.

BTW, I'm migrating the Rake docs off of Hieraki (just to cumbersome  
to use for real docs) and moving them to a static HTML website  
(generated by webgen currently, altho people have recommended  
nanoc).  Anybody interested in taking on a similar task for RubyGems?

-- 
-- Jim Weirich
-- jim.weirich at gmail.com





From todd.fisher at gmail.com  Thu Jan 17 10:48:00 2008
From: todd.fisher at gmail.com (Todd Fisher)
Date: Thu, 17 Jan 2008 10:48:00 -0500
Subject: [Rubygems-developers] removing a source
Message-ID: <e9a952090801170748m4bf0c31eg2f0e25942fba580a@mail.gmail.com>

I'm having trouble removing a specific source within my local gem
repository.  Here's what I did.
gem sources --add http://gems.myinternalhost.com:8808/

Now later I'm in a place where I can no longer access
gems.myinternalhost.com:8808/ or it has gone down.  I'd like to install a
new gem from gems.rubyforge.org.  I can do this but the only way I found was
by adding the --source gems.rubyforge.org, causing the operation to ignore
all other sources.  But what I would really like to do is remove the bad
source...

I've tried this:

rm $GEM_HOME/source_cache
rm -f $GEM_HOME/cache/*

gem sources --remove http://gems.myinternalhost.com:8808/
gem source --remove http://gems.myinternalhost.com:8808/

Here's what I get:

sudo gem source --remove http://gems.myinternalhost.com:8808/
Bulk updating Gem source index for: http://gems.rubyforge.org
ERROR:  While executing gem ... (Gem::RemoteFetcher::FetchError)
    getaddrinfo: No address associated with nodename (SocketError)
        getting size of http://gems.myinternalhost.com:8808/Marshal.4.8

 gem --version
1.0.1

Any ideas?

Thanks,
Todd
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/rubygems-developers/attachments/20080117/e9651d92/attachment.html 

From jeremy at hinegardner.org  Thu Jan 17 11:03:51 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Thu, 17 Jan 2008 09:03:51 -0700
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
	<C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>
Message-ID: <20080117160351.GB7244@hinegardner.org>

On Thu, Jan 17, 2008 at 09:59:28AM -0500, Jim Weirich wrote:
> BTW, I'm migrating the Rake docs off of Hieraki (just to cumbersome  
> to use for real docs) and moving them to a static HTML website  
> (generated by webgen currently, altho people have recommended  
> nanoc).  Anybody interested in taking on a similar task for RubyGems?

Jim, you should give webby a try too : http://webby.rubyforge.org/

And I could be intereseted in helping out with the RubyGems
documentation.

enjoy,

-jeremy

-- 
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From jim.weirich at gmail.com  Thu Jan 17 11:34:59 2008
From: jim.weirich at gmail.com (Jim Weirich)
Date: Thu, 17 Jan 2008 11:34:59 -0500
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <20080117160351.GB7244@hinegardner.org>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
	<C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>
	<20080117160351.GB7244@hinegardner.org>
Message-ID: <726376BF-DEA7-4199-9703-269F255B24C1@gmail.com>

On Jan 17, 2008, at 11:03 AM, Jeremy Hinegardner wrote:
> Jim, you should give webby a try too : http://webby.rubyforge.org/


Sigh.  Everytime I mention this, someone suggests a DIFFERENT static  
website generator.  Who would have thought there was so much  
competition in this area.

So, why webby over nanoc and webgen?

-- 
-- Jim Weirich
-- jim.weirich at gmail.com





From jeremy at hinegardner.org  Thu Jan 17 12:15:34 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Thu, 17 Jan 2008 10:15:34 -0700
Subject: [Rubygems-developers] [OT] Static site generators
In-Reply-To: <726376BF-DEA7-4199-9703-269F255B24C1@gmail.com>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
	<C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>
	<20080117160351.GB7244@hinegardner.org>
	<726376BF-DEA7-4199-9703-269F255B24C1@gmail.com>
Message-ID: <20080117171534.GD7244@hinegardner.org>

On Thu, Jan 17, 2008 at 11:34:59AM -0500, Jim Weirich wrote:
> On Jan 17, 2008, at 11:03 AM, Jeremy Hinegardner wrote:
>> Jim, you should give webby a try too : http://webby.rubyforge.org/
>
>
> Sigh.  Everytime I mention this, someone suggests a DIFFERENT static 
> website generator.  Who would have thought there was so much competition in 
> this area.

Yeah, it can get pretty wild, I've tried out most of them over time and
these days I've settled on webby.

> So, why webby over nanoc and webgen?

So, I turned to my left and asked the author of webgen why webby over
nanoc and webgen (Disclaimer, I work Tim Pease, author of webby):

His reply, webby is:
    
    * completely based upon Rake tasks 
    * has an autobuild loop so that pages get regenerated when their sources is saved
    * is simpler than webgen 
        - and I agree, webby is much easier to extend for custom 
          purposes, I've written both webgen and webby customisations and 
          it was much easier to do so in webby.
    * has rails style helpers for linking to internal pages
    * utilizes heel and launchy to make previewing of your site easy
    * has growl notifications if you want them
    * has CodeRay and Graphviz helpers
    * generated content can be filtered through tidy so it looks nice

Of course, as always, I would look over the options. I know Tim and Denis
(author of nanoc) have been chatting back and forth and enjoying the
competition :-).  In fact I think nanoc recently had a release that
added some of the features from webby.

enjoy,

-jeremy

-- 
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From avatar at spellboundnet.com  Thu Jan 17 14:35:01 2008
From: avatar at spellboundnet.com (Donavan Pantke)
Date: Thu, 17 Jan 2008 14:35:01 -0500
Subject: [Rubygems-developers] problem packaging rubybgems 1.0.1,
 install --prefix changed?
In-Reply-To: <20080117095820.GO16797@pixel.global-banlist.de>
References: <20080116200031.GA7244@hinegardner.org>	<20080116202038.GN16797@pixel.global-banlist.de>	<200801162207.29395.avatar@spellboundnet.com>
	<20080117095820.GO16797@pixel.global-banlist.de>
Message-ID: <478FADE5.2070601@spellboundnet.com>

Marcus Rueckert wrote:
> On 2008-01-16 22:07:29 -0500, Donavan Pantke wrote:
>   
>> OK, I can understand wanting a buildroot-style prefix for installing RubyGems 
>> itself, but I don't understand the practical need for a buildroot for 
>> installing gems. The only difference I see with a buildroot install and 
>> using --install-dir is that the binaries are dropped in the bin directory of 
>> the install dir instead of the system bindir. Wouldn't the following be 
>> perfectly legit in in a spec file?
>>
>> mkdir -p $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/
>> mkdir -p $RPM_BUILD_ROOT/%{bindir}
>> gem install -i $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/ some_gem
>>     
>
> if you go that road, you want:
> gem install --install-dir=%{buildroot}$(gem environment gemdir) foo.gem
>
>   
>> mv $RPM_BUILD_ROOT/usr/lib/ruby/site_lib/1.8/some_gem_bin 
>> $RPM_BUILD_ROOT/%{bindir}
>>
>> This is the kind of predicatable behavior that I've been trying to accomplish 
>> with my previous patches. Let me know what you think.
>>     
>
> oh it makes a difference.
> if you use --install-dir it copies the bin files to a wrong place. uses
> a different lib dir (and totally ignores stuff like biarch systems)
>   
Can you give a good example gem that exhibits all the bad behavior? At
this point I'm very curious about differentiating behaviors, and want to
see if those problems could be eliminated.
> --build-root just prefixes everything with the temporary root dir and we
> are done. no manually copying stuff around. no constructing paths
> manually. and all that.
>   
If we could get behaviors to match, --build-root would just be a synonym
for --install-dir and --bin-dir (Which I'm in the process of writing).
> now that i upgraded my package to 1.0.1, i will checkout your patches.
> from the previous mail i think i will use some kind of custom default.rb
> during testing. (similar to my -rvendor-specific stuff for non-gem ruby
> libs) so the user gets the normal gem dir and my stuff ends up in the
> vendor_gem dir. but binaries should e.g. always be installed into the
> ruby bin dir. i will report back.
>   
I have some ideas about this, I'll post those off list because it's
totally off-topic for this list.

Donavan
>     darix
>
>   


From thewoolleyman at gmail.com  Thu Jan 17 19:16:14 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Thu, 17 Jan 2008 17:16:14 -0700
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <20080117160351.GB7244@hinegardner.org>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
	<C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>
	<20080117160351.GB7244@hinegardner.org>
Message-ID: <e3af97510801171616h50b5b6b0j66da8801cb4a8b5b@mail.gmail.com>

On Jan 17, 2008 9:03 AM, Jeremy Hinegardner <jeremy at hinegardner.org> wrote:
> Jim, you should give webby a try too : http://webby.rubyforge.org/

I also don't understand what's wrong with webgen (and coincidentally
had someone suggest webby to me this morning).  Webgen is popular,
works fine, and is well-supported by a responsive author (the most
important feature, IMHO).

>
> And I could be intereseted in helping out with the RubyGems
> documentation.

Me too.  I put in a request to edit on the current site, but never
heard anything back.

-- Chad

From drbrain at segment7.net  Fri Jan 18 21:09:31 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Fri, 18 Jan 2008 18:09:31 -0800
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <726376BF-DEA7-4199-9703-269F255B24C1@gmail.com>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>
	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>
	<C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>
	<20080117160351.GB7244@hinegardner.org>
	<726376BF-DEA7-4199-9703-269F255B24C1@gmail.com>
Message-ID: <44D9C61F-0D6D-42DC-997D-AA76B3B0D120@segment7.net>

On Jan 17, 2008, at 08:34 AM, Jim Weirich wrote:
> On Jan 17, 2008, at 11:03 AM, Jeremy Hinegardner wrote:
>> Jim, you should give webby a try too : http://webby.rubyforge.org/
>
> Sigh.  Everytime I mention this, someone suggests a DIFFERENT static
> website generator.  Who would have thought there was so much
> competition in this area.
>
> So, why webby over nanoc and webgen?


While we're sighing over dueling static website generators, why not  
use rdoc*?

This way users can browse documentation with `gem server` even when  
they don't have the internet.

At the very least, it would be nice if we could re-use whatever other  
static files that are generated for documentation in RDoc.

* Yes, there are various problems with RDoc, but this is a forward- 
looking statement, as I have partially addressed that nasty frames  
issue, and am working my way up towards doing something about  
Gem::Specification's undocumented methods.

From djberg96 at gmail.com  Fri Jan 18 21:24:14 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Fri, 18 Jan 2008 19:24:14 -0700
Subject: [Rubygems-developers] Docs found by google seem old.
In-Reply-To: <44D9C61F-0D6D-42DC-997D-AA76B3B0D120@segment7.net>
References: <Pine.GSO.4.64.0801151707120.24884@brains.eng.cse.dmu.ac.uk>	<4AF4967D-D504-4214-974F-EECF052F9C36@segment7.net>	<C0DDF8F7-D1DA-4629-93AC-C8A4B5FFC736@gmail.com>	<20080117160351.GB7244@hinegardner.org>	<726376BF-DEA7-4199-9703-269F255B24C1@gmail.com>
	<44D9C61F-0D6D-42DC-997D-AA76B3B0D120@segment7.net>
Message-ID: <47915F4E.6020208@gmail.com>

Eric Hodel wrote:
> On Jan 17, 2008, at 08:34 AM, Jim Weirich wrote:
>> On Jan 17, 2008, at 11:03 AM, Jeremy Hinegardner wrote:
>>> Jim, you should give webby a try too : http://webby.rubyforge.org/
>> Sigh.  Everytime I mention this, someone suggests a DIFFERENT static
>> website generator.  Who would have thought there was so much
>> competition in this area.
>>
>> So, why webby over nanoc and webgen?
> 
> 
> While we're sighing over dueling static website generators, why not  
> use rdoc*?
> 
> This way users can browse documentation with `gem server` even when  
> they don't have the internet.
> 
> At the very least, it would be nice if we could re-use whatever other  
> static files that are generated for documentation in RDoc.
> 
> * Yes, there are various problems with RDoc, but this is a forward- 
> looking statement, as I have partially addressed that nasty frames  
> issue, and am working my way up towards doing something about  
> Gem::Specification's undocumented methods.

What do you think of this?

http://deveiate.org/projects/Darkfish-Rdoc

Regards,

Dan

From tobiasreif at pinkjuice.com  Tue Jan 22 06:57:46 2008
From: tobiasreif at pinkjuice.com (Tobi Reif)
Date: Tue, 22 Jan 2008 12:57:46 +0100
Subject: [Rubygems-developers] Failing gem commands; reinstalling RubyGems
Message-ID: <20080122115746.GA4616@linux.local>

Hi

These gem commands fail:

$ sudo gem install rails -y
/usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in
`gem_original_require': no such file to load -- sources
+(LoadError)
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/source_info_cache.rb:6
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/remote_installer.rb:12
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems.rb:112:in `manage_gems'
        from /usr/local/bin/gem:10
$ sudo gem install rails --source http://gems.rubyonrails.org
[same error]
$ sudo gem list
[same error]

Updating RubyGems fails too:

$ sudo gem update --system
/usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in
`gem_original_require': no such file to load -- sources
+(LoadError)
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/source_info_cache.rb:6
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/remote_installer.rb:12
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
        from /usr/local/lib/ruby/site_ruby/1.8/rubygems.rb:112:in `manage_gems'
        from /usr/local/bin/gem:10
$

It seems I need to reinstall RubyGems.

What's the cleanest way to uninstall RubyGems?

My installation dir is /usr/local/lib/ruby/site_ruby/gems. Will it
suffice to remove this dir? Or should I also remove the dirs listed
here? http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/122616
Or more?

Here's how I had installed RubyGems:

$ wget http://rubyforge.org/frs/download.php/20989/rubygems-0.9.4.tgz
$ tar -xzf rubygems-0.9.4.tgz
$ su -
# cat > /etc/profile.local
GEM_HOME=/usr/local/lib/ruby/site_ruby/gems
# ln -s /usr/local/bin $GEM_HOME/bin
# cd /home/tobi/del/compile/gems/rubygems-0.9.4/
# export GEM_HOME=/usr/local/lib/ruby/site_ruby/gems
# ruby setup.rb all --prefix=/usr/local --siteruby=/usr/
# gem env
RubyGems Environment:
  - VERSION: 0.9.4 (0.9.4)
  - INSTALLATION DIRECTORY: /usr/local/lib/ruby/site_ruby/gems
  - GEM PATH:
     - /usr/local/lib/ruby/site_ruby/gems
  - REMOTE SOURCES:
     - http://gems.rubyforge.org
# rm -f /usr/local/lib/ruby/site_ruby/gems/source_cache
# gem update --system
# gem install rails --include-dependencies
# gem install mongrel

--siteruby probably was /usr/local/lib/ruby/site_ruby.

Tobi


From martin.krauskopf at gmail.com  Wed Jan 23 07:24:36 2008
From: martin.krauskopf at gmail.com (Martin Krauskopf)
Date: Wed, 23 Jan 2008 13:24:36 +0100
Subject: [Rubygems-developers] Question on GEM_PATH
In-Reply-To: <1ACBC952-C871-49CC-8AF1-B2C2E122C865@segment7.net>
References: <e3af97510712152305o1c946acaya2394583d40ad366@mail.gmail.com>
	<1ACBC952-C871-49CC-8AF1-B2C2E122C865@segment7.net>
Message-ID: <fn7bia$gds$1@ger.gmane.org>

Eric Hodel wrote:
> On Dec 15, 2007, at 23:05 PM, Chad Woolley wrote:
>> When I use GEM_PATH to add an additional local repository to be used,
>> what dirs must be in that repository?  gems?  specifications? docs?
>> cache? others?
> 
> The directory doesn't even need to exist.  RubyGems will create it and  
> everything it needs inside it when you reference it, provided you have  
> permission to do so.

Hi, reviving little old thread,

I've noticed the automatic repository creation on demand. My question is 
whether there is an 'official' way to initialize/create a new repository 
'just now', without waiting for a 'demand'.
My use-case is that in the IDE I'll let user adjust a GEM_PATH. Ff the 
user enter a non-existing or empty repository, I'll show the dialog 
saying something like:

   "The directory is not a valid RubyGems repository, do you want to set
    it up?"  [Yes] [No]

If 'Yes' what should I run? I tried just a 'gem env' and it will 
create/setup gems repositories in all non-repository directories on the 
GEM_PATH.
So may I rely on 'gem env' or is there a more correct way?

Thanks,
	m.


From halostatue at gmail.com  Wed Jan 23 08:39:23 2008
From: halostatue at gmail.com (Austin Ziegler)
Date: Wed, 23 Jan 2008 08:39:23 -0500
Subject: [Rubygems-developers] Passing LDFLAGS?
Message-ID: <9e7db9110801230539l7b11579v91ad71dbce0b529c@mail.gmail.com>

I'm trying to install the FXRuby RubyGem while using Fox installed in
/opt/local via MacPorts and the Leopard-default Ruby. I'm passing the
following:

sudo env CPPFLAGS="-I/opt/local/include" ARCHFLAGS="-arch i386" \
         LDFLAGS="-L/opt/local/lib" gem install fxruby

CPPFLAGS is sent across just fine:

g++ -I. -I. \
        -I/System/Library/Frameworks/Ruby.framework/[...] -I. \
        -DHAVE_SYS_TIME_H -DHAVE_SIGNAL_H \
        -I/usr/local/include/fxscintilla -I/usr/local/include/fox-1.6 \
        -I/opt/local/include/fox-1.6 -I/opt/local/include/fxscintilla \
        -fno-common -arch i386 -Os -pipe -fno-common -O0 -Iinclude \
        -DWITH_FXSCINTILLA -DHAVE_FOX_1_6   -c \
        unregisterOwnedObjects.cpp

But LDFLAGS isn't being respected at all:

cc -arch i386 -pipe -bundle -o fox16.bundle librb.o core_wrap.o \
       dc_wrap.o dialogs_wrap.o frames_wrap.o fx3d_wrap.o FXRbApp.o \
       FXRbDataTarget.o FXRbGLViewer.o FXRuby.o iconlist_wrap.o \
       icons_wrap.o image_wrap.o impl.o label_wrap.o layout_wrap.o \
       list_wrap.o markfuncs.o mdi_wrap.o menu_wrap.o scintilla_wrap.o \
       table_wrap.o text_wrap.o treelist_wrap.o ui_wrap.o \
       unregisterOwnedObjects.o -L"." \
       -L"/System/Library/Frameworks/Ruby.framework/[...]" \
       -L"/usr/local/lib" -L"/usr/X11R6/lib" -L. -arch i386    -lruby \
       -lfxscintilla -lFOX-1.6 -lGLU -lX11 -lXext -lz -lstdc++ \
       -lpthread -ldl -lm

which results in:

    ld: library not found for -lfxscintilla
    collect2: ld returned 1 exit status
    make: *** [fox16.bundle] Error 1

(I have to do "-arch i386" because at least one of the dep libraries in
MacPorts won't build universal, which means that Fox won't build
universal. As soon as I have time, I'll be taking that up with them,
hopefully with a fix.)

Why isn't LDFLAGS being passed along to make in the Gem?

-austin
-- 
Austin Ziegler * halostatue at gmail.com * http://www.halostatue.ca/
               * austin at halostatue.ca * http://www.halostatue.ca/feed/
               * austin at zieglers.ca

From drbrain at segment7.net  Wed Jan 23 16:39:34 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Wed, 23 Jan 2008 13:39:34 -0800
Subject: [Rubygems-developers] Passing LDFLAGS?
In-Reply-To: <9e7db9110801230539l7b11579v91ad71dbce0b529c@mail.gmail.com>
References: <9e7db9110801230539l7b11579v91ad71dbce0b529c@mail.gmail.com>
Message-ID: <69963100-F0D8-4C46-8B9A-849575E7D1D2@segment7.net>

On Jan 23, 2008, at 05:39 AM, Austin Ziegler wrote:

> I'm trying to install the FXRuby RubyGem while using Fox installed in
> /opt/local via MacPorts and the Leopard-default Ruby. I'm passing the
> following:
>
> sudo env CPPFLAGS="-I/opt/local/include" ARCHFLAGS="-arch i386" \
>         LDFLAGS="-L/opt/local/lib" gem install fxruby
>
> CPPFLAGS is sent across just fine:
>
> g++ -I. -I. \
>        -I/System/Library/Frameworks/Ruby.framework/[...] -I. \
>        -DHAVE_SYS_TIME_H -DHAVE_SIGNAL_H \
>        -I/usr/local/include/fxscintilla -I/usr/local/include/fox-1.6 \
>        -I/opt/local/include/fox-1.6 -I/opt/local/include/fxscintilla \
>        -fno-common -arch i386 -Os -pipe -fno-common -O0 -Iinclude \
>        -DWITH_FXSCINTILLA -DHAVE_FOX_1_6   -c \
>        unregisterOwnedObjects.cpp
>
> But LDFLAGS isn't being respected at all:
>
> cc -arch i386 -pipe -bundle -o fox16.bundle librb.o core_wrap.o \
>       dc_wrap.o dialogs_wrap.o frames_wrap.o fx3d_wrap.o FXRbApp.o \
>       FXRbDataTarget.o FXRbGLViewer.o FXRuby.o iconlist_wrap.o \
>       icons_wrap.o image_wrap.o impl.o label_wrap.o layout_wrap.o \
>       list_wrap.o markfuncs.o mdi_wrap.o menu_wrap.o  
> scintilla_wrap.o \
>       table_wrap.o text_wrap.o treelist_wrap.o ui_wrap.o \
>       unregisterOwnedObjects.o -L"." \
>       -L"/System/Library/Frameworks/Ruby.framework/[...]" \
>       -L"/usr/local/lib" -L"/usr/X11R6/lib" -L. -arch i386    -lruby \
>       -lfxscintilla -lFOX-1.6 -lGLU -lX11 -lXext -lz -lstdc++ \
>       -lpthread -ldl -lm
>
> which results in:
>
>    ld: library not found for -lfxscintilla
>    collect2: ld returned 1 exit status
>    make: *** [fox16.bundle] Error 1
>
> (I have to do "-arch i386" because at least one of the dep libraries  
> in
> MacPorts won't build universal, which means that Fox won't build
> universal. As soon as I have time, I'll be taking that up with them,
> hopefully with a fix.)
>
> Why isn't LDFLAGS being passed along to make in the Gem?

Does it get passed without sudo?

AFAIK, nothing is changed in the LDFLAGS within rubygems.

From drbrain at segment7.net  Wed Jan 23 16:40:27 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Wed, 23 Jan 2008 13:40:27 -0800
Subject: [Rubygems-developers] Question on GEM_PATH
In-Reply-To: <fn7bia$gds$1@ger.gmane.org>
References: <e3af97510712152305o1c946acaya2394583d40ad366@mail.gmail.com>
	<1ACBC952-C871-49CC-8AF1-B2C2E122C865@segment7.net>
	<fn7bia$gds$1@ger.gmane.org>
Message-ID: <24E37BFB-482A-4930-9B64-8E2284BD1A2D@segment7.net>

On Jan 23, 2008, at 04:24 AM, Martin Krauskopf wrote:
> Eric Hodel wrote:
>> On Dec 15, 2007, at 23:05 PM, Chad Woolley wrote:
>>> When I use GEM_PATH to add an additional local repository to be  
>>> used,
>>> what dirs must be in that repository?  gems?  specifications? docs?
>>> cache? others?
>>
>> The directory doesn't even need to exist.  RubyGems will create it  
>> and
>> everything it needs inside it when you reference it, provided you  
>> have
>> permission to do so.
>
> Hi, reviving little old thread,
>
> I've noticed the automatic repository creation on demand. My  
> question is
> whether there is an 'official' way to initialize/create a new  
> repository
> 'just now', without waiting for a 'demand'.
> My use-case is that in the IDE I'll let user adjust a GEM_PATH. Ff the
> user enter a non-existing or empty repository, I'll show the dialog
> saying something like:
>
>   "The directory is not a valid RubyGems repository, do you want to  
> set
>    it up?"  [Yes] [No]
>
> If 'Yes' what should I run? I tried just a 'gem env' and it will
> create/setup gems repositories in all non-repository directories on  
> the
> GEM_PATH.
> So may I rely on 'gem env' or is there a more correct way?

Gem.ensuregem_subdirectories

From drbrain at segment7.net  Wed Jan 23 16:41:56 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Wed, 23 Jan 2008 13:41:56 -0800
Subject: [Rubygems-developers] removing a source
In-Reply-To: <e9a952090801170748m4bf0c31eg2f0e25942fba580a@mail.gmail.com>
References: <e9a952090801170748m4bf0c31eg2f0e25942fba580a@mail.gmail.com>
Message-ID: <36D88F53-1743-4673-A060-5BBBB81FF9CD@segment7.net>

On Jan 17, 2008, at 07:48 AM, Todd Fisher wrote:
> I'm having trouble removing a specific source within my local gem  
> repository.  Here's what I did.
>
> gem sources --add http://gems.myinternalhost.com:8808/
>
> Now later I'm in a place where I can no longer access  
> gems.myinternalhost.com:8808/  or it has gone down.  I'd like to  
> install a new gem from gems.rubyforge.org.  I can do this but the  
> only way I found was by adding the --source gems.rubyforge.org,  
> causing the operation to ignore all other sources.  But what I would  
> really like to do is remove the bad source...
>
> I've tried this:
>
> rm $GEM_HOME/source_cache
> rm -f $GEM_HOME/cache/*
>
> gem sources --remove http://gems.myinternalhost.com:8808/
> gem source --remove http://gems.myinternalhost.com:8808/
>
> Here's what I get:
>
> sudo gem source --remove http://gems.myinternalhost.com:8808/
> Bulk updating Gem source index for: http://gems.rubyforge.org
> ERROR:  While executing gem ... (Gem::RemoteFetcher::FetchError)
>     getaddrinfo: No address associated with nodename (SocketError)
>         getting size of http://gems.myinternalhost.com:8808/Marshal. 
> 4.8
>
>  gem --version
> 1.0.1
>
> Any ideas?

Does `gem sources -c` fix the problem?

From halostatue at gmail.com  Wed Jan 23 17:55:17 2008
From: halostatue at gmail.com (Austin Ziegler)
Date: Wed, 23 Jan 2008 17:55:17 -0500
Subject: [Rubygems-developers] Passing LDFLAGS?
In-Reply-To: <69963100-F0D8-4C46-8B9A-849575E7D1D2@segment7.net>
References: <9e7db9110801230539l7b11579v91ad71dbce0b529c@mail.gmail.com>
	<69963100-F0D8-4C46-8B9A-849575E7D1D2@segment7.net>
Message-ID: <9e7db9110801231455v42ed5797t26fa0e5646331f16@mail.gmail.com>

On Jan 23, 2008 4:39 PM, Eric Hodel <drbrain at segment7.net> wrote:
> On Jan 23, 2008, at 05:39 AM, Austin Ziegler wrote:
>> Why isn't LDFLAGS being passed along to make in the Gem?
> Does it get passed without sudo?

No difference as far as I can tell. It looks like I *could* sneak it
in with ARCHFLAGS, but that doesn't seem right.

  % sudo bash
  # CPPFLAGS="-I/opt/local/include" ARCHFLAGS="-arch i386" \
        LDFLAGS="-L/opt/local/lib" gem install fxruby

g++ -I. -I. -I/System/Library/Frameworks/Ruby.framework/[...] \
	-I. -DHAVE_SYS_TIME_H -DHAVE_SIGNAL_H \
	-I/usr/local/include/fxscintilla -I/usr/local/include/fox-1.6 \
	-I/opt/local/include/fox-1.6 -I/opt/local/include/fxscintilla \
	-fno-common -arch i386 -Os -pipe -fno-common -O0 -Iinclude \
	-DWITH_FXSCINTILLA -DHAVE_FOX_1_6   -c unregisterOwnedObjects.cpp
cc -arch i386 -pipe -bundle -o fox16.bundle librb.o core_wrap.o \
	 dc_wrap.o dialogs_wrap.o frames_wrap.o fx3d_wrap.o FXRbApp.o \
	 FXRbDataTarget.o FXRbGLViewer.o FXRuby.o iconlist_wrap.o \
	 icons_wrap.o image_wrap.o impl.o label_wrap.o layout_wrap.o \
	 list_wrap.o markfuncs.o mdi_wrap.o menu_wrap.o scintilla_wrap.o \
	 table_wrap.o text_wrap.o treelist_wrap.o ui_wrap.o \
	 unregisterOwnedObjects.o -L"."
	 -L"/System/Library/Frameworks/Ruby.framework/[...]" \
	 -L"/usr/local/lib" -L"/usr/X11R6/lib" -L. -arch i386 -lruby \
	 -lfxscintilla -lFOX-1.6 -lGLU -lX11 -lXext -lz -lstdc++ -lpthread \
	 -ldl -lm
ld: library not found for -lfxscintilla
collect2: ld returned 1 exit status

> AFAIK, nothing is changed in the LDFLAGS within rubygems.

-austin
-- 
Austin Ziegler * halostatue at gmail.com * http://www.halostatue.ca/
               * austin at halostatue.ca * http://www.halostatue.ca/feed/
               * austin at zieglers.ca

From martin.krauskopf at gmail.com  Thu Jan 24 04:56:30 2008
From: martin.krauskopf at gmail.com (Martin Krauskopf)
Date: Thu, 24 Jan 2008 10:56:30 +0100
Subject: [Rubygems-developers] Question on GEM_PATH
In-Reply-To: <24E37BFB-482A-4930-9B64-8E2284BD1A2D@segment7.net>
References: <e3af97510712152305o1c946acaya2394583d40ad366@mail.gmail.com>	<1ACBC952-C871-49CC-8AF1-B2C2E122C865@segment7.net>	<fn7bia$gds$1@ger.gmane.org>
	<24E37BFB-482A-4930-9B64-8E2284BD1A2D@segment7.net>
Message-ID: <fn9n8h$s3b$1@ger.gmane.org>

Eric Hodel wrote:
> On Jan 23, 2008, at 04:24 AM, Martin Krauskopf wrote:
>> Hi, reviving little old thread,
>>
>> I've noticed the automatic repository creation on demand. My  
>> question is
>> whether there is an 'official' way to initialize/create a new  
>> repository
>> 'just now', without waiting for a 'demand'.
>> My use-case is that in the IDE I'll let user adjust a GEM_PATH. Ff the
>> user enter a non-existing or empty repository, I'll show the dialog
>> saying something like:
>>
>>   "The directory is not a valid RubyGems repository, do you want to  
>> set
>>    it up?"  [Yes] [No]
>>
>> If 'Yes' what should I run? I tried just a 'gem env' and it will
>> create/setup gems repositories in all non-repository directories on  
>> the
>> GEM_PATH.
>> So may I rely on 'gem env' or is there a more correct way?
> 
> Gem.ensuregem_subdirectories

Thanks Eric,

the method is probably usable (== public) since 1.0.x(?) (I haven't take 
a look into the history). I need to be backward compatible with at least 
0.9.4+.
Probably I'll use a workaround like:

   module Gem; class << self; public :ensure_gem_subdirectories; end; end
   Gem.ensure_gem_subdirectories '/new/gem/repo'

which should work for 0.9.x as well.

Thanks for the tip,

	m.


From thewoolleyman at gmail.com  Fri Jan 25 13:04:52 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Fri, 25 Jan 2008 11:04:52 -0700
Subject: [Rubygems-developers] warnings from open-uri
Message-ID: <e3af97510801251004l7beff3e4ga0b0a71a79f692a7@mail.gmail.com>

Can I get some feedback on this:

http://rubyforge.org/tracker/index.php?func=detail&aid=17042&group_id=126&atid=575

It spews warnings constantly during our test and app execution.  I can
take a shot at a patch, but I'm unsure where to start.  What is the
issue, and what is the right way to fix it?  Namespace the rubygems
version?  How would that work, and what would it break?

Thanks,
-- Chad

From martin.krauskopf at gmail.com  Sun Jan 27 18:22:47 2008
From: martin.krauskopf at gmail.com (Martin Krauskopf)
Date: Mon, 28 Jan 2008 00:22:47 +0100
Subject: [Rubygems-developers] GEM_PATH is ignored when looking for
	installed dependencies
Message-ID: <fnj3k2$ptj$1@ger.gmane.org>

Hi,

I'm encountering strange behavior wrt. to Gem Path. Best to show an
example (ruby-debug depends on ruby-debug-base):

   $ gem env
   RubyGems Environment:
   [...]
     - INSTALLATION DIRECTORY: /space/ruby/gem-repo
   [...]
     - GEM PATHS:
       - /space/ruby/gem-repo-c
       - /space/ruby/gem-repo
   [...]

   $ ls -d /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3
   /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3
   $ ls -d /space/ruby/gem-repo/gems/ruby-debug-*
   ls: /space/ruby/gem-repo/gems/ruby-debug-*: No such file or directory
   $ gem install ruby-debug -v 0.9.3 # ERROR HERE IN INSTALLING -base
   Building native extensions.  This could take a while...
   Successfully installed ruby-debug-base-0.9.3
   Successfully installed ruby-debug-0.9.3
   2 gems installed
   $ ls -d /space/ruby/gem-repo/gems/ruby-debug-*
   /space/ruby/gem-repo/gems/ruby-debug-0.9.3
   /space/ruby/gem-repo/gems/ruby-debug-base-0.9.3 # REDUNDANT

Why is ruby-debug-base installed second time into main repository
(/space/ruby/gem-repo) when it is already installed in the 'secondary'
repository (/space/ruby/gem-repo-c).

$ gem install ruby-debug -v 0.9.3

in the example above should not install ruby-debug-base (second time).
Did I misunderstood Gem Path usage or is it bug?

Should not a DependencyInstaller#install fill source_index from all
repositories on the Gem Path?

Thanks,
	m.


From thewoolleyman at gmail.com  Mon Jan 28 03:11:36 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Mon, 28 Jan 2008 01:11:36 -0700
Subject: [Rubygems-developers] GEM_PATH is ignored when looking for
	installed dependencies
In-Reply-To: <fnj3k2$ptj$1@ger.gmane.org>
References: <fnj3k2$ptj$1@ger.gmane.org>
Message-ID: <e3af97510801280011m5c24eb99ma5a75e0cda14ca9@mail.gmail.com>

On Jan 27, 2008 4:22 PM, Martin Krauskopf <martin.krauskopf at gmail.com> wrote:
>    $ ls -d /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3
>    /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3
>    $ ls -d /space/ruby/gem-repo/gems/ruby-debug-*
>    ls: /space/ruby/gem-repo/gems/ruby-debug-*: No such file or directory
>    $ gem install ruby-debug -v 0.9.3 # ERROR HERE IN INSTALLING -base
>    Building native extensions.  This could take a while...
>    Successfully installed ruby-debug-base-0.9.3
>    Successfully installed ruby-debug-0.9.3
>    2 gems installed
>    $ ls -d /space/ruby/gem-repo/gems/ruby-debug-*
>    /space/ruby/gem-repo/gems/ruby-debug-0.9.3
>    /space/ruby/gem-repo/gems/ruby-debug-base-0.9.3 # REDUNDANT

That looks like unexpected behavior and a bug, unless there's
something with the ruby-debug-base already installed in gem-repo-c.

What other files are already under /space/ruby/gem-repo-c/gems for
ruby-debug-base (are there specs/cache/etc)?  I don't even know if
this matters - what exactly is used when creating the source cache -
the gem files, the spec, ???

Also, what are the exact commands you used to install
/space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3?  A fully
reproducible error scenario would help.

Thanks,
-- Chad

From tobiasreif at pinkjuice.com  Mon Jan 28 03:42:00 2008
From: tobiasreif at pinkjuice.com (Tobi Reif)
Date: Mon, 28 Jan 2008 09:42:00 +0100
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling	RubyGems
In-Reply-To: <20080122115746.GA4616@linux.local>
References: <20080122115746.GA4616@linux.local>
Message-ID: <20080128084200.GA3989@linux.local>

On Tue 2008-01-22 I wrote:
> Hi
> 
> These gem commands fail:
> 
> $ sudo gem install rails -y
> /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in
> `gem_original_require': no such file to load -- sources
> +(LoadError)
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/source_info_cache.rb:6
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/remote_installer.rb:12
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems.rb:112:in `manage_gems'
>         from /usr/local/bin/gem:10
> $ sudo gem install rails --source http://gems.rubyonrails.org
> [same error]
> $ sudo gem list
> [same error]
> 
> Updating RubyGems fails too:
> 
> $ sudo gem update --system
> /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in
> `gem_original_require': no such file to load -- sources
> +(LoadError)
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/source_info_cache.rb:6
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/remote_installer.rb:12
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `gem_original_require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:27:in `require'
>         from /usr/local/lib/ruby/site_ruby/1.8/rubygems.rb:112:in `manage_gems'
>         from /usr/local/bin/gem:10
> $
> 
> It seems I need to reinstall RubyGems.
> 
> What's the cleanest way to uninstall RubyGems?
> 
> My installation dir is /usr/local/lib/ruby/site_ruby/gems. Will it
> suffice to remove this dir? Or should I also remove the dirs listed
> here? http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/122616
> Or more?
> 
> Here's how I had installed RubyGems:
> 
> $ wget http://rubyforge.org/frs/download.php/20989/rubygems-0.9.4.tgz
> $ tar -xzf rubygems-0.9.4.tgz
> $ su -
> # cat > /etc/profile.local
> GEM_HOME=/usr/local/lib/ruby/site_ruby/gems
> # ln -s /usr/local/bin $GEM_HOME/bin
> # cd /home/tobi/del/compile/gems/rubygems-0.9.4/
> # export GEM_HOME=/usr/local/lib/ruby/site_ruby/gems
> # ruby setup.rb all --prefix=/usr/local --siteruby=/usr/
> # gem env
> RubyGems Environment:
>   - VERSION: 0.9.4 (0.9.4)
>   - INSTALLATION DIRECTORY: /usr/local/lib/ruby/site_ruby/gems
>   - GEM PATH:
>      - /usr/local/lib/ruby/site_ruby/gems
>   - REMOTE SOURCES:
>      - http://gems.rubyforge.org
> # rm -f /usr/local/lib/ruby/site_ruby/gems/source_cache
> # gem update --system
> # gem install rails --include-dependencies
> # gem install mongrel
> 
> --siteruby probably was /usr/local/lib/ruby/site_ruby.

Anyone?

Tobi


From thewoolleyman at gmail.com  Mon Jan 28 03:52:17 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Mon, 28 Jan 2008 01:52:17 -0700
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling RubyGems
In-Reply-To: <20080122115746.GA4616@linux.local>
References: <20080122115746.GA4616@linux.local>
Message-ID: <e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>

On Jan 22, 2008 4:57 AM, Tobi Reif <tobiasreif at pinkjuice.com> wrote:
> It seems I need to reinstall RubyGems.
>
> What's the cleanest way to uninstall RubyGems?

If you are only interested in getting a working gems install, I would
download the latest rubygems tar.gz, run setup.rb manually and forget
about update system via an older 'gem update --system' command.

If you want to know why the 'gem update --system' command is failing
in rubygems 0.9.4, then try to reproduce the problem with the latest
(or next-to-latest) rubygems release.  If it is only a bug in the
older release, which has been fixed in newer releases, there's not
much to do...

-- Chad

From martin.krauskopf at gmail.com  Mon Jan 28 05:38:18 2008
From: martin.krauskopf at gmail.com (Martin Krauskopf)
Date: Mon, 28 Jan 2008 11:38:18 +0100
Subject: [Rubygems-developers] GEM_PATH is ignored when looking for
	installed dependencies
In-Reply-To: <e3af97510801280011m5c24eb99ma5a75e0cda14ca9@mail.gmail.com>
References: <fnj3k2$ptj$1@ger.gmane.org>
	<e3af97510801280011m5c24eb99ma5a75e0cda14ca9@mail.gmail.com>
Message-ID: <fnkb6k$onu$1@ger.gmane.org>

Chad Woolley wrote:
> On Jan 27, 2008 4:22 PM, Martin Krauskopf <martin.krauskopf at gmail.com> wrote:
>>    $ ls -d /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3
>>    /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3
>>    $ ls -d /space/ruby/gem-repo/gems/ruby-debug-*
>>    ls: /space/ruby/gem-repo/gems/ruby-debug-*: No such file or directory
>>    $ gem install ruby-debug -v 0.9.3 # ERROR HERE IN INSTALLING -base
>>    Building native extensions.  This could take a while...
>>    Successfully installed ruby-debug-base-0.9.3
>>    Successfully installed ruby-debug-0.9.3
>>    2 gems installed
>>    $ ls -d /space/ruby/gem-repo/gems/ruby-debug-*
>>    /space/ruby/gem-repo/gems/ruby-debug-0.9.3
>>    /space/ruby/gem-repo/gems/ruby-debug-base-0.9.3 # REDUNDANT
> 
> That looks like unexpected behavior and a bug, unless there's
> something with the ruby-debug-base already installed in gem-repo-c.

No it was not. See the scenario below.

> What other files are already under /space/ruby/gem-repo-c/gems for
> ruby-debug-base (are there specs/cache/etc)?  I don't even know if
> this matters - what exactly is used when creating the source cache -
> the gem files, the spec, ???

source index is build from spec files in the specification directory. 
But the problem seems to be that it is build just from the main 
repository in Gem::DependencyInstaller#installer. Other repositories are 
ignored differently to run-time where they are 'seen' as expected.

> Also, what are the exact commands you used to install
> /space/ruby/gem-repo-c/gems/ruby-debug-base-0.9.3?

See below.

> A fully reproducible error scenario would help.

Here it is. Reproducible scenario with new fresh repos. Should be 
reproducible on whatever machine:

   $ mkdir -p /space/ruby/gem-repos
   $ cd /space/ruby/gem-repos
   $ export 
GEM_PATH="/space/ruby/gem-repos/gems:/space/ruby/gem-repos/gems-c"
   $ export GEM_HOME="/space/ruby/gem-repos/gems"
   $ gem env # this will also initializes/creates repositories
   RubyGems Environment:
     - RUBYGEMS VERSION: 1.0.1 (1.0.1)
     - RUBY VERSION: 1.8.6 (2007-09-24 patchlevel 111) [i686-linux]
     - INSTALLATION DIRECTORY: /space/ruby/gem-repos/gems
     - RUBY EXECUTABLE: /space/ruby/ruby-1.8.6-p111/bin/ruby1.8.6-p111
     - RUBYGEMS PLATFORMS:
       - ruby
       - x86-linux
     - GEM PATHS:
        - /space/ruby/gem-repos/gems
        - /space/ruby/gem-repos/gems-c
     - GEM CONFIGURATION:
        - :update_sources => true
        - :verbose => true
        - :benchmark => false
        - :backtrace => false
        - :bulk_threshold => 1000
        - "gem" => "--no-rdoc --no-ri"
     - REMOTE SOURCES:
        - http://gems.rubyforge.org
   $ ls
   gems/  gems-c/
   $ gem install -i /space/ruby/gem-repos/gems-c ruby-debug-base -v 0.9.3
   Bulk updating Gem source index for: http://gems.rubyforge.org
   Building native extensions.  This could take a while...
   Successfully installed ruby-debug-base-0.9.3
   1 gem installed
   $ gem install ruby-debug -v 0.9.3
   Building native extensions.  This could take a while...
   Successfully installed ruby-debug-base-0.9.3 # [1]
   Successfully installed ruby-debug-0.9.3
   2 gems installed

[1] this is wrong, already installed in repository on Gem Path.

So now we have ruby-debug-base installed two-times. Also note if you do:

$ gem uninstall ruby-debug-base

It will uninstall the gem from the main repository and leave the second 
copy in the 'secondary' repository - which is quite ok; and debugger 
still works as expected. So the second installation is really redundant.

Also this can be reproduce with all other gems, e.g. rake.

Regards,
	m.

PS: I could try to peek in and fix this if nobody else. But first I will 
need to checkout trunk and see how the process of tests, etc. works in 
RubyGems development process. I suppose that the fix will be simple as 
traversing all the repositories, not just main, when filling source_index.


From tobiasreif at pinkjuice.com  Mon Jan 28 08:20:29 2008
From: tobiasreif at pinkjuice.com (Tobi Reif)
Date: Mon, 28 Jan 2008 14:20:29 +0100
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling	RubyGems
In-Reply-To: <e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>
References: <20080122115746.GA4616@linux.local>
	<e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>
Message-ID: <20080128132029.GB7971@linux.local>

On Mon 2008-01-28 Chad Woolley wrote:
> On Jan 22, 2008 4:57 AM, Tobi Reif <tobiasreif at pinkjuice.com> wrote:
> > It seems I need to reinstall RubyGems.
> >
> > What's the cleanest way to uninstall RubyGems?
> 
> If you are only interested in getting a working gems install, I would
> download the latest rubygems tar.gz, run setup.rb manually and forget
> about update system via an older 'gem update --system' command.
> 
> If you want to know why the 'gem update --system' command is failing
> in rubygems 0.9.4, then try to reproduce the problem with the latest
> (or next-to-latest) rubygems release.  If it is only a bug in the
> older release, which has been fixed in newer releases, there's not
> much to do...

Thanks for your reply Chad.

I agree, it seems I'll have to go the .tgz/setup.rb route. But
before I do that I'd like to uninstall RubyGems completely. What would
be the complete list of dirs and files to delete? (according to how I
had installed RubyGems - see my original post).

RubyGems developers etc, is the following the complete list of
files/dirs to delete?

Files:
/usr/local/bin/gem
/usr/local/bin/gemri
/usr/local/bin/gem_mirror
/usr/local/bin/gem_server
/usr/local/bin/gemwhich
/usr/local/bin/gemlock
/usr/local/bin/update_rubygems
/usr/local/bin/index_gem_repository.rb
/usr/local/lib/ruby/site_ruby/1.8/ubygems.rb
/usr/local/lib/ruby/site_ruby/1.8/rubygems.rb
/usr/local/lib/ruby/site_ruby/1.8/gemconfigure.rb

Dirs:
/usr/local/lib/ruby/gems
/usr/local/lib/ruby/site_ruby/1.8/rubygems
/usr/local/lib/ruby/site_ruby/gems
/home/tobi/.gem

Or is there some automated way to uninstall RubyGems?

Tobi


From thewoolleyman at gmail.com  Mon Jan 28 12:57:20 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Mon, 28 Jan 2008 10:57:20 -0700
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling RubyGems
In-Reply-To: <20080128132029.GB7971@linux.local>
References: <20080122115746.GA4616@linux.local>
	<e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>
	<20080128132029.GB7971@linux.local>
Message-ID: <e3af97510801280957u4c9a6affje4f472474c38a386@mail.gmail.com>

On 1/28/08, Tobi Reif <tobiasreif at pinkjuice.com> wrote:
> Or is there some automated way to uninstall RubyGems?

I don't think there's an automated way, there are some threads in the
archives with links to methods.

However, installing over an old version usually works fine and cleans
up everything that needs to be cleaned up - I'd just try that.

-- Chad

From jeremy at hinegardner.org  Mon Jan 28 13:09:17 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Mon, 28 Jan 2008 11:09:17 -0700
Subject: [Rubygems-developers] rubygems.org down ?
Message-ID: <20080128180917.GU7244@hinegardner.org>

I'm getting a 503 Service Temporarily Unavailable from rubygems.org.  Is
this a planned outage?

enjoy,

-jeremy

--
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From chad at chadfowler.com  Mon Jan 28 13:13:34 2008
From: chad at chadfowler.com (Chad Fowler)
Date: Mon, 28 Jan 2008 12:13:34 -0600
Subject: [Rubygems-developers] rubygems.org down ?
In-Reply-To: <20080128180917.GU7244@hinegardner.org>
References: <20080128180917.GU7244@hinegardner.org>
Message-ID: <c67184de0801281013x612672d9la276592985f845ed@mail.gmail.com>

My fault.  Fixing.

Thanks,
Chad

On 1/28/08, Jeremy Hinegardner <jeremy at hinegardner.org> wrote:
> I'm getting a 503 Service Temporarily Unavailable from rubygems.org.  Is
> this a planned outage?
>
> enjoy,
>
> -jeremy
>
> --
> ========================================================================
>  Jeremy Hinegardner                              jeremy at hinegardner.org
>
> _______________________________________________
> Rubygems-developers mailing list
> Rubygems-developers at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubygems-developers
>

From tobiasreif at pinkjuice.com  Mon Jan 28 14:12:55 2008
From: tobiasreif at pinkjuice.com (Tobi Reif)
Date: Mon, 28 Jan 2008 20:12:55 +0100
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling	RubyGems
In-Reply-To: <e3af97510801280957u4c9a6affje4f472474c38a386@mail.gmail.com>
References: <20080122115746.GA4616@linux.local>
	<e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>
	<20080128132029.GB7971@linux.local>
	<e3af97510801280957u4c9a6affje4f472474c38a386@mail.gmail.com>
Message-ID: <20080128191255.GA3988@linux.local>

On Mon 2008-01-28 Chad Woolley wrote:
[...]
> However, installing over an old version [...] cleans
> up everything that needs to be cleaned up [...]

Hmmm, that's hard to imagine in cases where the existing RubyGems
installation is in a different location than the new one.

I really would like to know what the recommended way is for
uninstalling / reinstalling RubyGems. Rich, Jim, Chad, Eric, what's
the "official" recommendation?

Tobi


From Daniel.Berger at qwest.com  Mon Jan 28 14:57:04 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Mon, 28 Jan 2008 13:57:04 -0600
Subject: [Rubygems-developers] Mirror fallbacks (was: rubygems.org down ?)
In-Reply-To: <20080128180917.GU7244@hinegardner.org>
References: <20080128180917.GU7244@hinegardner.org>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D2F24@ITOMAE2KM01.AD.QINTRA.COM>

> -----Original Message-----
> From: rubygems-developers-bounces at rubyforge.org 
> [mailto:rubygems-developers-bounces at rubyforge.org] On Behalf 
> Of Jeremy Hinegardner
> Sent: Monday, January 28, 2008 11:09 AM
> To: rubygems-developers at rubyforge.org
> Subject: [Rubygems-developers] rubygems.org down ?
> 
> I'm getting a 503 Service Temporarily Unavailable from 
> rubygems.org.  Is this a planned outage?

Is there a way to configure rubygems to fallback to a mirror in the
event of an error like this?

With Perl's CPAN, for example, you select a bunch of mirrors the first
time you run the CPAN shell. From then on, if there's a problem
downloading from your first selection it defaults to the next mirror,
and so on, until it either succeeds or goes through all of your mirrors.

Can we do that? If so, how?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From jeremy at hinegardner.org  Mon Jan 28 15:16:58 2008
From: jeremy at hinegardner.org (Jeremy Hinegardner)
Date: Mon, 28 Jan 2008 13:16:58 -0700
Subject: [Rubygems-developers] Mirror fallbacks (was: rubygems.org
	down	?)
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D2F24@ITOMAE2KM01.AD.QINTRA.COM>
References: <20080128180917.GU7244@hinegardner.org>
	<7524A45A1A5B264FA4809E2156496CFB023D2F24@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <20080128201658.GV7244@hinegardner.org>

On Mon, Jan 28, 2008 at 01:57:04PM -0600, Berger, Daniel wrote:
> > -----Original Message-----
> > From: rubygems-developers-bounces at rubyforge.org 
> > [mailto:rubygems-developers-bounces at rubyforge.org] On Behalf 
> > Of Jeremy Hinegardner
> > Sent: Monday, January 28, 2008 11:09 AM
> > To: rubygems-developers at rubyforge.org
> > Subject: [Rubygems-developers] rubygems.org down ?
> > 
> > I'm getting a 503 Service Temporarily Unavailable from 
> > rubygems.org.  Is this a planned outage?
> 
> Is there a way to configure rubygems to fallback to a mirror in the
> event of an error like this?
> 
> With Perl's CPAN, for example, you select a bunch of mirrors the first
> time you run the CPAN shell. From then on, if there's a problem
> downloading from your first selection it defaults to the next mirror,
> and so on, until it either succeeds or goes through all of your mirrors.
> 
> Can we do that? If so, how?

Correct me if I'm wrong, but rubygems.org is not where the gems are
distributed.  They are hit via gems.rubyforge.org. Those two addresses
resolve to different IP's.  rubygems.org is the documentation.

enjoy,

-jeremy

-- 
========================================================================
 Jeremy Hinegardner                              jeremy at hinegardner.org 


From Daniel.Berger at qwest.com  Mon Jan 28 15:22:46 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Mon, 28 Jan 2008 14:22:46 -0600
Subject: [Rubygems-developers] Mirror fallbacks (was: rubygems.orgdown	?)
In-Reply-To: <20080128201658.GV7244@hinegardner.org>
References: <20080128180917.GU7244@hinegardner.org><7524A45A1A5B264FA4809E2156496CFB023D2F24@ITOMAE2KM01.AD.QINTRA.COM>
	<20080128201658.GV7244@hinegardner.org>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D2F27@ITOMAE2KM01.AD.QINTRA.COM>

> -----Original Message-----
> From: rubygems-developers-bounces at rubyforge.org 
> [mailto:rubygems-developers-bounces at rubyforge.org] On Behalf 
> Of Jeremy Hinegardner
> Sent: Monday, January 28, 2008 1:17 PM
> To: rubygems-developers at rubyforge.org
> Subject: Re: [Rubygems-developers] Mirror fallbacks (was: 
> rubygems.orgdown ?)
> 
> On Mon, Jan 28, 2008 at 01:57:04PM -0600, Berger, Daniel wrote:
> > > -----Original Message-----
> > > From: rubygems-developers-bounces at rubyforge.org
> > > [mailto:rubygems-developers-bounces at rubyforge.org] On Behalf Of 
> > > Jeremy Hinegardner
> > > Sent: Monday, January 28, 2008 11:09 AM
> > > To: rubygems-developers at rubyforge.org
> > > Subject: [Rubygems-developers] rubygems.org down ?
> > > 
> > > I'm getting a 503 Service Temporarily Unavailable from 
> rubygems.org.  
> > > Is this a planned outage?
> > 
> > Is there a way to configure rubygems to fallback to a mirror in the 
> > event of an error like this?
> > 
> > With Perl's CPAN, for example, you select a bunch of 
> mirrors the first 
> > time you run the CPAN shell. From then on, if there's a problem 
> > downloading from your first selection it defaults to the 
> next mirror, 
> > and so on, until it either succeeds or goes through all of 
> your mirrors.
> > 
> > Can we do that? If so, how?
> 
> Correct me if I'm wrong, but rubygems.org is not where the 
> gems are distributed.  They are hit via gems.rubyforge.org. 
> Those two addresses resolve to different IP's.  rubygems.org 
> is the documentation.
> 
> enjoy,
> 
> -jeremy

Oh, geez, I misread that.

But, I'm still curious about my original question. :)

Thanks,

Dan 


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From thewoolleyman at gmail.com  Mon Jan 28 15:48:03 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Mon, 28 Jan 2008 13:48:03 -0700
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling RubyGems
In-Reply-To: <20080128191255.GA3988@linux.local>
References: <20080122115746.GA4616@linux.local>
	<e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>
	<20080128132029.GB7971@linux.local>
	<e3af97510801280957u4c9a6affje4f472474c38a386@mail.gmail.com>
	<20080128191255.GA3988@linux.local>
Message-ID: <e3af97510801281248n1e32288yf7a5a39ec427c2b5@mail.gmail.com>

On 1/28/08, Tobi Reif <tobiasreif at pinkjuice.com> wrote:
> On Mon 2008-01-28 Chad Woolley wrote:
> [...]
> > However, installing over an old version [...] cleans
> > up everything that needs to be cleaned up [...]
>
> Hmmm, that's hard to imagine in cases where the existing RubyGems
> installation is in a different location than the new one.

Ah, sorry I forgot that part.

>
> I really would like to know what the recommended way is for
> uninstalling / reinstalling RubyGems. Rich, Jim, Chad, Eric, what's
> the "official" recommendation?

Yes, I'd like to hear the latest and greatest official uninstall
procedure as well...

From tom at infoether.com  Mon Jan 28 19:38:20 2008
From: tom at infoether.com (Tom Copeland)
Date: Mon, 28 Jan 2008 19:38:20 -0500
Subject: [Rubygems-developers] Mirror fallbacks (was:	rubygems.orgdown	?)
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D2F27@ITOMAE2KM01.AD.QINTRA.COM>
References: <20080128180917.GU7244@hinegardner.org>
	<7524A45A1A5B264FA4809E2156496CFB023D2F24@ITOMAE2KM01.AD.QINTRA.COM>
	<20080128201658.GV7244@hinegardner.org>
	<7524A45A1A5B264FA4809E2156496CFB023D2F27@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <1201567100.11241.1.camel@bugs.hal>


On Mon, 2008-01-28 at 14:22 -0600, Berger, Daniel wrote:
> > > With Perl's CPAN, for example, you select a bunch of 
> > mirrors the first 
> > > time you run the CPAN shell. From then on, if there's a problem 
> > > downloading from your first selection it defaults to the 
> > next mirror, 
> > > and so on, until it either succeeds or goes through all of 
> > your mirrors.
> > > 
> > > Can we do that? If so, how?

Yup.  We'd need to keep a list of the current mirrors on
gems.rubyforge.org/some/uri.  The gem client would need to fetch the
list (maybe caching it locally) and get the gem from a random mirror or
the closest one geographically or whatever.  

Of course, this would mean that we wouldn't have accurate stats on
rubyforge.org any more, but, meh.

Yours,

Tom



From tobiasreif at pinkjuice.com  Tue Jan 29 05:14:17 2008
From: tobiasreif at pinkjuice.com (Tobi Reif)
Date: Tue, 29 Jan 2008 11:14:17 +0100
Subject: [Rubygems-developers] Failing gem commands;
	reinstalling	RubyGems
In-Reply-To: <e3af97510801281248n1e32288yf7a5a39ec427c2b5@mail.gmail.com>
References: <20080122115746.GA4616@linux.local>
	<e3af97510801280052y175938ddv857721bfec103b14@mail.gmail.com>
	<20080128132029.GB7971@linux.local>
	<e3af97510801280957u4c9a6affje4f472474c38a386@mail.gmail.com>
	<20080128191255.GA3988@linux.local>
	<e3af97510801281248n1e32288yf7a5a39ec427c2b5@mail.gmail.com>
Message-ID: <20080129101417.GA4184@linux.local>

On Mon 2008-01-28 Chad Woolley wrote:
[...]
> > I really would like to know what the recommended way is for
> > uninstalling / reinstalling RubyGems. Rich, Jim, Chad, Eric, what's
> > the "official" recommendation?
> 
> Yes, I'd like to hear the latest and greatest official uninstall
> procedure as well...

If there is no automated way, perhaps one could be added?

RMagick for example has uninstall.rb (I've never tried it) and a file
called InstalledFiles.

Tobi


From transfire at gmail.com  Wed Jan 30 12:43:48 2008
From: transfire at gmail.com (Trans)
Date: Wed, 30 Jan 2008 09:43:48 -0800 (PST)
Subject: [Rubygems-developers] WARNINGS
Message-ID: <66575480-4b72-450b-8bf9-0a4bd9d228b7@m34g2000hsf.googlegroups.com>

Why is Rubygems now issuing warnings when building, eg.

  WARNING:  no rubyforge_project specified
  WARNING:  RDoc will not be generated (has_rdoc == false)

I've never seen it do this before. How to I silence it? (Using ruby,
not command line).

Thanks,
T.

From luislavena at gmail.com  Wed Jan 30 12:48:06 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 30 Jan 2008 15:48:06 -0200
Subject: [Rubygems-developers] WARNINGS
In-Reply-To: <66575480-4b72-450b-8bf9-0a4bd9d228b7@m34g2000hsf.googlegroups.com>
References: <66575480-4b72-450b-8bf9-0a4bd9d228b7@m34g2000hsf.googlegroups.com>
Message-ID: <71166b3b0801300948j19fa1fbfjdb7d83274388a41e@mail.gmail.com>

On Jan 30, 2008 3:43 PM, Trans <transfire at gmail.com> wrote:
> Why is Rubygems now issuing warnings when building, eg.
>
>   WARNING:  no rubyforge_project specified
>   WARNING:  RDoc will not be generated (has_rdoc == false)
>
> I've never seen it do this before. How to I silence it? (Using ruby,
> not command line).
>

These warnings where introduced due bad gem specifications.

Setting the correct values in the gem specification will silent these warnings.

-- 
Luis Lavena
Multimedia systems
-
A common mistake that people make when trying to design
something completely foolproof is to underestimate
the ingenuity of complete fools.
Douglas Adams

From me at susanpotter.net  Wed Jan 30 13:32:52 2008
From: me at susanpotter.net (Susan Potter)
Date: Wed, 30 Jan 2008 12:32:52 -0600
Subject: [Rubygems-developers] RubyGems Dependency Type
Message-ID: <1076b0d90801301032hb5ba5d1qf93d5287172d3a21@mail.gmail.com>

Hi,

Back in November on this list there was a discussion skirting around a
feature I have been forced to work-around for my enterprise client.
My work-around is a separate Gem and only allows my solution to work
on our internal systems because it is installed everywhere RubyGems
is.  It uses Ruby's ability to extend/reopen RubyGem classes to append
properties/methods to it, specifically the Gem::Specification and
Gem::Dependency classes as well as the CLI functionality.

The thread I refer to was titled: What is right and wrong with
dependencies definitions?

The thread discussion was initiated by someone named Luis (I think),
but below I quote Eric Hodel and Chad Woolley to provide people a
refresher or background to the uninitiated (see below the ---- line).

While I think Chad Woolley's GemInstaller is useful in some
situations, the problem I am trying to solve is to describe the kind
of dependency a gem is for a project.  For example, it is not right to
say that rspec will only ever be a test dependency, since it would be
needed by any rspec extensions at runtime, not just at test time.  I
am also not personally bowled over with the idea of starting a server
to establish whether I need to install further dependencies.  The
reason for this is the separation of work roles that generally exist
in the server-side environment (system administrator,
deployer/developer, etc.).

I would rather these concerns were taken care of at gem install time.
I see GemInstaller's purpose more useful for one-off scripts or UI
based applications rather than server-side applications.  Server-side
software is primarily my domain, so please keep that in mind.

The following is the scheme I used at my client (but with a few
client-specific things thrown in that I will not discuss here as they
aren't relevant for solving *this* particular problem):
* each dependency when defined has a "type".  I have used the types
(with some modifications) that Maven allows developers to define.
Yes, yes, I know the Ruby world is infinitely better than the Java
world, but Maven (which happens to be in the Java world) has actually
found a plausible solution in my view.  Maven's dependency types are:
compile (default), provided (for JEE contained apps), test.  I have
personally used the following types for Ruby to make more sense for
our world: runtime (default), compile (for Ruby/C extension purposes),
test.
* on "gem install" (unless otherwise told) the runtime dependencies
are installed only.  I added a flag to the gem CLI (again through
extension).  Of course, this again only works in our internal, fully
controlled system.

Here is my proposal open to feedback since my background is primarily
in the server-side enterprise space (I am directly answering Eric
Hodel's questions from below):
* should developer dependencies by installed by default?
>> No
* what does the command-line option look like?
>> gem install -Dall or gem install -Dtest or gem install -Dcompile
>> test installs runtime AND test dependencies.  compile installs
runtime AND compile dependencies.  all installs all three: runtime,
compile and test.
* what happens on uninstall?
>> ???  The client implementation I have currently only uninstalls the
given gem(s), but this needs further thinking as I realize this is not
optimum.
* what should `gem check` do if all dependencies aren't installed?
>> in my proposed implementation I have found it useful to only check
runtime by default, but if -D is specified it will check those as per
the CLI option noted above.

To be honest, I think the thing that matters most is that RubyGems
provides a way to describe this metadata and then we can worry about
tools and facilities to wrap around this later.  If people do not want
to set the extra dependency type, that is fine, we default it to
runtime and Gem developers aren't any worse off than they are now.

If you want to do extra things based on RAILS_ENV or MERB_ENV or
another environmental setting you can do so with something like
GemInstaller or another RubyGems "extension".  In fact, I think simply
adding the metadata property of Gem::Dependency#type (ok, we use #kind
because of #type history) to RubyGems gives greater flexibility rather
than only providing one facility (e.g. GemInstaller, that you
essentially have to be married to).  We can even defer how people
handle installing using these dependency types to third party Gems
instead of involving RubyGems in that business.

I am willing and able (with about 3-5 hours a week to commit to open
source contributions across the board) to do most of the grunt work
and submit a patch for RubyGems project to do the capturing of the
metadata and change the gem CLI behavior depending on accepted
proposal.  Unfortunately I cannot just open source the client work I
have already done since it is considered proprietary.

Thanks for listening,
Susan
-----
Previous thread on this mailing list...
Thread: What is right and wrong with dependencies definitions?
Month: November 2007

====
From: drbrain at segment7.net (Eric Hodel)
Date: Fri, 9 Nov 2007 14:47:51 -0800

Things that need to be discussed are:
* should developer dependencies by installed by default?
* what does the command-line option look like?
* what happens on uninstall?
* what should `gem check` do if all dependencies aren't installed?

Before attempting a patch, I think at least these questions should be
answered.  There are probably more, since getting it to work well
requires changes and hooks in lots of places.

Its really easy to say "gems should have developer dependencies".
Its a lot of work to make it a reality.
====

====
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Fri, 9 Nov 2007 17:08:30 -0700

Regardless, there are valid points on both sides.  If building this
support into RubyGems really is too hard (and I believe you), then we
need alternate approaches.

Ok, pimp-mode on: I think that my GemInstaller tool
(http://geminstaller.rubyforge.org) is perfect for this - it centrally
manaages all your dependencies via an erb-parsed file (or files).
This allows you to explicitly manage your dependencies, and have them
behave differently based on whatever criteria you want - RAILS_ENV,
hostname, environment variables, whatever.  If you have a startup
check in your app that installs and loads the appropriate gems in
development or test mode (rake or test_helper), then you can only
include actual runtime dependencies in your gem, but ensure that
anyone who wants to build/test your app still has the gems they need.
====

====
From: drbrain at segment7.net (Eric Hodel)
Date: Fri, 9 Nov 2007 17:33:08 -0800

> I agree that this will probably be really hard to implement in
> rubygems, and I'll believe you that it's not worth it.  However,
> there's still valid reasons for both sides of the argument.

I'm not saying its not worth doing, but it's much, much easier said
than done.
====

-- 
mailto:me at susanpotter.net
www:http://susanpotter.net
blog:http://snakesgemscoffee.susanpotter.net
linkedin:http://linkedin.com/in/susanpotter

From transfire at gmail.com  Wed Jan 30 16:10:18 2008
From: transfire at gmail.com (Trans)
Date: Wed, 30 Jan 2008 13:10:18 -0800 (PST)
Subject: [Rubygems-developers] WARNINGS
In-Reply-To: <71166b3b0801300948j19fa1fbfjdb7d83274388a41e@mail.gmail.com>
References: <66575480-4b72-450b-8bf9-0a4bd9d228b7@m34g2000hsf.googlegroups.com>
	<71166b3b0801300948j19fa1fbfjdb7d83274388a41e@mail.gmail.com>
Message-ID: <1cc495f5-2c7e-4f36-bac1-90b117c843c9@j78g2000hsd.googlegroups.com>



On Jan 30, 12:48 pm, "Luis Lavena" <luislav... at gmail.com> wrote:
> On Jan 30, 2008 3:43 PM, Trans <transf... at gmail.com> wrote:
>
> > Why is Rubygems now issuing warnings when building, eg.
>
> >   WARNING:  no rubyforge_project specified
> >   WARNING:  RDoc will not be generated (has_rdoc == false)
>
> > I've never seen it do this before. How to I silence it? (Using ruby,
> > not command line).
>
> These warnings where introduced due bad gem specifications.
>
> Setting the correct values in the gem specification will silent these warnings.

That's not so. The has_rdoc is set to false for a reason. Plus one
might not have a rubyforge project.

T.

From luislavena at gmail.com  Wed Jan 30 16:15:39 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 30 Jan 2008 19:15:39 -0200
Subject: [Rubygems-developers] WARNINGS
In-Reply-To: <1cc495f5-2c7e-4f36-bac1-90b117c843c9@j78g2000hsd.googlegroups.com>
References: <66575480-4b72-450b-8bf9-0a4bd9d228b7@m34g2000hsf.googlegroups.com>
	<71166b3b0801300948j19fa1fbfjdb7d83274388a41e@mail.gmail.com>
	<1cc495f5-2c7e-4f36-bac1-90b117c843c9@j78g2000hsd.googlegroups.com>
Message-ID: <71166b3b0801301315y3ab050ex2bb79abb22caf027@mail.gmail.com>

On Jan 30, 2008 7:10 PM, Trans <transfire at gmail.com> wrote:
> On Jan 30, 12:48 pm, "Luis Lavena" <luislav... at gmail.com> wrote:
>
> > Setting the correct values in the gem specification will silent these warnings.
>
> That's not so. The has_rdoc is set to false for a reason. Plus one
> might not have a rubyforge project.
>

Then is a bug and you should report it to the tracker:

http://rubyforge.org/tracker/?group_id=126

Thank you for your time :)

-- 
Luis Lavena
Multimedia systems
-
A common mistake that people make when trying to design
something completely foolproof is to underestimate
the ingenuity of complete fools.
Douglas Adams

From thewoolleyman at gmail.com  Wed Jan 30 16:16:36 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Wed, 30 Jan 2008 14:16:36 -0700
Subject: [Rubygems-developers] RubyGems Dependency Type
In-Reply-To: <1076b0d90801301032hb5ba5d1qf93d5287172d3a21@mail.gmail.com>
References: <1076b0d90801301032hb5ba5d1qf93d5287172d3a21@mail.gmail.com>
Message-ID: <e3af97510801301316x46dc3f0pd34343731c293bc1@mail.gmail.com>

On Jan 30, 2008 11:32 AM, Susan Potter <me at susanpotter.net> wrote:
> I am willing and able (with about 3-5 hours a week to commit to open
> source contributions across the board) to do most of the grunt work
> and submit a patch for RubyGems project to do the capturing of the
> metadata and change the gem CLI behavior depending on accepted
> proposal.

Susan,

Strange you should bring this up again - just this week I was bringing
up this thread at work.  I know that Eric is in the camp of "install
everything everywhere", but the fact is that many people just are not
comfortable with that.  Especially people who are on call to support
important production servers.  And, the incident with soap4r gem
screwing up Rails simply by being installed proves that wierd,
unexpected, *impossible* things can happen, especially in the magical
world of Ruby.  You simply don't want any chance of that happening -
especially on production.  That's the spirit behind GemInstaller, and
it's ability to print out "rogue" gems.

I also totally agree that, while it is possible to do what you want
with ERB in a geminstaller.yml config file, that's a hack and
'marries' your infrastructure to GemInstaller.  I can't see why you'd
be opposed to that, but I digress ;)

Anyway, I think this sounds like a very reasonable proposal.  Couple
of comments/suggestions, though:

1. Preserving backward compatablity and default behavior for existing
gems/specs/dependencies is of utmost importance.  Functional sanity
check tests which automatically run against existing gems/specs from
rubyforge, and ensure that dependencies are still handled the same,
would be a good regression safety net.

2. What is the -D switch to install short for?  Naming is important -
maybe -T/--type? or -c/--category?  I kind of like category, not too
generic, not an overloaded term.

3. The area that will cause the most contention is probably the
default categories.  You suggested 'runtime', 'test', and 'compile'.
'runtime' and 'test' sound reasonable, but what is compile?  Are these
for platform-specific build dependencies?  What about build-time only
(non-runtime) dependencies for pure-ruby gems?  Maybe 'build' would be
a more generic choice than 'compile'.  Also, what about deploy- or
publish-time dependencies - such as hoe, rubyforge, webgen/webby, etc?
 Are these in the 'build' category as well?

4. What about dependencies that are required for more than one category?

5. What would the new gem .gemspec file format look like?
Specifically, what about dependencies in more than one category?  Do
you propose adding a new optional parameter to
Gem::Specification.add_dependency?  If this accepted a string or an
array, it would handle the multiple-category issue.

Again, looks good.  If you can make it work, it will be great.

Thanks,
-- Chad

From drbrain at segment7.net  Thu Jan 31 17:23:26 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 31 Jan 2008 14:23:26 -0800
Subject: [Rubygems-developers] WARNINGS
Message-ID: <87AE92A5-A3E4-4531-8A2A-46B0588171EA@segment7.net>

Oops, I only sent this to Trans.

On Jan 30, 2008, at 13:10 PM, Trans wrote:
> On Jan 30, 12:48 pm, "Luis Lavena" <luislav... at gmail.com> wrote:
>> On Jan 30, 2008 3:43 PM, Trans <transf... at gmail.com> wrote:
>>> Why is Rubygems now issuing warnings when building, eg.
>>
>>> WARNING:  no rubyforge_project specified
>>> WARNING:  RDoc will not be generated (has_rdoc == false)
>>
>>> I've never seen it do this before. How to I silence it? (Using ruby,
>>> not command line).
>>
>> These warnings where introduced due bad gem specifications.
>>
>> Setting the correct values in the gem specification will silent  
>> these warnings.
>
> That's not so. The has_rdoc is set to false for a reason.

Then you're a bad person.  Not generating RDoc/ri is hostile to your  
users.

> Plus one might not have a rubyforge project.

There's an "orphans" project on rubyforge for gems that are lacking a  
real project.  If you're building an internal gem, I think you can  
deal with it.

From drbrain at segment7.net  Thu Jan 31 18:50:18 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 31 Jan 2008 15:50:18 -0800
Subject: [Rubygems-developers] WARNINGS
In-Reply-To: <4b6f054f0801311456v5ab6299x85f36f8ef879dcb8@mail.gmail.com>
References: <87AE92A5-A3E4-4531-8A2A-46B0588171EA@segment7.net>
	<4b6f054f0801311456v5ab6299x85f36f8ef879dcb8@mail.gmail.com>
Message-ID: <7906F1ED-E1E4-43D7-B20E-2A78A6BEAB7F@segment7.net>

On Jan 31, 2008, at 14:56 PM, Trans wrote:
> On Jan 31, 2008 5:23 PM, Eric Hodel <drbrain at segment7.net> wrote:
>> On Jan 30, 2008, at 13:10 PM, Trans wrote:
>>> On Jan 30, 12:48 pm, "Luis Lavena" <luislav... at gmail.com> wrote:
>>>> On Jan 30, 2008 3:43 PM, Trans <transf... at gmail.com> wrote:
>>>>> Why is Rubygems now issuing warnings when building, eg.
>>>>
>>>>> WARNING:  no rubyforge_project specified
>>>>> WARNING:  RDoc will not be generated (has_rdoc == false)
>>>>
>>>>> I've never seen it do this before. How to I silence it? (Using  
>>>>> ruby,
>>>>> not command line).
>>>>
>>>> These warnings where introduced due bad gem specifications.
>>>>
>>>> Setting the correct values in the gem specification will silent
>>>> these warnings.
>>>
>>> That's not so. The has_rdoc is set to false for a reason.
>>
>> Then you're a bad person.
>
> I'm a bad person? Eric do your ever listen to the things you say?

Yes, and you take things too seriously.

> The reason I don't have rdocs generated for one of my projects is
> becuase it has special requirements for the docs.

Then you'll have to live with the warnings, or fix the special  
requirements.

>> Not generating RDoc/ri is hostile to your users.
>
> If this is what you think why is is even an option?

Backwards compatibility.

From discordantus at gmail.com  Thu Jan 31 20:33:37 2008
From: discordantus at gmail.com (Mark Hubbart)
Date: Thu, 31 Jan 2008 17:33:37 -0800
Subject: [Rubygems-developers] WARNINGS
In-Reply-To: <87AE92A5-A3E4-4531-8A2A-46B0588171EA@segment7.net>
References: <87AE92A5-A3E4-4531-8A2A-46B0588171EA@segment7.net>
Message-ID: <de63abca0801311733x418d766aq822ca57181e1c883@mail.gmail.com>

On Jan 31, 2008 2:23 PM, Eric Hodel <drbrain at segment7.net> wrote:
> Oops, I only sent this to Trans.
>
>
> On Jan 30, 2008, at 13:10 PM, Trans wrote:
> > On Jan 30, 12:48 pm, "Luis Lavena" <luislav... at gmail.com> wrote:
>
> >> On Jan 30, 2008 3:43 PM, Trans <transf... at gmail.com> wrote:
> >>> Why is Rubygems now issuing warnings when building, eg.
> >>
> >>> WARNING:  no rubyforge_project specified
> >>> WARNING:  RDoc will not be generated (has_rdoc == false)
> >>
> >>> I've never seen it do this before. How to I silence it? (Using ruby,
> >>> not command line).
> >>
> >> These warnings where introduced due bad gem specifications.
> >>
>
> >> Setting the correct values in the gem specification will silent
> >> these warnings.
> >
> > That's not so. The has_rdoc is set to false for a reason.
>
> Then you're a bad person.  Not generating RDoc/ri is hostile to your
> users.

harsh much?

Reasons you might not want to generate RDoc for a particular gem:

- Your interface is already documented. Say you write a lib that
speeds up the use of Ruby's built-in complex class. If there's nothing
but a speed improvement, why re-document the methods?
- There is no interface to document. Perhaps your library does magic
stuff in the background (say, logging performance stats), needing no
method calls whatsoever.
- Your entire interface is dynamic, and has no set method calls. Maybe
you prefer to explain it on a website.
- Some other reason, at the choice of the developer. Complain to them
if they're wrong.

Anyway, if it's always bad to not generate rdoc/ri, they should simply
take the flag out.

> > Plus one might not have a rubyforge project.
>
> There's an "orphans" project on rubyforge for gems that are lacking a
> real project.  If you're building an internal gem, I think you can
> deal with it.

The "real" project could be on sourceforge.org, or code.google.com.
But if a project name is that important, maybe rubygems should insert
the default one (the "orphans" project) whenever none is specified.

Anyway, it seems like these warnings should be given when building the
gem, not while installing it.

? Mark