From drbrain at segment7.net  Thu May  1 02:04:52 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Wed, 30 Apr 2008 23:04:52 -0700
Subject: [Rubygems-developers] Missing require from
	rubygems/remote_fetcher.rb "stringio"
In-Reply-To: <3454c9680804300125t2674d02dl40a531d65365e57d@mail.gmail.com>
References: <bd544b2d0804292013v318ea262q8d52428ebf60b76b@mail.gmail.com>
	<3454c9680804300051q6eb520dk9c07ff67e6d07cb9@mail.gmail.com>
	<3454c9680804300125t2674d02dl40a531d65365e57d@mail.gmail.com>
Message-ID: <C7BB0DFE-49B7-47E1-8FC6-2C7FB3AC8A32@segment7.net>

On Apr 30, 2008, at 01:25 AM, Vladimir Sizikov wrote:
> OK, some more info here.
>
> Ola Bini figured out why MRI works and JRuby fails in the first place.
> It turned out that yaml library in MRI requires StringIO, while JRuby
> didn't. This has been corrected so that JRuby is
> "side-effect-compatible" with MRI in that regard, and yaml now also
> require's StringIO.
>
> Still, adding explicit require 'stringio' to rubygems sources could be
> useful, making the code more robust and less dependent on particular
> implementation and side effects.

Ok.  Did somebody add the ticket to the RubyGems tracker?  Otherwise  
it might get forgotten by the time I have some RubyGems cycles come up.

From luislavena at gmail.com  Thu May  1 02:18:06 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Thu, 1 May 2008 03:18:06 -0300
Subject: [Rubygems-developers] Missing require from
	rubygems/remote_fetcher.rb "stringio"
In-Reply-To: <C7BB0DFE-49B7-47E1-8FC6-2C7FB3AC8A32@segment7.net>
References: <bd544b2d0804292013v318ea262q8d52428ebf60b76b@mail.gmail.com>
	<3454c9680804300051q6eb520dk9c07ff67e6d07cb9@mail.gmail.com>
	<3454c9680804300125t2674d02dl40a531d65365e57d@mail.gmail.com>
	<C7BB0DFE-49B7-47E1-8FC6-2C7FB3AC8A32@segment7.net>
Message-ID: <71166b3b0804302318v7fe79aeq67926ec519ba5e56@mail.gmail.com>

On Thu, May 1, 2008 at 3:04 AM, Eric Hodel <drbrain at segment7.net> wrote:
> >
>
>  Ok.  Did somebody add the ticket to the RubyGems tracker?  Otherwise it
> might get forgotten by the time I have some RubyGems cycles come up.
>

It seems so:

#19866: Rubygems fails with NameError: StringIO
http://rubyforge.org/tracker/index.php?func=detail&aid=19866&group_id=126&atid=575

I'm also adding the uninstaller stuff too, just in case ;-)

-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From jeremydeininger at gmail.com  Thu May  1 12:06:23 2008
From: jeremydeininger at gmail.com (Jeremy d)
Date: Thu, 1 May 2008 09:06:23 -0700
Subject: [Rubygems-developers] multiple specific versions of dependencies
Message-ID: <f3b220000805010906i1ebe77f3l16cc95c0344118cc@mail.gmail.com>

Hello guys,
I have an interesting use case for you, and I've been working to patch
RubyGems in an attempt to match this functionality.

User needs to install multiple specifically versioned gems in the
quickest way possible.  Example command line would be similar to this
"gem install mongrel-1.0 daemons-1.0.8 rails rake-0.7.1".

I've been working with the latest RubyGems to see if the patch would
still fit and it is relying on having more caching available in order
to hot-wire the dependency tree.  I've ported it most of the way, but
I have a failing test that specifies the installation of dependencies
out of order.  I've tracked it down to line 99 of
dependency_installer.rb, previously the gems_and_sources contained a
large list of specs, in which I could cherry-pick the right
dependency.   Now with the new caching, it only returns the latest
version of a dependency.  Is there a way I can still access a larger
list of dependencies from the cache and select the specific version
needed?

I'm attaching the patch for 1.0.1 so you can see it in action.  Also
included is the patch in progress for 1.1.1, with 1 failing test.  The
patch adds an entirely new gem command called quickinstall, to retain
original functionality as much as possible, while adding default
options for speedup.  To use it: gem quickinstall mongrel-1.0
daemons-1.0.8

Any advise on a better way to solve this use case for future versions
of RubyGems would be much appreciated!

-Jeremy
RightScale.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rubygems-quickinstall-patch.zip
Type: application/zip
Size: 11777 bytes
Desc: not available
URL: <http://rubyforge.org/pipermail/rubygems-developers/attachments/20080501/a959b367/attachment-0001.zip>

From scholz at scriptolutions.com  Sat May  3 11:23:36 2008
From: scholz at scriptolutions.com (Lothar Scholz)
Date: Sat, 3 May 2008 22:23:36 +0700
Subject: [Rubygems-developers] How do i get all available versions of a gem?
Message-ID: <1664397337.20080503222336@scriptolutions.com>

Hello,

I upgraded from 0.9 and when i now use "gem list --remote --details" it
seems that i don't get the complete list of available gem version.
Only something like

ruby-style (1.1.4)
    Supervised TCPServer, Yielding Listeners Easily

where previous versions had an output like:

ruby-style (1.1.4, 1.1.3, 1.1.2, 1.1.1)
    Supervised TCPServer, Yielding Listeners Easily

Is there any way to get this version numbers back? They are very
usefull in practice.

-- 
Best regards,
 Lothar Scholz              mailto:scholz at scriptolutions.com


From thewoolleyman at gmail.com  Sat May  3 13:07:23 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Sat, 3 May 2008 10:07:23 -0700
Subject: [Rubygems-developers] How do i get all available versions of a
	gem?
In-Reply-To: <1664397337.20080503222336@scriptolutions.com>
References: <1664397337.20080503222336@scriptolutions.com>
Message-ID: <e3af97510805031007hea55559r25d45791b84fedba@mail.gmail.com>

On Sat, May 3, 2008 at 8:23 AM, Lothar Scholz <scholz at scriptolutions.com> wrote:
>  I upgraded from 0.9 and when i now use "gem list --remote --details" it
>  seems that i don't get the complete list of available gem version.

Use --all. There were caching performance improvements in 1.1.1, only
the latest is retrieved by default now.

-- Chad

From ryand-ruby at zenspider.com  Wed May  7 00:10:57 2008
From: ryand-ruby at zenspider.com (Ryan Davis)
Date: Tue, 6 May 2008 21:10:57 -0700
Subject: [Rubygems-developers] developer dependencies
Message-ID: <B345C07B-361B-412D-A711-8367B9A15A65@zenspider.com>

I tasked John Barnette with getting developer dependencies implemented  
properly in rubygems a week or so ago... I'm happy to announce that we  
just committed his changes. Enjoy.


From luislavena at gmail.com  Wed May  7 01:08:52 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 7 May 2008 02:08:52 -0300
Subject: [Rubygems-developers] developer dependencies
In-Reply-To: <B345C07B-361B-412D-A711-8367B9A15A65@zenspider.com>
References: <B345C07B-361B-412D-A711-8367B9A15A65@zenspider.com>
Message-ID: <71166b3b0805062208r75cc84b8g91864ec99c8d3dd5@mail.gmail.com>

On Wed, May 7, 2008 at 1:10 AM, Ryan Davis <ryand-ruby at zenspider.com> wrote:
> I tasked John Barnette with getting developer dependencies implemented
> properly in rubygems a week or so ago... I'm happy to announce that we just
> committed his changes. Enjoy.
>

just two things:yay and wow!

Looking forward for the blog post describing how to abu... use this
new functionality :-)

-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From djberg96 at gmail.com  Sun May 11 13:18:37 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sun, 11 May 2008 11:18:37 -0600
Subject: [Rubygems-developers] The return of PLATFORM constants
Message-ID: <6037b70c0805111018t4356388fn6299d9875b84d43d@mail.gmail.com>

Hi,

I have a few C extensions where I would like to be able to provide a
way to specify a generic platform.

For example, the sys-cpu source is the same for both FreeBSD and OS X.
However, as things stand now I must create two separate gems - one for
each platform. Here's where a Gem::Platform::BSD constant would be
very handy. Or whatever you want to call it.

I'm in a similar situation with gems that aren't particular to any
version of a given platform or compiler. For example, the sys-cpu
source for Solaris is the same for any version of Solaris, be it 2.6
or 2.10. Here's where a Gem::Platform::SOLARIS would be very handy.

Is there any chance of adding these? If not, what are my options
outside of building a gem on every platform?

Thanks,

Dan

PS - I thought I brought this up before, but now I can't find the thread.

From thewoolleyman at gmail.com  Sun May 11 16:30:44 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Sun, 11 May 2008 13:30:44 -0700
Subject: [Rubygems-developers] Mirror problem with Selenium 1.0.7?
Message-ID: <e3af97510805111330r3716d98ev4f550c4060327f1d@mail.gmail.com>

The Selenium (NOT selenium) gem doesn't seem to have the 1.0.7 version
propogated to all mirrors for installation, but it is on the download
page (http://rubyforge.org/frs/?group_id=2789).  Where should this be
reported/addressed?

Thx,
-- Chad

From stephen.bannasch at deanbrook.org  Mon May 12 13:04:54 2008
From: stephen.bannasch at deanbrook.org (Stephen Bannasch)
Date: Mon, 12 May 2008 13:04:54 -0400
Subject: [Rubygems-developers] dependent gems not installed when using a
	--source gem repository
Message-ID: <p06240817c44e274b30bc@[63.138.152.170]>

I've got an alpha gem I am working on called make_models and I'm hosting it now on my own gem repository. Both main and arrayfields are gem dependencies of make_models.  If I download it and install from a file the dependencies are resolved.

I got this report from someone else who tried to install it from my gem repository without having the prerequisite gems:

>sudo gem install make_models --source=http://rails.dev.concord.org/gems/
>
>I get this error:
>
>(9:27:59) [tikal:/web/rails] pburney% sudo gem install make_models --source=http://rails.dev.concord.org/gems/
>Password:
>Bulk updating Gem source index for: <http://rails.dev.concord.org/gems/>http://rails.dev.concord.org/gems/
>ERROR:  Error installing make_models:
>       make_models requires main (>= 2.8.0)

Has anybody else seen this problem?

It could be something wrong with my gem since it's one of the first two gems I've made.

From scholz at scriptolutions.com  Tue May 13 10:44:58 2008
From: scholz at scriptolutions.com (Lothar Scholz)
Date: Tue, 13 May 2008 21:44:58 +0700
Subject: [Rubygems-developers] Is there any update to the documentation
Message-ID: <1679109981.20080513214458@scriptolutions.com>

Hello,

Seems that for a long time nobody touched http://rubygems.org/

As a third party tool vendor i have to say it's hard to keep up with
the development when even the reference is not updated from time to
time. Or did i miss some new location?



-- 
Best regards,
 Lothar Scholz              mailto:scholz at scriptolutions.com


From scholz at scriptolutions.com  Wed May 14 05:05:33 2008
From: scholz at scriptolutions.com (Lothar Scholz)
Date: Wed, 14 May 2008 16:05:33 +0700
Subject: [Rubygems-developers] Why is ruby gems always updating when using
	"--source" ?
Message-ID: <1118656429.20080514160533@scriptolutions.com>

Hello,

when i did something pretty easy like
gem specification ackbar --remote --verbose --source "http://gems.rubyforge.org"

Then i get an output:

Missing metadata for 3153 gems
Bulk updating Gem source index for: http://gems.rubyforge.org/
connection reset after 2 requests, retrying

and have to wait a long long time, but without --source it does it
immediately (well still slow but at least without any updates).

-- 
Best regards,
 Lothar Scholz              mailto:scholz at scriptolutions.com


From luislavena at gmail.com  Wed May 14 07:29:36 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 14 May 2008 08:29:36 -0300
Subject: [Rubygems-developers] Why is ruby gems always updating when
	using "--source" ?
In-Reply-To: <1118656429.20080514160533@scriptolutions.com>
References: <1118656429.20080514160533@scriptolutions.com>
Message-ID: <71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>

On Wed, May 14, 2008 at 6:05 AM, Lothar Scholz
<scholz at scriptolutions.com> wrote:
> Hello,
>
>  when i did something pretty easy like
>  gem specification ackbar --remote --verbose --source "http://gems.rubyforge.org"
>

gems.rubyforge.org is the default, first gem repository rubygems will
use for the commands you indicate.


>  and have to wait a long long time, but without --source it does it
>  immediately (well still slow but at least without any updates).
>

When you explicitly indicates it, it skip the current local cache it
have for it and try downloading again.

HTH,
-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From scholz at scriptolutions.com  Wed May 14 08:30:45 2008
From: scholz at scriptolutions.com (Lothar Scholz)
Date: Wed, 14 May 2008 19:30:45 +0700
Subject: [Rubygems-developers] Why is ruby gems always updating when
	using "--source" ?
In-Reply-To: <71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>
References: <1118656429.20080514160533@scriptolutions.com>
	<71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>
Message-ID: <1893598493.20080514193045@scriptolutions.com>

Hello Luis,

Wednesday, May 14, 2008, 6:29:36 PM, you wrote:

LL> On Wed, May 14, 2008 at 6:05 AM, Lothar Scholz
LL> <scholz at scriptolutions.com> wrote:
>> Hello,
>>
>>  when i did something pretty easy like
>>  gem specification ackbar --remote --verbose --source "http://gems.rubyforge.org"
>>

LL> gems.rubyforge.org is the default, first gem repository rubygems will
LL> use for the commands you indicate.

I know but it shouldn't matter.


>>  and have to wait a long long time, but without --source it does it
>>  immediately (well still slow but at least without any updates).
>>

LL> When you explicitly indicates it, it skip the current local cache it
LL> have for it and try downloading again.

First i do not understand why the local cache does not not have an
indicator for the repository url (it should find out that source is
just the same). I do not even understand what the purpose of this
cache is when an explicit "--remote" is added.

Second why does it want to update any kind of data when i just want to
retrieve a remote spec.

Sorry i think something is terrible broken here.

In terms of performance rubygems sucks huge at the moment.
With more then 3000 gems and 13000 versions we need a lot of
improvements.


-- 
Best regards,
 Lothar Scholz                mailto:scholz at scriptolutions.com


From luislavena at gmail.com  Wed May 14 08:40:46 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 14 May 2008 09:40:46 -0300
Subject: [Rubygems-developers] Why is ruby gems always updating when
	using "--source" ?
In-Reply-To: <1893598493.20080514193045@scriptolutions.com>
References: <1118656429.20080514160533@scriptolutions.com>
	<71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>
	<1893598493.20080514193045@scriptolutions.com>
Message-ID: <71166b3b0805140540k352c88eaob83dbe33b56d3bf0@mail.gmail.com>

On Wed, May 14, 2008 at 9:30 AM, Lothar Scholz
<scholz at scriptolutions.com> wrote:
> Hello Luis,
>
>  Wednesday, May 14, 2008, 6:29:36 PM, you wrote:
>
>  LL> On Wed, May 14, 2008 at 6:05 AM, Lothar Scholz
>
> LL> <scholz at scriptolutions.com> wrote:
>  >> Hello,
>  >>
>  >>  when i did something pretty easy like
>  >>  gem specification ackbar --remote --verbose --source "http://gems.rubyforge.org"
>  >>
>
>  LL> gems.rubyforge.org is the default, first gem repository rubygems will
>  LL> use for the commands you indicate.
>
>  I know but it shouldn't matter.
>

Don't get your point.

>
>
>  >>  and have to wait a long long time, but without --source it does it
>  >>  immediately (well still slow but at least without any updates).
>  >>
>
>  LL> When you explicitly indicates it, it skip the current local cache it
>  LL> have for it and try downloading again.
>
>  First i do not understand why the local cache does not not have an
>  indicator for the repository url (it should find out that source is
>  just the same). I do not even understand what the purpose of this
>  cache is when an explicit "--remote" is added.
>
>  Second why does it want to update any kind of data when i just want to
>  retrieve a remote spec.
>
>  Sorry i think something is terrible broken here.
>
>  In terms of performance rubygems sucks huge at the moment.
>  With more then 3000 gems and 13000 versions we need a lot of
>  improvements.
>

Mmm, I hardly understand your point, issued the same command here and
got ackbar spec results in 2 seconds... even with the bulk update...
and on Windows.

So maybe is kind of dumb to actually understand your point here, or
maybe you didn't explain it better.

-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From scholz at scriptolutions.com  Wed May 14 09:31:05 2008
From: scholz at scriptolutions.com (Lothar Scholz)
Date: Wed, 14 May 2008 20:31:05 +0700
Subject: [Rubygems-developers] Why is ruby gems always updating when
	using "--source" ?
In-Reply-To: <71166b3b0805140540k352c88eaob83dbe33b56d3bf0@mail.gmail.com>
References: <1118656429.20080514160533@scriptolutions.com>
	<71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>
	<1893598493.20080514193045@scriptolutions.com>
	<71166b3b0805140540k352c88eaob83dbe33b56d3bf0@mail.gmail.com>
Message-ID: <1544830969.20080514203105@scriptolutions.com>

Hello Luis,

>>  I know but it shouldn't matter.
>>

LL> Don't get your point.

Well if http://gems.rubyforge.org is the default for --source the
operation should be exactly the same if i use or leave the option.

>>
>>  In terms of performance rubygems sucks huge at the moment.
>>  With more then 3000 gems and 13000 versions we need a lot of
>>  improvements.
>>

LL> Mmm, I hardly understand your point, issued the same command here and
LL> got ackbar spec results in 2 seconds... even with the bulk update...
LL> and on Windows.

Okay how much data was transfered?

I'm not interested in seconds i'm sure it is the data transfer size
that kills the performance it. How much data is transfered?

At the moment on my system everything seems to be impossible to
install anything.

Just because every simple command wants to update my fucking local cache,
even when i don't see a technical reason why it needs to be updates.
For example installing a gem where i give the exact version number.
This should just do a http request download a gem file and unpack this
in the correct places. Where does it need any information about other
gems?

So why is it trying to do a 10min bulk update? Mostly the connection
is getting a time out anyway, so even waiting is impossible.

And it tries to do this on every second or third operation?
Even if i just did a remote listing (with --all --details)
just before this operation (which should have send enough information
for the rubygems client).

Well when i think i really don't understand why we should have a local
cache for ruby gems at. I don't see any real value. It just adds
overload and complexity. So can you tell me what the purpose of the
local cache is and keeping any information locally other the local
downloaded gem archive files?


-- 
Best regards,
 Lothar Scholz                mailto:scholz at scriptolutions.com


From hgs at dmu.ac.uk  Wed May 14 10:07:32 2008
From: hgs at dmu.ac.uk (Hugh Sasse)
Date: Wed, 14 May 2008 15:07:32 +0100 (WEST)
Subject: [Rubygems-developers] Why is ruby gems always updating when
 using "--source" ?
In-Reply-To: <1544830969.20080514203105@scriptolutions.com>
References: <1118656429.20080514160533@scriptolutions.com>
	<71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>
	<1893598493.20080514193045@scriptolutions.com>
	<71166b3b0805140540k352c88eaob83dbe33b56d3bf0@mail.gmail.com>
	<1544830969.20080514203105@scriptolutions.com>
Message-ID: <Pine.GSO.4.64.0805141450410.11194@brains.eng.cse.dmu.ac.uk>

On Wed, 14 May 2008, Lothar Scholz wrote:

> LL> Mmm, I hardly understand your point, issued the same command here and
> LL> got ackbar spec results in 2 seconds... even with the bulk update...
> LL> and on Windows.
> 
> Okay how much data was transfered?
> 
> I'm not interested in seconds i'm sure it is the data transfer size
> that kills the performance it. How much data is transfered?

There has been some improvement in this sort of thing over time.
The indices weren't originally compressed, if I recall correctly.
Now they are.  I submitted patches to Ruby to get HTTP to use compression
when the server supports it.  I think there has been a certain amount
of work done on using the differential features of HTTP to allow transmission
of fragments that changed rather than the whole thing.  I don't know
how far that got. And Rubyforge does automatically make use of
mirror sites as well, to try to spread the load  So things are improving.
> 
> At the moment on my system everything seems to be impossible to
> install anything.
> 
> Just because every simple command wants to update my fucking local cache,
> even when i don't see a technical reason why it needs to be updates.
> For example installing a gem where i give the exact version number.

I've not looked at the source recently, and can't remember the details
from when I did, but clearly when no version number is given, the index
must be obtained, to check what the newest is.  I suspect that exact
version numbers make use of the same code (because of programmer effort).

> This should just do a http request download a gem file and unpack this
> in the correct places. Where does it need any information about other
> gems?

If the gem is on Rubyforge you can get it "by hand", and then install it.
So you have a workaround for now.
        [...]
There have been a lot of pressing issues with rubygems of late
(different platform issues, what to do about case (in)?sensitive
names, etc) so I just think "other wheels are getting the grease"
at the moment.  But I don't think the points you raise will go away, 
so they will be addressed in due course.

        Hugh

From drbrain at segment7.net  Thu May 15 03:25:14 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 15 May 2008 00:25:14 -0700
Subject: [Rubygems-developers] Is there any update to the documentation
In-Reply-To: <1679109981.20080513214458@scriptolutions.com>
References: <1679109981.20080513214458@scriptolutions.com>
Message-ID: <BAA50322-9D9E-4892-B3E3-1DCA7236074C@segment7.net>

On May 13, 2008, at 07:44 AM, Lothar Scholz wrote:

> Hello,
>
> Seems that for a long time nobody touched http://rubygems.org/
>
> As a third party tool vendor i have to say it's hard to keep up with
> the development when even the reference is not updated from time to
> time. Or did i miss some new location?

`gem server`

From drbrain at segment7.net  Thu May 15 03:34:20 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 15 May 2008 00:34:20 -0700
Subject: [Rubygems-developers] Why is ruby gems always updating when
	using "--source" ?
In-Reply-To: <1893598493.20080514193045@scriptolutions.com>
References: <1118656429.20080514160533@scriptolutions.com>
	<71166b3b0805140429o1dd82401o3c89b4a1be3939fc@mail.gmail.com>
	<1893598493.20080514193045@scriptolutions.com>
Message-ID: <93FCB0D8-783A-4929-9F6D-366FC0AC4048@segment7.net>

On May 14, 2008, at 05:30 AM, Lothar Scholz wrote:
> Wednesday, May 14, 2008, 6:29:36 PM, you wrote:
> LL> On Wed, May 14, 2008 at 6:05 AM, Lothar Scholz
> LL> <scholz at scriptolutions.com> wrote:
>>> and have to wait a long long time, but without --source it does it
>>> immediately (well still slow but at least without any updates).
>
> LL> When you explicitly indicates it, it skip the current local  
> cache it
> LL> have for it and try downloading again.
>
> First i do not understand why the local cache does not not have an
> indicator for the repository url (it should find out that source is
> just the same). I do not even understand what the purpose of this
> cache is when an explicit "--remote" is added.
>
> Second why does it want to update any kind of data when i just want to
> retrieve a remote spec.
>
> Sorry i think something is terrible broken here.
>
> In terms of performance rubygems sucks huge at the moment.
> With more then 3000 gems and 13000 versions we need a lot of
> improvements.

The current metadata update system is a victim of RubyGems success.   
Please see trunk for work in progress towards an improved metadata  
update system.  It is not yet wired in, however.

From drbrain at segment7.net  Thu May 15 03:37:12 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 15 May 2008 00:37:12 -0700
Subject: [Rubygems-developers] dependent gems not installed when using a
	--source gem repository
In-Reply-To: <p06240817c44e274b30bc@[63.138.152.170]>
References: <p06240817c44e274b30bc@[63.138.152.170]>
Message-ID: <0F91F526-065D-4DB4-A1D1-963F843B110E@segment7.net>

On May 12, 2008, at 10:04 AM, Stephen Bannasch wrote:
> I've got an alpha gem I am working on called make_models and I'm  
> hosting it now on my own gem repository. Both main and arrayfields  
> are gem dependencies of make_models.  If I download it and install  
> from a file the dependencies are resolved.
>
> I got this report from someone else who tried to install it from my  
> gem repository without having the prerequisite gems:
>
>> sudo gem install make_models --source=http://rails.dev.concord.org/gems/
>>
>> I get this error:
>>
>> (9:27:59) [tikal:/web/rails] pburney% sudo gem install make_models  
>> --source=http://rails.dev.concord.org/gems/
>> Password:
>> Bulk updating Gem source index for: <http://rails.dev.concord.org/gems/ 
>> >http://rails.dev.concord.org/gems/
>> ERROR:  Error installing make_models:
>>      make_models requires main (>= 2.8.0)
>
> Has anybody else seen this problem?

There is no main gem in that gem repository.  Either add main to that  
repository, or include both that repository and your repository as  
sources.

From drbrain at segment7.net  Thu May 15 03:45:43 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 15 May 2008 00:45:43 -0700
Subject: [Rubygems-developers] The return of PLATFORM constants
In-Reply-To: <6037b70c0805111018t4356388fn6299d9875b84d43d@mail.gmail.com>
References: <6037b70c0805111018t4356388fn6299d9875b84d43d@mail.gmail.com>
Message-ID: <C5C78EB1-7705-41B5-9EEB-855205A01DF8@segment7.net>

On May 11, 2008, at 10:18 AM, Daniel Berger wrote:

> I have a few C extensions where I would like to be able to provide a
> way to specify a generic platform.
>
> For example, the sys-cpu source is the same for both FreeBSD and OS X.
> However, as things stand now I must create two separate gems - one for
> each platform. Here's where a Gem::Platform::BSD constant would be
> very handy. Or whatever you want to call it.

The same binary works on FreeBSD and OS X?  Or is it a source-only gem?

If it is a source-only gem, is there a way to handle this with  
extconf.rb instead?  While OS X has FreeBSD roots, the two don't have  
completely compatible internals, so its hard to describe this in  
RubyGems.

> I'm in a similar situation with gems that aren't particular to any
> version of a given platform or compiler. For example, the sys-cpu
> source for Solaris is the same for any version of Solaris, be it 2.6
> or 2.10. Here's where a Gem::Platform::SOLARIS would be very handy.
>
> Is there any chance of adding these? If not, what are my options
> outside of building a gem on every platform?

You should be able to hand-specify a platform with something like  
Gem::Platform.new 'x86-freebsd', but there seems to be a bug.

Currently Gem::Platform is rather strict, as it is easier to make it  
more permissive in future releases than less permissive in future  
releases.

From drbrain at segment7.net  Thu May 15 03:47:43 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 15 May 2008 00:47:43 -0700
Subject: [Rubygems-developers] multiple specific versions of dependencies
In-Reply-To: <f3b220000805010906i1ebe77f3l16cc95c0344118cc@mail.gmail.com>
References: <f3b220000805010906i1ebe77f3l16cc95c0344118cc@mail.gmail.com>
Message-ID: <F17AD93F-68A4-4462-A4AB-31243C843FDF@segment7.net>

On May 1, 2008, at 09:06 AM, Jeremy d wrote:

> Hello guys,
> I have an interesting use case for you, and I've been working to patch
> RubyGems in an attempt to match this functionality.
>
> User needs to install multiple specifically versioned gems in the
> quickest way possible.  Example command line would be similar to this
> "gem install mongrel-1.0 daemons-1.0.8 rails rake-0.7.1".
>
> I've been working with the latest RubyGems to see if the patch would
> still fit and it is relying on having more caching available in order
> to hot-wire the dependency tree.  I've ported it most of the way, but
> I have a failing test that specifies the installation of dependencies
> out of order.  I've tracked it down to line 99 of
> dependency_installer.rb, previously the gems_and_sources contained a
> large list of specs, in which I could cherry-pick the right
> dependency.   Now with the new caching, it only returns the latest
> version of a dependency.  Is there a way I can still access a larger
> list of dependencies from the cache and select the specific version
> needed?

Gem::SourceInfoCache.search takes an "all" paramater than is false by  
default.  When you detect a version, you should make it switch to true.

> I'm attaching the patch for 1.0.1 so you can see it in action.  Also
> included is the patch in progress for 1.1.1, with 1 failing test.  The
> patch adds an entirely new gem command called quickinstall, to retain
> original functionality as much as possible, while adding default
> options for speedup.  To use it: gem quickinstall mongrel-1.0
> daemons-1.0.8

I'll take a look at it over the weekend.

From drbrain at segment7.net  Thu May 15 09:33:39 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Thu, 15 May 2008 06:33:39 -0700
Subject: [Rubygems-developers] env shebang always for JRuby
In-Reply-To: <4815E1C9.7060403@sun.com>
References: <4815E1C9.7060403@sun.com>
Message-ID: <F17A1E24-9568-4ED9-8C52-F477328120A9@segment7.net>

On Apr 28, 2008, at 07:40 AM, Charles Oliver Nutter wrote:

> JRuby ships with its "jruby" command as a shell script or batch file  
> under Windows. This means that the default behavior for JRuby  
> installing RubyGems is to always use an "env" shebang. However there  
> are two problems with this right now:
>
> - we have to manually patch our copy of RubyGems every time we  
> upgrade, to make sure it env-shebangs by default
> - RubyGems hardcodes the env shebang string to "ruby" rather than  
> using the name of the current runtime's command (this may have been  
> fixed?)
>
> At any rate, I'm looking for suggestions or a RubyGems fix that will  
> save us having to manually patch this all the time. JRuby is not  
> going to have a native startup command any time soon, so the env  
> shebang is always going to be necessary.
>
> Thoughts?

We discussed this on IM privately, and I forgot to copy it here.

I'm planning on adding a file (or two) that RubyGems will look for and  
run to provide implementation-specific and distribution-specific  
defaults.  This would allow jruby or apple or whomever to provide  
consistent behavior.

From Daniel.Berger at qwest.com  Thu May 15 10:23:24 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Thu, 15 May 2008 09:23:24 -0500
Subject: [Rubygems-developers] The return of PLATFORM constants
In-Reply-To: <C5C78EB1-7705-41B5-9EEB-855205A01DF8@segment7.net>
References: <6037b70c0805111018t4356388fn6299d9875b84d43d@mail.gmail.com>
	<C5C78EB1-7705-41B5-9EEB-855205A01DF8@segment7.net>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D326D@ITOMAE2KM01.AD.QINTRA.COM>

 

> -----Original Message-----
> From: rubygems-developers-bounces at rubyforge.org 
> [mailto:rubygems-developers-bounces at rubyforge.org] On Behalf 
> Of Eric Hodel
> Sent: Thursday, May 15, 2008 1:46 AM
> To: rubygems-developers at rubyforge.org
> Subject: Re: [Rubygems-developers] The return of PLATFORM constants
> 
> On May 11, 2008, at 10:18 AM, Daniel Berger wrote:
> 
> > I have a few C extensions where I would like to be able to 
> provide a 
> > way to specify a generic platform.
> >
> > For example, the sys-cpu source is the same for both 
> FreeBSD and OS X.
> > However, as things stand now I must create two separate 
> gems - one for 
> > each platform. Here's where a Gem::Platform::BSD constant would be 
> > very handy. Or whatever you want to call it.
> 
> The same binary works on FreeBSD and OS X?  Or is it a 
> source-only gem?
> 
> If it is a source-only gem, is there a way to handle this 
> with extconf.rb instead?  While OS X has FreeBSD roots, the 
> two don't have completely compatible internals, so its hard 
> to describe this in RubyGems.

That's what I do. But what do I set the Gem::PLATFORM to? I can't just
use 'CURRENT', because FreeBSD won't pickup an OS X gem. I can't use
'RUBY' because there are other platforms, too.

The only solution I've got so far is to either create separate gems, or
just lump _all_ of the source files together and let extconf.rb sort it
out. The latter approach works, but the downside to that approach is
that I can't be specific about the rdoc files, so I end up a bunch of
extra .c files (from other platforms) that get picked up during rdoc
generation.
 
> > I'm in a similar situation with gems that aren't particular to any 
> > version of a given platform or compiler. For example, the sys-cpu 
> > source for Solaris is the same for any version of Solaris, 
> be it 2.6 
> > or 2.10. Here's where a Gem::Platform::SOLARIS would be very handy.
> >
> > Is there any chance of adding these? If not, what are my options 
> > outside of building a gem on every platform?
> 
> You should be able to hand-specify a platform with something 
> like Gem::Platform.new 'x86-freebsd', but there seems to be a bug.

I'll see what I can see. :)

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 martin.krauskopf at gmail.com  Fri May 16 03:31:58 2008
From: martin.krauskopf at gmail.com (Martin Krauskopf)
Date: Fri, 16 May 2008 09:31:58 +0200
Subject: [Rubygems-developers] Is there any update to the documentation
In-Reply-To: <BAA50322-9D9E-4892-B3E3-1DCA7236074C@segment7.net>
References: <1679109981.20080513214458@scriptolutions.com>
	<BAA50322-9D9E-4892-B3E3-1DCA7236074C@segment7.net>
Message-ID: <g0jd9e$64a$1@ger.gmane.org>

Eric Hodel wrote:
> On May 13, 2008, at 07:44 AM, Lothar Scholz wrote:
> 
>> Hello,
>>
>> Seems that for a long time nobody touched http://rubygems.org/
>>
>> As a third party tool vendor i have to say it's hard to keep up with
>> the development when even the reference is not updated from time to
>> time. Or did i miss some new location?
> 
> `gem server`

Hi Eric,

likely you misread Lothar's question. He did not ask how to get 
documentation for installed gems.

To Lothar: I have the similar problem. RubyGems sources and ChangeLog 
seems to be the best docs.

     m.


From stephen.bannasch at deanbrook.org  Fri May 16 13:00:10 2008
From: stephen.bannasch at deanbrook.org (Stephen Bannasch)
Date: Fri, 16 May 2008 13:00:10 -0400
Subject: [Rubygems-developers] dependent gems not installed when using a
 --source gem repository
In-Reply-To: <0F91F526-065D-4DB4-A1D1-963F843B110E@segment7.net>
References: <p06240817c44e274b30bc@[63.138.152.170]>
	<0F91F526-065D-4DB4-A1D1-963F843B110E@segment7.net>
Message-ID: <p06240800c453655b723f@[192.168.1.114]>

At 12:37 AM -0700 5/15/08, Eric Hodel wrote:
>On May 12, 2008, at 10:04 AM, Stephen Bannasch wrote:
>>I've got an alpha gem I am working on called make_models and I'm hosting it now on my own gem repository. Both main and arrayfields are gem dependencies of make_models.  If I download it and install from a file the dependencies are resolved.
>>
>>I got this report from someone else who tried to install it from my gem repository without having the prerequisite gems:
>>
>>>sudo gem install make_models --source=http://rails.dev.concord.org/gems/
>>>
>>>I get this error:
>>>
>>>(9:27:59) [tikal:/web/rails] pburney% sudo gem install make_models --source=http://rails.dev.concord.org/gems/
>>>Password:
>>>Bulk updating Gem source index for: <http://rails.dev.concord.org/gems/>http://rails.dev.concord.org/gems/
>>>ERROR:  Error installing make_models:
>>>     make_models requires main (>= 2.8.0)
>>
>>Has anybody else seen this problem?
>
>There is no main gem in that gem repository.  Either add main to that repository, or include both that repository and your repository as sources.

I had assumed that the installed source repositories (default and any installed with gem sources --add) would be included in the search strategy used for dependency resolution.

It wasn't immediately obvious how to to specify two or more gem repositories to use for resolving install dependencies without making them permanent.

I looked at the code in local_remote_options.rb:

  Gem::LocalRemoteOptions#add_source_option

  # Add the --source option
  def add_source_option
    accept_uri_http

    add_option(:"Local/Remote", '--source URL', URI::HTTP,
               'Use URL as the remote source for gems') do |value, options|
      if options[:added_source] then
        Gem.sources << value
      else
        options[:added_source] = true
        Gem.sources.replace [value]
      end
    end
  end

OK ... now I understand ...

Here's a doc diff that adds a bit more explanation about how it currently works:

$ svn diff lib/rubygems/local_remote_options.rb
Index: lib/rubygems/local_remote_options.rb
===================================================================
--- lib/rubygems/local_remote_options.rb	(revision 1715)
+++ lib/rubygems/local_remote_options.rb	(working copy)
@@ -73,7 +73,8 @@
     accept_uri_http
 
     add_option(:"Local/Remote", '--source URL', URI::HTTP,
-               'Use URL as the remote source for gems') do |value, options|
+               'Use URL as the remote source for gems instead of installed sources'
+               'this option may be used multiple times to specify additional sources') do |value, options|
       if options[:added_source] then
         Gem.sources << value
       else

----------------------------------------------------------------------------------

And for anybody finding this later with Google here's a specific answer:

I've already uninstalled my gem make_models. The gem arrayfields is a dependency for make_models but is not in my repository. I'm uninstalling it for this test.

  $ sudo gem uninstall arrayfields

Now specifying both the default rubygems repository and my own repository (where make_models is hosted) installs the make_models gem from my custom repository and arrayfields from the rubygems repository.

  $ sudo gem install make_models --source=http://gems.rubyforge.org/ --source=http://rails.dev.concord.org/gems/

  Updating metadata for 2 gems from http://rails.dev.concord.org/gems/
  ..
  complete
  Successfully installed arrayfields-4.5.0
  Successfully installed make_models-0.0.1.13254
  2 gems installed
  Installing ri documentation for make_models-0.0.1.13254...
  Installing RDoc documentation for make_models-0.0.1.13254...

From tim at kingtony.com.tw  Fri May 16 21:14:18 2008
From: tim at kingtony.com.tw (=?BIG5?B?w7mnyqSk?=)
Date: Sat, 17 May 2008 09:14:18 +0800
Subject: [Rubygems-developers] rubygems on UBUNTU server edition 8.04 is
	some troube
Message-ID: <83357d8b0805161814y4729c3b4tf8a6ebbbe5d2e5cc@mail.gmail.com>

Dear all:
when I did as follows:
timloo at ktshop:/usr/bin$ sudo gem1.8 install rails
*It will holding on too long. *

top - 09:08:19 up 4 days, 15:15,  4 users,  load average: 7.59, 7.87, 6.19
Tasks:  64 total,   1 running,  60 sleeping,   3 stopped,   0 zombie
Cpu(s):  0.1%us,  5.8%sy,  0.0%ni,  0.0%id, 93.7%wa,  0.1%hi,  0.3%si,
0.0%st
Mem:    109392k total,   107472k used,     1920k free,       76k buffers
Swap:   321260k total,   321260k used,        0k free,      924k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
20607 root      20   0 26584  21m   72 D  2.5 19.7   0:23.11 gem1.8

*and at the time, when pinging the server , It has failed .*

*But When I gemed  install rake, It is OK. *
timloo at ktshop:/usr/bin$ sudo gem1.8 install rake
[sudo] password for timloo:
Updating metadata for 13 gems from http://gems.rubyforge.org/
.............
complete
Successfully installed rake-0.8.1
1 gem installed
Installing ri documentation for rake-0.8.1...
Installing RDoc documentation for rake-0.8.1...

Any Help??

Thanks

Tim





-- 
???
???????????
???
??->(04)23353567-643
???->http://www.wretch.cc/blog/timloo
(????????)
MSN(GAIM)->timloo_7 at hotmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/rubygems-developers/attachments/20080517/c7c603c6/attachment.html>

From drbrain at segment7.net  Sat May 17 22:05:46 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Sat, 17 May 2008 19:05:46 -0700
Subject: [Rubygems-developers] Is there any update to the documentation
In-Reply-To: <g0jd9e$64a$1@ger.gmane.org>
References: <1679109981.20080513214458@scriptolutions.com>
	<BAA50322-9D9E-4892-B3E3-1DCA7236074C@segment7.net>
	<g0jd9e$64a$1@ger.gmane.org>
Message-ID: <DB9B7CCB-4331-4174-ABF8-E654412FBE5B@segment7.net>

On May 16, 2008, at 00:31 AM, Martin Krauskopf wrote:
> Eric Hodel wrote:
>> On May 13, 2008, at 07:44 AM, Lothar Scholz wrote:
>>> Hello,
>>>
>>> Seems that for a long time nobody touched http://rubygems.org/
>>>
>>> As a third party tool vendor i have to say it's hard to keep up with
>>> the development when even the reference is not updated from time to
>>> time. Or did i miss some new location?
>> `gem server`
>
> Hi Eric,
>
> likely you misread Lothar's question.

I did not.

> He did not ask how to get documentation for installed gems.

http://localhost:8808/#rubygems

From clifford.heath at gmail.com  Wed May 21 20:55:23 2008
From: clifford.heath at gmail.com (Clifford Heath)
Date: Thu, 22 May 2008 10:55:23 +1000
Subject: [Rubygems-developers] Binary gems or extension source - which and
	how?
Message-ID: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>

I've used newgem to build a gem that's a simple wrapper for the wbxml  
library,
and I'm not sure how to publish it. My Manifest includes the ext  
files, including
extconf.rb, but since newgem & hoe build the gemspec file on the fly,  
I haven't
been able to work out how to arrange for the extension to be built on  
install.

Can someone give me some pointers please? The reference is too detailed,
I'm looking for a recipe if possible.

Also, what support does gems provide for the naming procedure for binary
files? If I want to publish platform-specific binary gems in addition  
to the source
version, what's the procedure?

Clifford Heath.

From luislavena at gmail.com  Wed May 21 21:09:23 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Wed, 21 May 2008 22:09:23 -0300
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
Message-ID: <71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>

On Wed, May 21, 2008 at 9:55 PM, Clifford Heath
<clifford.heath at gmail.com> wrote:
> I've used newgem to build a gem that's a simple wrapper for the wbxml
> library,
> and I'm not sure how to publish it. My Manifest includes the ext files,
> including
> extconf.rb, but since newgem & hoe build the gemspec file on the fly, I
> haven't
> been able to work out how to arrange for the extension to be built on
> install.
>
> Can someone give me some pointers please? The reference is too detailed,
> I'm looking for a recipe if possible.
>

You need to tweak somehow the spec Hoe generates inside the Hoe.new
do...end block.

AFAIK, right now Hoe is not able to handle normal C extensions, only
inlined ones (the ones generated via RubyInline).

I think Aslak Hellesoy published something related to this subject:

http://blog.aslakhellesoy.com/2008/3/30/ruby-usb-gems

> Also, what support does gems provide for the naming procedure for binary
> files? If I want to publish platform-specific binary gems in addition to the
> source
> version, what's the procedure?

rubygems itself append the platform specific to the end part of the
filename it generates:

mygem-1.1.1-i386-mingw32.gem

HTH,
-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From clifford.heath at gmail.com  Wed May 21 23:25:25 2008
From: clifford.heath at gmail.com (Clifford Heath)
Date: Thu, 22 May 2008 13:25:25 +1000
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
Message-ID: <97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>

Thanks Luis. I got the gem to build automatically, but the
shared library load path doesn't seem to be set correctly.
The require_paths in the gemspec is ["lib", "ext/wbxml"],
but the .bundle built on MacOSX doesn't load when the
.rb in the lib directory says "require 'wbxml.so'".

Any idea what the story is here?

If anyone wants to check it out, it's the wbxml project on
RubyForge. You'll need libwbxml-dev (or port install wbxml2)
to build it.

The answer to getting the gem to build extensions on install
was to modify the config/hoe.rb generated by "newgem", adding

# Arrange to build the extension on install:
$hoe.spec.extensions = ['ext/wbxml/extconf.rb']

Clifford Heath.

On 22/05/2008, at 11:09 AM, Luis Lavena wrote:

> On Wed, May 21, 2008 at 9:55 PM, Clifford Heath
> <clifford.heath at gmail.com> wrote:
>> I've used newgem to build a gem that's a simple wrapper for the wbxml
>> library,
>> and I'm not sure how to publish it. My Manifest includes the ext  
>> files,
>> including
>> extconf.rb, but since newgem & hoe build the gemspec file on the  
>> fly, I
>> haven't
>> been able to work out how to arrange for the extension to be built on
>> install.
>>
>> Can someone give me some pointers please? The reference is too  
>> detailed,
>> I'm looking for a recipe if possible.
>>
>
> You need to tweak somehow the spec Hoe generates inside the Hoe.new
> do...end block.
>
> AFAIK, right now Hoe is not able to handle normal C extensions, only
> inlined ones (the ones generated via RubyInline).
>
> I think Aslak Hellesoy published something related to this subject:
>
> http://blog.aslakhellesoy.com/2008/3/30/ruby-usb-gems
>
>> Also, what support does gems provide for the naming procedure for  
>> binary
>> files? If I want to publish platform-specific binary gems in  
>> addition to the
>> source
>> version, what's the procedure?
>
> rubygems itself append the platform specific to the end part of the
> filename it generates:
>
> mygem-1.1.1-i386-mingw32.gem
>
> HTH,
> -- 
> Luis Lavena
> AREA 17
> -
> Human beings, who are almost unique in having the ability to learn  
> from
> the experience of others, are also remarkable for their apparent
> disinclination to do so.
> Douglas Adams
> _______________________________________________
> Rubygems-developers mailing list
> Rubygems-developers at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubygems-developers


From clifford.heath at gmail.com  Wed May 21 23:47:20 2008
From: clifford.heath at gmail.com (Clifford Heath)
Date: Thu, 22 May 2008 13:47:20 +1000
Subject: [Rubygems-developers] Fixed: Re: Binary gems or extension source -
	which and how?
In-Reply-To: <71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
Message-ID: <331C78CA-222B-4A1D-B389-4097E749B5B4@gmail.com>

Re my previous problem: I had a file "lib/wbxml.rb", which contained

require 'wbxml.so'

and the shared library was under the ext directory (dirname was in
required_paths), but the fact that the file basename of the file was
the same prevented rubygems from loading it.

The .rb contained nothing of interest, so I removed it and the problem
has gone away. A trap for young players... thanks Luis for offering
help.

Clifford Heath.

From luislavena at gmail.com  Thu May 22 01:02:45 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Thu, 22 May 2008 02:02:45 -0300
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
Message-ID: <71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>

On Thu, May 22, 2008 at 12:25 AM, Clifford Heath
<clifford.heath at gmail.com> wrote:
> Thanks Luis. I got the gem to build automatically, but the
> shared library load path doesn't seem to be set correctly.
> The require_paths in the gemspec is ["lib", "ext/wbxml"],
> but the .bundle built on MacOSX doesn't load when the
> .rb in the lib directory says "require 'wbxml.so'".
>
> Any idea what the story is here?
>

What are you trying to do?

you mention a bundle file, but then a require 'wbxml.so'?

first of you need to remove extensions from your requires, that breaks
compatibility since OSX uses 'bundle' and Linux/Windows uses 'so' as
extensions for the shared objects.

> If anyone wants to check it out, it's the wbxml project on
> RubyForge. You'll need libwbxml-dev (or port install wbxml2)
> to build it.
>

I've checked your Rakefile:

http://wbxml.rubyforge.org/svn/Rakefile

And you're assuming a lot about the extension (and the extension of it too).

I suggest you take a look at DrNic post about newgem generators
related to building C extensions:

http://drnicwilliams.com/2008/04/01/writing-c-extensions-in-rubygems/

> The answer to getting the gem to build extensions on install
> was to modify the config/hoe.rb generated by "newgem", adding
>
> # Arrange to build the extension on install:
> $hoe.spec.extensions = ['ext/wbxml/extconf.rb']
>

As I commented before, Hoe was not aimed to create gems that bundles
extensions out of the box.

Are you trying to build a native (pre-build) platform specific gem? or
are you trying to include in your gem the C extension?

-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From luislavena at gmail.com  Thu May 22 01:04:09 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Thu, 22 May 2008 02:04:09 -0300
Subject: [Rubygems-developers] Fixed: Re: Binary gems or extension
	source - which and how?
In-Reply-To: <331C78CA-222B-4A1D-B389-4097E749B5B4@gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<331C78CA-222B-4A1D-B389-4097E749B5B4@gmail.com>
Message-ID: <71166b3b0805212204s4ec5fe02l1eeb0af64613aa6@mail.gmail.com>

On Thu, May 22, 2008 at 12:47 AM, Clifford Heath
<clifford.heath at gmail.com> wrote:
> Re my previous problem: I had a file "lib/wbxml.rb", which contained
>
> require 'wbxml.so'
>
> and the shared library was under the ext directory (dirname was in
> required_paths), but the fact that the file basename of the file was
> the same prevented rubygems from loading it.
>
> The .rb contained nothing of interest, so I removed it and the problem
> has gone away. A trap for young players... thanks Luis for offering
> help.
>

Great!, too bad I just sent a reply... I follow the FIFO workflow, so
you already answered most of what I commented on that email ;-)

Take care,
-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From clifford.heath at gmail.com  Thu May 22 01:55:58 2008
From: clifford.heath at gmail.com (Clifford Heath)
Date: Thu, 22 May 2008 15:55:58 +1000
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
	<71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>
Message-ID: <674C89FF-8044-416D-9FE4-0542B90055EB@gmail.com>

On 22/05/2008, at 3:02 PM, Luis Lavena wrote:
> you mention a bundle file, but then a require 'wbxml.so'?

Actually there's something in Rubygems that will load a bundle
even if you require .so - I think it'll also load a DLL on windows
if you ask for an so... just to make life easier.

> first of you need to remove extensions from your requires, that breaks
> compatibility since OSX uses 'bundle' and Linux/Windows uses 'so' as
> extensions for the shared objects.

I don't think so. Quite a few gems require *.so, and work cross- 
platform.
Look at RMagick for example.

> I suggest you take a look at DrNic post about newgem generators
> related to building C extensions:
>
> http://drnicwilliams.com/2008/04/01/writing-c-extensions-in-rubygems/

Yes, did that yesterday, he doesn't answer the questions I asked.

> Are you trying to build a native (pre-build) platform specific gem?

I was... not now. It's what I did in the past with the one-wire gem too,
especially since Windows users often don't have a C compiler. I still
haven't learnt how to do that properly, but for now, it's not needed.

Thanks,

Clifford Heath.

From drnicwilliams at gmail.com  Thu May 22 04:22:23 2008
From: drnicwilliams at gmail.com (Nic Williams)
Date: Thu, 22 May 2008 18:22:23 +1000
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <674C89FF-8044-416D-9FE4-0542B90055EB@gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
	<71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>
	<674C89FF-8044-416D-9FE4-0542B90055EB@gmail.com>
Message-ID: <44b555bb0805220122g624e227drc4f35da701eb6962@mail.gmail.com>

If there are any lessons to be learned, and things to be added to the newgem
generator I'd love to have them. The original generator for extensions was
an encapsulation of what I had learned at Euruko so it might not be all
encompassing.
Another option is to replace the Hoe block with Echoe (see echoe gem). I
think it has more support for extensions, but again I don't know what that
extra support is :)

Nic

On Thu, May 22, 2008 at 3:55 PM, Clifford Heath <clifford.heath at gmail.com>
wrote:

> On 22/05/2008, at 3:02 PM, Luis Lavena wrote:
>
>> you mention a bundle file, but then a require 'wbxml.so'?
>>
>
> Actually there's something in Rubygems that will load a bundle
> even if you require .so - I think it'll also load a DLL on windows
> if you ask for an so... just to make life easier.
>
>  first of you need to remove extensions from your requires, that breaks
>> compatibility since OSX uses 'bundle' and Linux/Windows uses 'so' as
>> extensions for the shared objects.
>>
>
> I don't think so. Quite a few gems require *.so, and work cross-platform.
> Look at RMagick for example.
>
>  I suggest you take a look at DrNic post about newgem generators
>> related to building C extensions:
>>
>> http://drnicwilliams.com/2008/04/01/writing-c-extensions-in-rubygems/
>>
>
> Yes, did that yesterday, he doesn't answer the questions I asked.
>
>  Are you trying to build a native (pre-build) platform specific gem?
>>
>
> I was... not now. It's what I did in the past with the one-wire gem too,
> especially since Windows users often don't have a C compiler. I still
> haven't learnt how to do that properly, but for now, it's not needed.
>
> Thanks,
>
> Clifford Heath.
>
> _______________________________________________
> Rubygems-developers mailing list
> Rubygems-developers at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubygems-developers
>



-- 
Dr Nic Williams
* Learn TextMate for Rails2 -
http://peepcode.com/products/textmate-for-rails-2 *
Training in Ruby/Rails/JavaScript - http://drnicacademy.com
Fun with Ruby/Rails/Javascript - http://drnicwilliams.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/rubygems-developers/attachments/20080522/9983e26a/attachment.html>

From luislavena at gmail.com  Thu May 22 04:37:57 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Thu, 22 May 2008 05:37:57 -0300
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <44b555bb0805220122g624e227drc4f35da701eb6962@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
	<71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>
	<674C89FF-8044-416D-9FE4-0542B90055EB@gmail.com>
	<44b555bb0805220122g624e227drc4f35da701eb6962@mail.gmail.com>
Message-ID: <71166b3b0805220137o421d4e85he6a9eda2ec25e3f2@mail.gmail.com>

On Thu, May 22, 2008 at 5:22 AM, Nic Williams <drnicwilliams at gmail.com> wrote:
> If there are any lessons to be learned, and things to be added to the newgem
> generator I'd love to have them. The original generator for extensions was
> an encapsulation of what I had learned at Euruko so it might not be all
> encompassing.

I think there is no solution to fit all the scenarios. (One tool to
rule them all?).

Some developers want 1 extension into their gems (ext/), sometimes
other put another nested level (ext/my_ext), there are others that
have more than 1 extension (ext/{foo,bar}) so can be difficult to fit
all these scenarios out of the box.

> Another option is to replace the Hoe block with Echoe (see echoe gem). I
> think it has more support for extensions, but again I don't know what that
> extra support is :)

Evan added good support to it, mostly extracted of our interaction
with Mongrel gem. You will also get stuff like:

Platform.make, Platform.windows?, Platform.gcc? or Platform.msvc?

These handy definitions can let you: exclude or include dependencies
or task in your Rakefile based on GCC, windows or microsoft compilers
(also java?)

Regards,
-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From jftucker at gmail.com  Thu May 22 05:06:30 2008
From: jftucker at gmail.com (James Tucker)
Date: Thu, 22 May 2008 10:06:30 +0100
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <71166b3b0805220137o421d4e85he6a9eda2ec25e3f2@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
	<71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>
	<674C89FF-8044-416D-9FE4-0542B90055EB@gmail.com>
	<44b555bb0805220122g624e227drc4f35da701eb6962@mail.gmail.com>
	<71166b3b0805220137o421d4e85he6a9eda2ec25e3f2@mail.gmail.com>
Message-ID: <29F22984-2A5F-4B04-9948-6EAB48791FD3@gmail.com>


On 22 May 2008, at 09:37, Luis Lavena wrote:

> On Thu, May 22, 2008 at 5:22 AM, Nic Williams  
> <drnicwilliams at gmail.com> wrote:
>> If there are any lessons to be learned, and things to be added to  
>> the newgem
>> generator I'd love to have them. The original generator for  
>> extensions was
>> an encapsulation of what I had learned at Euruko so it might not be  
>> all
>> encompassing.
>
> I think there is no solution to fit all the scenarios. (One tool to
> rule them all?).

I use this one for situations where there is no pre-built binary gem.  
It's really designed to run on our local gem server (which includes a  
windows build environment), and then it packs up any gems that change:

http://gem-repackager.rubyforge.org/svn/lib/gem_repackager.rb

It probably needs more work, and requires root on most *nix systems.  
It'll also need some tweaks to work with rubyinline, which I've yet to  
complete (thus the lack of release).

Anyway, have fun.

From clifford.heath at gmail.com  Thu May 22 05:24:17 2008
From: clifford.heath at gmail.com (Clifford Heath)
Date: Thu, 22 May 2008 19:24:17 +1000
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <44b555bb0805220122g624e227drc4f35da701eb6962@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<97E4C97A-2C0A-4988-8DA2-903DF9FCE1FA@gmail.com>
	<71166b3b0805212202r13d445d5he8ffab8653f4461d@mail.gmail.com>
	<674C89FF-8044-416D-9FE4-0542B90055EB@gmail.com>
	<44b555bb0805220122g624e227drc4f35da701eb6962@mail.gmail.com>
Message-ID: <B0A71DCF-E85D-48A3-B2FC-C58ECB5596C7@gmail.com>

On 22/05/2008, at 6:22 PM, Nic Williams wrote:
> If there are any lessons to be learned, and things to be added to  
> the newgem generator I'd love to have them.

Ahh, Hi Nic - congrats on becoming a Dad (again?)!

There were two things that tricked me:

If your extension has the name as the gem, and you have a lib/ 
<gemname>.rb file
that's supposed to load it, it doesn't (at least on Leopard). You need  
to give the
extension and the .rb file different basenames. Perhaps on LInux it's  
ok because
the shared library has a libXYZ.so name, I don't know.

The other thing, and perhaps this could go into the extconf_generator,  
is that you
need to add something like this to your config/hoe.rake:

# Arrange to build the extension on install:
$hoe.spec.extensions = ['ext/wbxml/extconf.rb']

where wbxml is a wildcard for your extensions. Otherwise hoe won't  
build them when
the gem gets installed.

Thanks for newgem BTW, it has saved me time several times now. This is  
the first
time I've used it with a gem having a C extension.

In the area of documentation, I used pure rubyforge for source and  
publishing because
I wasn't sure about how integrated the publishing is with source on  
github and the
releases going to rubyforge. If it's all fine and "just works"  
something should say so.

Some more clues or recipes on how to generate a source gem for most  
platforms,
and a (say) Windows gem for deployment there (where folk often don't  
have a
compiler installed) would be good too.

Clifford Heath.

From ryand-ruby at zenspider.com  Thu May 22 17:21:10 2008
From: ryand-ruby at zenspider.com (Ryan Davis)
Date: Thu, 22 May 2008 14:21:10 -0700
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
Message-ID: <F050CE1F-26F2-4804-9733-58D8174C37FE@zenspider.com>


On May 21, 2008, at 18:09 , Luis Lavena wrote:

> You need to tweak somehow the spec Hoe generates inside the Hoe.new
> do...end block.
>
> AFAIK, right now Hoe is not able to handle normal C extensions, only
> inlined ones (the ones generated via RubyInline).


not true:

# -*- ruby -*-

require 'rubygems'
require 'hoe'

EXT = "ext/blah.#{Hoe::DLEXT}"

class Blah
   VERSION = '1.0.0'
end

Hoe.new('blah', Blah::VERSION) do |p|
   p.developer('FIX', 'FIX at example.com')

   p.spec_extras[:extensions] = "ext/extconf.rb"
   p.clean_globs << EXT << "ext/*.o" << "ext/Makefile"
end

task :test => EXT

file EXT => ["ext/extconf.rb", "ext/blah.c"] do
   Dir.chdir "ext" do
     ruby "extconf.rb"
     sh "make"
   end
end

# vim: syntax=Ruby


From luislavena at gmail.com  Thu May 22 18:23:24 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Thu, 22 May 2008 19:23:24 -0300
Subject: [Rubygems-developers] Binary gems or extension source - which
	and how?
In-Reply-To: <F050CE1F-26F2-4804-9733-58D8174C37FE@zenspider.com>
References: <3EF7463B-AF92-4074-9098-CE20FCC2E7CD@gmail.com>
	<71166b3b0805211809u2bf3eae4h4d7f5bf44629bcef@mail.gmail.com>
	<F050CE1F-26F2-4804-9733-58D8174C37FE@zenspider.com>
Message-ID: <71166b3b0805221523o57c86ba7jf4ab8c8bda5e04b5@mail.gmail.com>

On Thu, May 22, 2008 at 6:21 PM, Ryan Davis <ryand-ruby at zenspider.com> wrote:
>
> On May 21, 2008, at 18:09 , Luis Lavena wrote:
>
>> You need to tweak somehow the spec Hoe generates inside the Hoe.new
>> do...end block.
>>
>> AFAIK, right now Hoe is not able to handle normal C extensions, only
>> inlined ones (the ones generated via RubyInline).
>
>
> not true:
>
> # -*- ruby -*-
>
> [snip]

I should said "as far memory remember the last time I looked at the code" :-P

I personally don't use Hoe and the last time I checked it only dealed
with Inlined extensions in pre-compiled form (was me who added it).

:-D

-- 
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From charles.nutter at sun.com  Mon May 26 22:48:07 2008
From: charles.nutter at sun.com (Charles Oliver Nutter)
Date: Mon, 26 May 2008 21:48:07 -0500
Subject: [Rubygems-developers] A couple patches to help JRuby
Message-ID: <483B7667.3040802@sun.com>

FYI, there's a couple bugs with patches that hopefully can get into the 
next maintenance release. We've modified them locally, but we'd like to 
avoid the juggling act we have to do when upgrading:

[#19866] Rubygems fails with NameError: StringIO
http://rubyforge.org/tracker/index.php?func=detail&aid=19866&group_id=126&atid=575

This one occurs because under MRI, YAML pulls in StringIO. Under JRuby, 
it does not, since we have our own YAML implementation. Other 
implementations may or may not have the same problem, so adding an 
explicit require 'stringio' seems like a good idea.

[#20379] RubyGems only looks at platform for win_platform? method, and 
so doesn't install .bat files under JRuby
http://rubyforge.org/tracker/index.php?func=detail&aid=20379&group_id=126&atid=575

JRuby's platform is always "java", but we have the correct OS name in 
rbconfig's "host_os" property. So this fix just adds a second check to 
the win_platform? method, allowing .bat files to install correctly with 
JRuby on Windows.

Any objections to including in the next maintenance release?

- Charlie

From charles.nutter at sun.com  Mon May 26 23:03:07 2008
From: charles.nutter at sun.com (Charles Oliver Nutter)
Date: Mon, 26 May 2008 22:03:07 -0500
Subject: [Rubygems-developers] A couple patches to help JRuby
In-Reply-To: <483B7667.3040802@sun.com>
References: <483B7667.3040802@sun.com>
Message-ID: <483B79EB.2060903@sun.com>

Charles Oliver Nutter wrote:
> FYI, there's a couple bugs with patches that hopefully can get into the 
> next maintenance release. We've modified them locally, but we'd like to 
> avoid the juggling act we have to do when upgrading:

Here's another:

[#20380] RubyGems installs using wrong executable name in JRuby under 
Windows
http://rubyforge.org/tracker/index.php?func=detail&aid=20380&group_id=126&atid=575

In this case, because RubyGems always sticks EXEEXT on the end of 
RUBY_INSTALL_NAME, we end up with an executable named jruby.bat.exe, 
which of course doesn't run. I don't have a good patch for this one, so 
I'm looking for suggestions. Either RubyGems needs to not append when 
RUBY_INSTALL_NAME has the full executable name, or JRuby needs to 
instead report just "jruby" as RUBY_INSTALL_NAME" and then "bat" as 
EXEEXT. However the latter seems a little weird, especially if other 
code is expecting EXEEXT to be "exe".

Thoughts?

- Charlie

From drbrain at segment7.net  Tue May 27 14:59:30 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Tue, 27 May 2008 11:59:30 -0700
Subject: [Rubygems-developers] A couple patches to help JRuby
In-Reply-To: <483B7667.3040802@sun.com>
References: <483B7667.3040802@sun.com>
Message-ID: <64C1EBEF-0ED2-4D7C-A6A8-634B57C511B5@segment7.net>

On May 26, 2008, at 19:48 PM, Charles Oliver Nutter wrote:

> FYI, there's a couple bugs with patches that hopefully can get into  
> the next maintenance release. We've modified them locally, but we'd  
> like to avoid the juggling act we have to do when upgrading:
>
> [#19866] Rubygems fails with NameError: StringIO
> http://rubyforge.org/tracker/index.php?func=detail&aid=19866&group_id=126&atid=575
>
> [#20379] RubyGems only looks at platform for win_platform? method,  
> and so doesn't install .bat files under JRuby
> http://rubyforge.org/tracker/index.php?func=detail&aid=20379&group_id=126&atid=575
>
> Any objections to including in the next maintenance release?

None.  I've marked them as high priority so I shouldn't forget.

From drbrain at segment7.net  Tue May 27 14:59:59 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Tue, 27 May 2008 11:59:59 -0700
Subject: [Rubygems-developers] A couple patches to help JRuby
In-Reply-To: <483B79EB.2060903@sun.com>
References: <483B7667.3040802@sun.com> <483B79EB.2060903@sun.com>
Message-ID: <CAF7E67D-F48B-4674-BC00-B101DD37964E@segment7.net>

On May 26, 2008, at 20:03 PM, Charles Oliver Nutter wrote:

> Charles Oliver Nutter wrote:
>> FYI, there's a couple bugs with patches that hopefully can get into  
>> the next maintenance release. We've modified them locally, but we'd  
>> like to avoid the juggling act we have to do when upgrading:
>
> Here's another:
>
> [#20380] RubyGems installs using wrong executable name in JRuby  
> under Windows
> http://rubyforge.org/tracker/index.php?func=detail&aid=20380&group_id=126&atid=575
>
> In this case, because RubyGems always sticks EXEEXT on the end of  
> RUBY_INSTALL_NAME, we end up with an executable named jruby.bat.exe,  
> which of course doesn't run. I don't have a good patch for this one,  
> so I'm looking for suggestions. Either RubyGems needs to not append  
> when RUBY_INSTALL_NAME has the full executable name, or JRuby needs  
> to instead report just "jruby" as RUBY_INSTALL_NAME" and then "bat"  
> as EXEEXT. However the latter seems a little weird, especially if  
> other code is expecting EXEEXT to be "exe".

Might be able to handle this one with the default overrides we  
discussed.

From drbrain at segment7.net  Tue May 27 17:06:23 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Tue, 27 May 2008 14:06:23 -0700
Subject: [Rubygems-developers] HEADSUP: trunk unusable, new metadata updater
Message-ID: <D07198A8-3DBD-4D65-83FC-C758A612F6AC@segment7.net>

I'm making a checkpoint commit for a new metadata updater that removes  
the bulk update phase.  (lib/rubygems/spec_fetcher.rb)

Tom will be testing out the new indexer in the next day or two, so  
trunk will be unusable against any gem repositories until then.

Also, backwards compatibility with the bulk-update-style index hasn't  
been hooked up.

From thewoolleyman at gmail.com  Tue May 27 21:57:22 2008
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Tue, 27 May 2008 18:57:22 -0700
Subject: [Rubygems-developers] HEADSUP: trunk unusable,
	new metadata updater
In-Reply-To: <D07198A8-3DBD-4D65-83FC-C758A612F6AC@segment7.net>
References: <D07198A8-3DBD-4D65-83FC-C758A612F6AC@segment7.net>
Message-ID: <e3af97510805271857x10269a27yfb7ec28aeeab861d@mail.gmail.com>

On Tue, May 27, 2008 at 2:06 PM, Eric Hodel <drbrain at segment7.net> wrote:
> I'm making a checkpoint commit for a new metadata updater that removes the
> bulk update phase.  (lib/rubygems/spec_fetcher.rb)

Can you make a tag for the last stable revision?

From drbrain at segment7.net  Tue May 27 22:30:06 2008
From: drbrain at segment7.net (Eric Hodel)
Date: Tue, 27 May 2008 19:30:06 -0700
Subject: [Rubygems-developers] HEADSUP: trunk unusable,
	new metadata updater
In-Reply-To: <e3af97510805271857x10269a27yfb7ec28aeeab861d@mail.gmail.com>
References: <D07198A8-3DBD-4D65-83FC-C758A612F6AC@segment7.net>
	<e3af97510805271857x10269a27yfb7ec28aeeab861d@mail.gmail.com>
Message-ID: <20326ED7-5228-4DED-AE67-0FEE2A60B4A4@segment7.net>

On May 27, 2008, at 18:57 PM, Chad Woolley wrote:
> On Tue, May 27, 2008 at 2:06 PM, Eric Hodel <drbrain at segment7.net>  
> wrote:
>> I'm making a checkpoint commit for a new metadata updater that  
>> removes the
>> bulk update phase.  (lib/rubygems/spec_fetcher.rb)
>
> Can you make a tag for the last stable revision?

Tom has deployed the refactored and updated indexer, so now only  
backwards compatibility is missing.  I'll be rolling that in tomorrow,  
so I think it will be ok.