From drbrain at segment7.net Tue Jul 5 16:23:36 2011 From: drbrain at segment7.net (Eric Hodel) Date: Tue, 5 Jul 2011 13:23:36 -0700 Subject: [Rubygems-developers] test (sorry) Message-ID: <084E14E5-828A-4C78-A28C-FD37B2027054@segment7.net> From aaron at tenderlovemaking.com Tue Jul 5 19:29:29 2011 From: aaron at tenderlovemaking.com (Aaron Patterson) Date: Tue, 5 Jul 2011 16:29:29 -0700 Subject: [Rubygems-developers] Gemspec does not allow an empty require_paths list Message-ID: <20110705232929.GA80850@higgins.local> Because of this validation: https://github.com/rubygems/rubygems/blob/master/lib/rubygems/specification.rb#L2189-2192 today it's not possible to generate a gemspec with an empty `require_paths`. I would like to build shell a gem that contain no code, it merely depends on other gems. It's possible to build these shell gems today. However, a directory in the `required_paths` list does not exist, people who install the gem will get an rdoc warning: https://github.com/rails/rails/issues/1646 Eric proposed a workaround of always shipping an empty directory, but I don't think that is a good long term solution. It requires that gem packagers to know that rdoc behaves in a certain way when provided with a directory that doesn't exist, and to ship an empty directory to avoid that warning. Instead, I would rather see this validation removed so that I can specify an empty list of require paths and avoid this warning. I've put together a patch, but I'm not sure that it's complete. Without applying the patch, I get errors running 1.8.7. The errors and failures do not seem to increase after applying my patch, but I cannot be sure I've not broken something. https://gist.github.com/1065473 -- Aaron Patterson http://tenderlovemaking.com/ From evan at fallingsnow.net Tue Jul 5 20:50:13 2011 From: evan at fallingsnow.net (Evan Phoenix) Date: Tue, 5 Jul 2011 17:50:13 -0700 Subject: [Rubygems-developers] Gemspec does not allow an empty require_paths list In-Reply-To: <20110705232929.GA80850@higgins.local> References: <20110705232929.GA80850@higgins.local> Message-ID: <2C10124738484E04862142D24F67B041@gmail.com> See below. -- Evan Phoenix // evan at fallingsnow.net On Tuesday, July 5, 2011 at 4:29 PM, Aaron Patterson wrote: > Because of this validation: > > https://github.com/rubygems/rubygems/blob/master/lib/rubygems/specification.rb#L2189-2192 > > today it's not possible to generate a gemspec with an empty > `require_paths`. I would like to build shell a gem that contain no > code, it merely depends on other gems. Could you provide some info about your use case? Does this new gem exist only to specify dependencies and therefore act as a meta-gem of sorts? > > It's possible to build these shell gems today. However, a directory > in the `required_paths` list does not exist, people who install the > gem will get an rdoc warning: > > https://github.com/rails/rails/issues/1646 > > Eric proposed a workaround of always shipping an empty directory, but I don't > think that is a good long term solution. It requires that gem packagers > to know that rdoc behaves in a certain way when provided with a > directory that doesn't exist, and to ship an empty directory to avoid > that warning. > > Instead, I would rather see this validation removed so that I can > specify an empty list of require paths and avoid this warning. > > I've put together a patch, but I'm not sure that it's complete. Without > applying the patch, I get errors running 1.8.7. The errors and failures > do not seem to increase after applying my patch, but I cannot be sure > I've not broken something. > > https://gist.github.com/1065473 > > -- > Aaron Patterson > http://tenderlovemaking.com/ > _______________________________________________ > Rubygems-developers mailing list > http://rubyforge.org/projects/rubygems > Rubygems-developers at rubyforge.org (mailto:Rubygems-developers at rubyforge.org) > http://rubyforge.org/mailman/listinfo/rubygems-developers From jftucker at gmail.com Tue Jul 5 23:10:43 2011 From: jftucker at gmail.com (James Tucker) Date: Tue, 5 Jul 2011 20:10:43 -0700 Subject: [Rubygems-developers] Gemspec does not allow an empty require_paths list In-Reply-To: <2C10124738484E04862142D24F67B041@gmail.com> References: <20110705232929.GA80850@higgins.local> <2C10124738484E04862142D24F67B041@gmail.com> Message-ID: <3D43B54F-18BA-43A4-AA14-73A5F2004B1E@gmail.com> On Jul 5, 2011, at 5:50 PM, Evan Phoenix wrote: > See below. > > -- > Evan Phoenix // evan at fallingsnow.net > > > On Tuesday, July 5, 2011 at 4:29 PM, Aaron Patterson wrote: > >> Because of this validation: >> >> https://github.com/rubygems/rubygems/blob/master/lib/rubygems/specification.rb#L2189-2192 >> >> today it's not possible to generate a gemspec with an empty >> `require_paths`. I would like to build shell a gem that contain no >> code, it merely depends on other gems. > Could you provide some info about your use case? Does this new gem exist only to specify dependencies and therefore act as a meta-gem of sorts? It may not be Aarons use case, but I've done this before a handful of times for ease of distribution, just packing a very simple binfile into a gem. Example: http://pastie.textmate.org/private/yb1u0pr4w9ywnoztlr39a Yes, this could be written in another language. Yes, it could be packed into a file in lib, and other things. It could be distributed some way other than a gem. It's just convenient to be able to use a cross-platform package manager for a simple purpose such as this, though. From aaron at tenderlovemaking.com Wed Jul 6 00:02:38 2011 From: aaron at tenderlovemaking.com (Aaron Patterson) Date: Tue, 5 Jul 2011 21:02:38 -0700 Subject: [Rubygems-developers] Gemspec does not allow an empty require_paths list In-Reply-To: <2C10124738484E04862142D24F67B041@gmail.com> References: <20110705232929.GA80850@higgins.local> <2C10124738484E04862142D24F67B041@gmail.com> Message-ID: <20110706040238.GA81761@mobile-166-129-152-114.mycingular.net> On Tue, Jul 05, 2011 at 05:50:13PM -0700, Evan Phoenix wrote: > On Tuesday, July 5, 2011 at 4:29 PM, Aaron Patterson wrote: > > > Because of this validation: > > > > https://github.com/rubygems/rubygems/blob/master/lib/rubygems/specification.rb#L2189-2192 > > > > today it's not possible to generate a gemspec with an empty > > `require_paths`. I would like to build shell a gem that contain no > > code, it merely depends on other gems. > Could you provide some info about your use case? Does this new gem exist only to specify dependencies and therefore act as a meta-gem of sorts? Yes, it's the rails gem. The rails gem is just a meta-gem. It doesn't contain any code, just depends on the rest of the gems. -- Aaron Patterson http://tenderlovemaking.com/ From djberg96 at gmail.com Wed Jul 6 09:28:16 2011 From: djberg96 at gmail.com (Daniel Berger) Date: Wed, 6 Jul 2011 08:28:16 -0500 Subject: [Rubygems-developers] universal vs x86 Message-ID: <001201cc3be0$91371100$b3a53300$@com> Hi, I recently pushed a gem (sys-admin) where I set the cpu to "universal" for the windows version, yet I now see it set to "x86". What happened? Dan From luislavena at gmail.com Wed Jul 6 10:45:09 2011 From: luislavena at gmail.com (Luis Lavena) Date: Wed, 6 Jul 2011 11:45:09 -0300 Subject: [Rubygems-developers] universal vs x86 In-Reply-To: <001201cc3be0$91371100$b3a53300$@com> References: <001201cc3be0$91371100$b3a53300$@com> Message-ID: On Wed, Jul 6, 2011 at 10:28 AM, Daniel Berger wrote: > Hi, > > I recently pushed a gem (sys-admin) where I set the cpu to "universal" for > the windows version, yet I now see it set to "x86". > I had no issue generating a gem with "universal-mingw32" locally, gem spec worked and displayed it properly, even gem installation worked: https://gist.github.com/598746009349b717bdea I did gem spec on sys-admin-1.5.5-x86-mingw32.gem and it contained x86-mingw32 -- Luis Lavena AREA 17 - Perfection in design is achieved not when there is nothing more to add, but rather when there is nothing more to take away. Antoine de Saint-Exup?ry From luislavena at gmail.com Fri Jul 8 16:31:25 2011 From: luislavena at gmail.com (Luis Lavena) Date: Fri, 8 Jul 2011 17:31:25 -0300 Subject: [Rubygems-developers] RubyGems is taking pwd/gems as source of gems? Message-ID: This is something I noticed today working with trunk: GEM_HOME and GEM_PATH both point to V:/gems If I'm in V:/ gem list display the gem platforms twice: V:\>gem list sqlite3 sqlite3 (1.3.3 x86-mingw32 x86-mingw32) V:\>gem list rails rails (3.0.9 ruby) === Moving out to another directory, this does not happen: V:\emptyapp>gem list rails rails (3.0.9) V:\emptyapp>gem list sqlite3 sqlite3 (1.3.3 x86-mingw32) Why "gems" in current directory is taking such importance? V:\emptyapp>ruby -v ruby 1.9.3dev (2011-07-08 trunk 32461) [i386-mingw32] V:\emptyapp>gem --version 1.8.5.1 -- Luis Lavena AREA 17 - Perfection in design is achieved not when there is nothing more to add, but rather when there is nothing more to take away. Antoine de Saint-Exup?ry From noreply at rubyforge.org Sat Jul 9 18:31:20 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Sat, 9 Jul 2011 18:31:20 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29313 ] Installer outputs BOM from History.txt Message-ID: <20110709223120.C5525185834E@rubyforge.org> Bugs item #29313, was opened at 2011-07-10 00:31 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29313&group_id=126 Category: RubyGems installer (setup.rb) Group: None Status: Open Resolution: None Priority: 3 Submitted By: Alois Mahdal (aloismahdal) Assigned to: Nobody (None) Summary: Installer outputs BOM from History.txt Initial Comment: After installation, changelog item for last version is displayed to console. However, the file is in UTF-8 with BOM (byte order mark, "EF BB BF") and the BOM is output to console as well, which results in display of three "ugly" characters. Excerpt from output: -----8<----- d:\var\tmp\ror\rubygems-1.8.5\rubygems-1.8.5>ruby setup.rb RubyGems 1.8.5 installed ???=== 1.8.4 / 2011-05-31 * 1 minor enhancement: * The -u option to 'update local source cache' is official deprecated. * Remove has_rdoc deprecations from Specification. * 2 bug fixes: [...] ----->8----- Suggestions: * Filter out the BOM on output * Convert History.txt to ANSI/ISO (History.txt seems to be the only UTF-8 file around anyway) ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29313&group_id=126 From noreply at rubyforge.org Sat Jul 9 18:36:13 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Sat, 9 Jul 2011 18:36:13 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29313 ] Installer outputs BOM from History.txt Message-ID: <20110709223614.80FCB1858300@rubyforge.org> Bugs item #29313, was opened at 2011-07-10 00:31 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29313&group_id=126 Category: RubyGems installer (setup.rb) Group: None >Status: Deleted Resolution: None Priority: 3 Submitted By: Alois Mahdal (aloismahdal) Assigned to: Nobody (None) Summary: Installer outputs BOM from History.txt Initial Comment: After installation, changelog item for last version is displayed to console. However, the file is in UTF-8 with BOM (byte order mark, "EF BB BF") and the BOM is output to console as well, which results in display of three "ugly" characters. Excerpt from output: -----8<----- d:\var\tmp\ror\rubygems-1.8.5\rubygems-1.8.5>ruby setup.rb RubyGems 1.8.5 installed ???=== 1.8.4 / 2011-05-31 * 1 minor enhancement: * The -u option to 'update local source cache' is official deprecated. * Remove has_rdoc deprecations from Specification. * 2 bug fixes: [...] ----->8----- Suggestions: * Filter out the BOM on output * Convert History.txt to ANSI/ISO (History.txt seems to be the only UTF-8 file around anyway) ---------------------------------------------------------------------- >Comment By: Alois Mahdal (aloismahdal) Date: 2011-07-10 00:36 Message: (Sorry for OT guys but the note about github should be *way* more outstanding) Moving issue to GitHub. ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29313&group_id=126 From noreply at rubyforge.org Sun Jul 10 08:54:33 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Sun, 10 Jul 2011 08:54:33 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29163 ] Invalid specs resulting from incorrect syck time parsing Message-ID: <20110710125433.2EDD418583A3@rubyforge.org> Bugs item #29163, was opened at 2011-05-04 23:14 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29163&group_id=126 Category: `gem` commands (other) Group: next Status: Closed Resolution: Accepted Priority: 3 Submitted By: Bruno Michel (nono) Assigned to: Evan Phoenix (evan) Summary: Invalid specs resulting from incorrect syck time parsing Initial Comment: Hi, I have the bug described on https://github.com/rubygems/rubygems/pull/57 with Ruby 1.9.2-p180 and Rubygems v1.8.0. If I generate a gem with `gem build`, the `date` field in the metadata of the generated gem is: date: 2011-05-04 00:00:00.000000000Z If I push this gem on rubygems.org and try to install it remotely (same PC), this does not works: ERROR: While executing gem ... (NameError) uninitialized constant Psych::Syck ---------------------------------------------------------------------- Comment By: Aryk Grosz (gotskill10) Date: 2011-07-10 12:54 Message: Evan, were your fixes in 0f4067a still in the most current version of RubyGems. I was looking at 1.8.5 and don't see this. I'm using the latest version of RubyGems and struggling with this issue. ---------------------------------------------------------------------- Comment By: Dan Peterson (dpiddy) Date: 2011-06-22 16:59 Message: Should the fix be extended to `marshal_dump` as well? I'm running into a problem on our internal gems repo where built-gem metadata is being uploaded with unquoted ='s. When `gem generate_index` is run, the metadata is loaded which turns the = into DefaultKey. That's then marshaled out as part of the quick specs which breaks clients that try to load them. Applying the same fix to `marshal_dump` let quick specs be generated that didn't have DefaultKey refs in them even though the incoming gem specs were broken. ---------------------------------------------------------------------- Comment By: Bruno Michel (nono) Date: 2011-06-01 20:53 Message: Oki, now I understand from where this DefaultKey come. Thanks ---------------------------------------------------------------------- Comment By: Evan Phoenix (evan) Date: 2011-06-01 20:45 Message: Yep. You can see the syck bug by running this code: YAML.load "- - =\n - 1\n" You'll see that syck puts DefaultKey in there, which is just plain wrong. ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-06-01 20:42 Message: The problem is that psych will generate a perfectly valid string that starts with an unquoted '='. Syck has a bug parsing that and winds up using DefaultKey. ---------------------------------------------------------------------- Comment By: Bruno Michel (nono) Date: 2011-06-01 19:13 Message: Thanks for the explanation. The gem was regenerated with a Ruby 1.8 and a new version was pushed. If I want to build the next version with Ruby 1.9, how can I can be sure the new gem don't have a YAML::Syck::DefaultKey? ---------------------------------------------------------------------- Comment By: Evan Phoenix (evan) Date: 2011-06-01 18:09 Message: I've fixed this by having the DefaultKey class always available and having Requirement fix up any reference to it. Fixed in github:0f4067a ---------------------------------------------------------------------- Comment By: Evan Phoenix (evan) Date: 2011-06-01 16:16 Message: This is because html_spellchecker has an instance of YAML::Syck::DefaultKey in its gemspec. These should not be in gemspecs and the gem needs to be regenerated. I'm looking into a way to mitigate this in the mean time. ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-05-24 10:20 Message: I think it is your cache... That URL looks OK via: curl http://gems:Nasom4 at gems.af83.com/specs.4.8.gz | gunzip -dc | ruby -Ilib -rubygems -e 'p Marshal.load $stdin.read' ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-05-24 10:12 Message: Bruno, this installs fine for me as you can see below... From your output, I suspect you might have a bad source in your list. Specifically "gems.af83.com" looks like it might be bad for you. Can you do `gem sources -c` to ensure that your sources are default and your caches are cleared and try again to verify? % gem i -i xxx html_spellchecker --version=0.1.1 -V --backtrace GET http://rubygems.org/specs.4.8.gz 302 Found GET http://production.s3.rubygems.org/specs.4.8.gz 200 OK GET http://rubygems.org/quick/Marshal.4.8/html_spellchecker-0.1.1.gemspec.rz ... and so on to completion ... ---------------------------------------------------------------------- Comment By: Bruno Michel (nono) Date: 2011-05-24 09:43 Message: Hi, for example: ~% gem install html_spellchecker --version=0.1.1 -V --backtrace GET http://rubygems.org/specs.4.8.gz 302 Found GET http://production.s3.rubygems.org/specs.4.8.gz 304 Not Modified GET http://gems:Nasom4 at gems.af83.com/specs.4.8.gz 200 OK ERROR: While executing gem ... (NameError) uninitialized constant Psych::Syck /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/specification.rb:640:in `load' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/specification.rb:640:in `_load' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:134:in `load' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:134:in `fetch_spec' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:95:in `block in fetch_with_errors' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:94:in `map' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:94:in `fetch_with_errors' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/dependency_installer.rb:110:in `find_gems_with_sources' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/dependency_installer.rb:228:in `find_spec_by_name_and_version' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/dependency_installer.rb:259:in `install' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/commands/install_command.rb:121:in `block in execute' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/commands/install_command.rb:115:in `each' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/commands/install_command.rb:115:in `execute' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/command.rb:278:in `invoke' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/command_manager.rb:147:in `process_args' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/command_manager.rb:117:in `run' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/gem_runner.rb:65:in `run' /home/nono/.rvm/rubies/ruby-1.9.2-p180/bin/gem:21:in `
' zsh: exit 1 ~% gem -v 1.8.3 ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-05-24 00:38 Message: I need a full backtrace and the gem name+version to properly investigate. Your error doesn't jive with the date problem. (If I'm remembering the error correctly). ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29163&group_id=126 From noreply at rubyforge.org Sun Jul 10 08:58:34 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Sun, 10 Jul 2011 08:58:34 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29163 ] Invalid specs resulting from incorrect syck time parsing Message-ID: <20110710125834.B415618583A4@rubyforge.org> Bugs item #29163, was opened at 2011-05-04 23:14 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29163&group_id=126 Category: `gem` commands (other) Group: next Status: Closed Resolution: Accepted Priority: 3 Submitted By: Bruno Michel (nono) Assigned to: Evan Phoenix (evan) Summary: Invalid specs resulting from incorrect syck time parsing Initial Comment: Hi, I have the bug described on https://github.com/rubygems/rubygems/pull/57 with Ruby 1.9.2-p180 and Rubygems v1.8.0. If I generate a gem with `gem build`, the `date` field in the metadata of the generated gem is: date: 2011-05-04 00:00:00.000000000Z If I push this gem on rubygems.org and try to install it remotely (same PC), this does not works: ERROR: While executing gem ... (NameError) uninitialized constant Psych::Syck ---------------------------------------------------------------------- Comment By: Aryk Grosz (gotskill10) Date: 2011-07-10 12:58 Message: Nevermind my last comment. I just noticed that it was in master. Do you know when the next version of rubygems might get released? I noticed those fixes weren't in 1.8.5. Is there a way to run an edge version of rubygems? ---------------------------------------------------------------------- Comment By: Aryk Grosz (gotskill10) Date: 2011-07-10 12:54 Message: Evan, were your fixes in 0f4067a still in the most current version of RubyGems. I was looking at 1.8.5 and don't see this. I'm using the latest version of RubyGems and struggling with this issue. ---------------------------------------------------------------------- Comment By: Dan Peterson (dpiddy) Date: 2011-06-22 16:59 Message: Should the fix be extended to `marshal_dump` as well? I'm running into a problem on our internal gems repo where built-gem metadata is being uploaded with unquoted ='s. When `gem generate_index` is run, the metadata is loaded which turns the = into DefaultKey. That's then marshaled out as part of the quick specs which breaks clients that try to load them. Applying the same fix to `marshal_dump` let quick specs be generated that didn't have DefaultKey refs in them even though the incoming gem specs were broken. ---------------------------------------------------------------------- Comment By: Bruno Michel (nono) Date: 2011-06-01 20:53 Message: Oki, now I understand from where this DefaultKey come. Thanks ---------------------------------------------------------------------- Comment By: Evan Phoenix (evan) Date: 2011-06-01 20:45 Message: Yep. You can see the syck bug by running this code: YAML.load "- - =\n - 1\n" You'll see that syck puts DefaultKey in there, which is just plain wrong. ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-06-01 20:42 Message: The problem is that psych will generate a perfectly valid string that starts with an unquoted '='. Syck has a bug parsing that and winds up using DefaultKey. ---------------------------------------------------------------------- Comment By: Bruno Michel (nono) Date: 2011-06-01 19:13 Message: Thanks for the explanation. The gem was regenerated with a Ruby 1.8 and a new version was pushed. If I want to build the next version with Ruby 1.9, how can I can be sure the new gem don't have a YAML::Syck::DefaultKey? ---------------------------------------------------------------------- Comment By: Evan Phoenix (evan) Date: 2011-06-01 18:09 Message: I've fixed this by having the DefaultKey class always available and having Requirement fix up any reference to it. Fixed in github:0f4067a ---------------------------------------------------------------------- Comment By: Evan Phoenix (evan) Date: 2011-06-01 16:16 Message: This is because html_spellchecker has an instance of YAML::Syck::DefaultKey in its gemspec. These should not be in gemspecs and the gem needs to be regenerated. I'm looking into a way to mitigate this in the mean time. ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-05-24 10:20 Message: I think it is your cache... That URL looks OK via: curl http://gems:Nasom4 at gems.af83.com/specs.4.8.gz | gunzip -dc | ruby -Ilib -rubygems -e 'p Marshal.load $stdin.read' ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-05-24 10:12 Message: Bruno, this installs fine for me as you can see below... From your output, I suspect you might have a bad source in your list. Specifically "gems.af83.com" looks like it might be bad for you. Can you do `gem sources -c` to ensure that your sources are default and your caches are cleared and try again to verify? % gem i -i xxx html_spellchecker --version=0.1.1 -V --backtrace GET http://rubygems.org/specs.4.8.gz 302 Found GET http://production.s3.rubygems.org/specs.4.8.gz 200 OK GET http://rubygems.org/quick/Marshal.4.8/html_spellchecker-0.1.1.gemspec.rz ... and so on to completion ... ---------------------------------------------------------------------- Comment By: Bruno Michel (nono) Date: 2011-05-24 09:43 Message: Hi, for example: ~% gem install html_spellchecker --version=0.1.1 -V --backtrace GET http://rubygems.org/specs.4.8.gz 302 Found GET http://production.s3.rubygems.org/specs.4.8.gz 304 Not Modified GET http://gems:Nasom4 at gems.af83.com/specs.4.8.gz 200 OK ERROR: While executing gem ... (NameError) uninitialized constant Psych::Syck /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/specification.rb:640:in `load' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/specification.rb:640:in `_load' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:134:in `load' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:134:in `fetch_spec' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:95:in `block in fetch_with_errors' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:94:in `map' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/spec_fetcher.rb:94:in `fetch_with_errors' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/dependency_installer.rb:110:in `find_gems_with_sources' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/dependency_installer.rb:228:in `find_spec_by_name_and_version' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/dependency_installer.rb:259:in `install' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/commands/install_command.rb:121:in `block in execute' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/commands/install_command.rb:115:in `each' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/commands/install_command.rb:115:in `execute' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/command.rb:278:in `invoke' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/command_manager.rb:147:in `process_args' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/command_manager.rb:117:in `run' /home/nono/.rvm/rubies/ruby-1.9.2-p180/lib/ruby/site_ruby/1.9.1/rubygems/gem_runner.rb:65:in `run' /home/nono/.rvm/rubies/ruby-1.9.2-p180/bin/gem:21:in `
' zsh: exit 1 ~% gem -v 1.8.3 ---------------------------------------------------------------------- Comment By: Ryan Davis (zenspider) Date: 2011-05-24 00:38 Message: I need a full backtrace and the gem name+version to properly investigate. Your error doesn't jive with the date problem. (If I'm remembering the error correctly). ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29163&group_id=126 From noreply at rubyforge.org Wed Jul 13 13:54:04 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Wed, 13 Jul 2011 13:54:04 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29316 ] ruby_code case not handled: Regexp Message-ID: <20110713175404.3255A1218275@rubyforge.org> Bugs item #29316, was opened at 2011-07-13 10:54 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 Category: `gem install` command Group: None Status: Open Resolution: None Priority: 3 Submitted By: Gonzalo Garramuno (gga) Assigned to: Nobody (None) Summary: ruby_code case not handled: Regexp Initial Comment: ERROR: While executing gem ... (Gem::Exception) ruby_code case not handled: Regexp When installing the ifmapper gem in rubygems.org in ubuntu's ruby-1.9.1-full. Trying to build the gem with the same gem1.9 (-v 1.8.5), I obtain a gem that when installed reports: ERROR: While executing gem ... (Gem::Package::FormatError) corrupt gem (Gem::Package::TarInvalidError: tar is corrupt, name contains null byte) in /media/Linux/src/bin/ifmapper/ifmapper-1.1.4/ifmapper-1.1.4.gem ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 From noreply at rubyforge.org Wed Jul 13 13:54:45 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Wed, 13 Jul 2011 13:54:45 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29316 ] ruby_code case not handled: Regexp Message-ID: <20110713175448.894DC1218275@rubyforge.org> Bugs item #29316, was opened at 2011-07-13 10:54 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 Category: `gem install` command Group: None Status: Open Resolution: None Priority: 3 Submitted By: Gonzalo Garramuno (gga) Assigned to: Nobody (None) Summary: ruby_code case not handled: Regexp Initial Comment: ERROR: While executing gem ... (Gem::Exception) ruby_code case not handled: Regexp When installing the ifmapper gem in rubygems.org in ubuntu's ruby-1.9.1-full. Trying to build the gem with the same gem1.9 (-v 1.8.5), I obtain a gem that when installed reports: ERROR: While executing gem ... (Gem::Package::FormatError) corrupt gem (Gem::Package::TarInvalidError: tar is corrupt, name contains null byte) in /media/Linux/src/bin/ifmapper/ifmapper-1.1.4/ifmapper-1.1.4.gem ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 From noreply at rubyforge.org Wed Jul 13 14:17:42 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Wed, 13 Jul 2011 14:17:42 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29316 ] ruby_code case not handled: Regexp Message-ID: <20110713181742.2226518583A4@rubyforge.org> Bugs item #29316, was opened at 2011-07-13 14:54 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 Category: `gem install` command Group: None Status: Open Resolution: None Priority: 3 Submitted By: Gonzalo Garramuno (gga) Assigned to: Nobody (None) Summary: ruby_code case not handled: Regexp Initial Comment: ERROR: While executing gem ... (Gem::Exception) ruby_code case not handled: Regexp When installing the ifmapper gem in rubygems.org in ubuntu's ruby-1.9.1-full. Trying to build the gem with the same gem1.9 (-v 1.8.5), I obtain a gem that when installed reports: ERROR: While executing gem ... (Gem::Package::FormatError) corrupt gem (Gem::Package::TarInvalidError: tar is corrupt, name contains null byte) in /media/Linux/src/bin/ifmapper/ifmapper-1.1.4/ifmapper-1.1.4.gem ---------------------------------------------------------------------- Comment By: Luis Lavena (luislavena) Date: 2011-07-13 15:17 Message: Hi! Believe it or not, we're moving to github's issue tracker! The pressure to move has been loud and clear, but for the longest time it didn't support the way we worked. Now that it has freeform labels it can support our workflow adequately. There is just one problem: Your bug is here and we don't want to lose you! What we'd like you to do is fairly painless and should only take a few minutes: Please file the _exact_ same ticket on github and provide a link back to this crufty old ticket so we can refer to any previous discussion that occurred. Once the new ticket is made, put a forwarding link to it here and close this ticket out. In a couple weeks, we'll track down the old tickets that are still open and decide what to do with them. We could have done this fairly automatically, but we really want you to do it so you're attached to the ticket and we can still communicate with you and keep a record of it. https://github.com/rubygems/rubygems/issues Thanks, The RubyGems Team ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 From noreply at rubyforge.org Wed Jul 13 14:18:18 2011 From: noreply at rubyforge.org (noreply at rubyforge.org) Date: Wed, 13 Jul 2011 14:18:18 -0400 (EDT) Subject: [Rubygems-developers] [ rubygems-Bugs-29316 ] ruby_code case not handled: Regexp Message-ID: <20110713181818.95CFC1218275@rubyforge.org> Bugs item #29316, was opened at 2011-07-13 14:54 You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 Category: `gem install` command Group: None Status: Open Resolution: None Priority: 3 Submitted By: Gonzalo Garramuno (gga) Assigned to: Nobody (None) Summary: ruby_code case not handled: Regexp Initial Comment: ERROR: While executing gem ... (Gem::Exception) ruby_code case not handled: Regexp When installing the ifmapper gem in rubygems.org in ubuntu's ruby-1.9.1-full. Trying to build the gem with the same gem1.9 (-v 1.8.5), I obtain a gem that when installed reports: ERROR: While executing gem ... (Gem::Package::FormatError) corrupt gem (Gem::Package::TarInvalidError: tar is corrupt, name contains null byte) in /media/Linux/src/bin/ifmapper/ifmapper-1.1.4/ifmapper-1.1.4.gem ---------------------------------------------------------------------- >Comment By: Luis Lavena (luislavena) Date: 2011-07-13 15:18 Message: Hello, Please report this to GitHub, the tracker is not even linked on RubyForge page anymore. ---------------------------------------------------------------------- Comment By: Luis Lavena (luislavena) Date: 2011-07-13 15:17 Message: Hi! Believe it or not, we're moving to github's issue tracker! The pressure to move has been loud and clear, but for the longest time it didn't support the way we worked. Now that it has freeform labels it can support our workflow adequately. There is just one problem: Your bug is here and we don't want to lose you! What we'd like you to do is fairly painless and should only take a few minutes: Please file the _exact_ same ticket on github and provide a link back to this crufty old ticket so we can refer to any previous discussion that occurred. Once the new ticket is made, put a forwarding link to it here and close this ticket out. In a couple weeks, we'll track down the old tickets that are still open and decide what to do with them. We could have done this fairly automatically, but we really want you to do it so you're attached to the ticket and we can still communicate with you and keep a record of it. https://github.com/rubygems/rubygems/issues Thanks, The RubyGems Team ---------------------------------------------------------------------- You can respond by visiting: http://rubyforge.org/tracker/?func=detail&atid=575&aid=29316&group_id=126 From transfire at gmail.com Fri Jul 15 08:55:57 2011 From: transfire at gmail.com (Trans) Date: Fri, 15 Jul 2011 05:55:57 -0700 (PDT) Subject: [Rubygems-developers] Searching for files among gems Message-ID: <8d0d3cde-1655-46e7-b354-617addc1cb20@c29g2000yqd.googlegroups.com> How would one use the RubyGems API to handle the following: 1. find matching files in gems 2. that are within the require_paths of the gems 3. where the gems are active gems 4. or, if not active, then the latest version of a gem Is there anything in the API that would make this easier, or even trivial? Or is the only way a brute force approach? Thanks. From djberg96 at gmail.com Fri Jul 15 09:24:26 2011 From: djberg96 at gmail.com (Daniel Berger) Date: Fri, 15 Jul 2011 07:24:26 -0600 Subject: [Rubygems-developers] platform.rb regex concerns Message-ID: <4E203F8A.40604@gmail.com> Hi, I noticed in platform.rb we have what appears to a redundant regular expression and some things that are too strict. Regarding the redundant regex I see this on line 75-76 of platform.rb: when /^dotnet$/ then [ 'dotnet', nil ] when /^dotnet([\d.]*)/ then [ 'dotnet', $1 ] What case does the first one cover that the second one doesn't? Also, it seems the regexen for aix, freebsd, hpux, openbsd and solaris are too strict. If no version is specified the os will come back as "unknown". For example: irb(main):001:0> Gem::Platform.new("universal-solaris10.2") => # irb(main):002:0> Gem::Platform.new("universal-solaris") => # Other platforms allow an OS to be set with a nil version. Why not those? I can get around it manually, but would rather not have to. I would suggest that these be modified to make the digits optional as they currently are for darwin. Regards, Dan From ryand-ruby at zenspider.com Sun Jul 17 04:11:15 2011 From: ryand-ruby at zenspider.com (Ryan Davis) Date: Sun, 17 Jul 2011 01:11:15 -0700 Subject: [Rubygems-developers] platform.rb regex concerns In-Reply-To: <4E203F8A.40604@gmail.com> References: <4E203F8A.40604@gmail.com> Message-ID: <0D29E5EB-E330-489E-89E2-F3509F87E154@zenspider.com> On Jul 15, 2011, at 06:24 , Daniel Berger wrote: > I noticed in platform.rb we have what appears to a redundant regular expression and some things that are too strict. Regarding the redundant regex I see this on line 75-76 of platform.rb: > > when /^dotnet$/ then [ 'dotnet', nil ] > when /^dotnet([\d.]*)/ then [ 'dotnet', $1 ] > > What case does the first one cover that the second one doesn't? I assume that the latter should be "+", not "*"? Sounds like a valid bug to me. > Also, it seems the regexen for aix, freebsd, hpux, openbsd and solaris are too strict. If no version is specified the os will come back as "unknown". For example: > > irb(main):001:0> Gem::Platform.new("universal-solaris10.2") > => # > > irb(main):002:0> Gem::Platform.new("universal-solaris") > => # > > Other platforms allow an OS to be set with a nil version. Why not those? I can get around it manually, but would rather not have to. I would suggest that these be modified to make the digits optional as they currently are for darwin. Probably also bugs. Please file a ticket and feel free to assign to me. From drbrain at segment7.net Thu Jul 21 18:00:00 2011 From: drbrain at segment7.net (Eric Hodel) Date: Thu, 21 Jul 2011 15:00:00 -0700 Subject: [Rubygems-developers] Gemspec does not allow an empty require_paths list In-Reply-To: <20110705232929.GA80850@higgins.local> References: <20110705232929.GA80850@higgins.local> Message-ID: <1F044E90-42D7-4EBC-B080-A91913A1C212@segment7.net> On Jul 5, 2011, at 4:29 PM, Aaron Patterson wrote: > Eric proposed a workaround of always shipping an empty directory, but I don't > think that is a good long term solution. It requires that gem packagers > to know that rdoc behaves in a certain way when provided with a > directory that doesn't exist, and to ship an empty directory to avoid > that warning. > > Instead, I would rather see this validation removed so that I can > specify an empty list of require paths and avoid this warning. > > I've put together a patch, but I'm not sure that it's complete. Without > applying the patch, I get errors running 1.8.7. The errors and failures > do not seem to increase after applying my patch, but I cannot be sure > I've not broken something. > > https://gist.github.com/1065473 I think the lack of a require path should be a warning instead of an error. Most people won't want to use this feature. I don't think the errors are your fault. From ryand-ruby at zenspider.com Thu Jul 21 18:43:28 2011 From: ryand-ruby at zenspider.com (Ryan Davis) Date: Thu, 21 Jul 2011 15:43:28 -0700 Subject: [Rubygems-developers] [ANN] isolate 3.1.1 Released Message-ID: <245F71BE-612D-44BA-9D9F-24387FD9F40D@zenspider.com> isolate version 3.1.1 has been released! * Isolate is a very simple RubyGems sandbox. It provides a way to express and automatically install your project's Gem dependencies. Changes: ### 3.1.1 / 2011-07-21 * 2 bug fixes: * Deprecate is now Gem::Deprecate. Try both. * Don't install gems for other environments. (rhburrows) From celoserpa at gmail.com Thu Jul 21 22:55:55 2011 From: celoserpa at gmail.com (Marcelo de Moraes Serpa) Date: Thu, 21 Jul 2011 21:55:55 -0500 Subject: [Rubygems-developers] Using rubygems for another language/platform Message-ID: Hey guys, I love rubygems, I think it's the best packaging system out there. I've been spoiled by it throughout my 5 years of working with Ruby. I was wondering if it could be extended / modified to serve packages for other languages. Let's say I develop a new language X, and start creating new libs. Instead of creating a new packaging system, I'd like to adapt RubyGems to work with it. Doesn't matter if ruby is needed. I know that ultimately it would be possible, I just don't know how hard would it be and if other people already though about that. Perhaps someone is already doing something similar out there :) Thanks in advance, Marcelo. From drbrain at segment7.net Fri Jul 22 12:24:54 2011 From: drbrain at segment7.net (Eric Hodel) Date: Fri, 22 Jul 2011 09:24:54 -0700 Subject: [Rubygems-developers] Using rubygems for another language/platform In-Reply-To: References: Message-ID: <245CA821-77F7-4C2B-BE55-32F00E8D0F38@segment7.net> On Jul 21, 2011, at 7:55 PM, Marcelo de Moraes Serpa wrote: > I love rubygems, I think it's the best packaging system out there. I've been > spoiled by it throughout my 5 years of working with Ruby. > > I was wondering if it could be extended / modified to serve packages for > other languages. Let's say I develop a new language X, and start creating > new libs. Instead of creating a new packaging system, I'd like to adapt > RubyGems to work with it. Doesn't matter if ruby is needed. Since RubyGems is written in Ruby using it for another programming language seems a little odd. Requiring Ruby to install packages for other languages seems odd. > I know that ultimately it would be possible, I just don't know how hard > would it be and if other people already though about that. Perhaps someone > is already doing something similar out there :) As far as I know, nobody has ported RubyGems or its packages to another language. It's certainly possible to port RubyGems to another language, but it would probably be better to steal the ideas of RubyGems instead. A .gem is a tar file containing a chunk of gzipped metadata and a gzipped tar file with the contents. The tar format is easy to implement in ruby and libz bindings can be found most anywhere. If I were writing RubyGems from scratch I wouldn't make the specification behave the same way. In particular, it's too easy for people building packages to touch things they're not supposed to like specification_version. Also, RubyGems' commands were built then API was extracted from the commands long after. I would aggressively refactor API out of the commands as they were being built. From nick at quaran.to Fri Jul 22 14:11:25 2011 From: nick at quaran.to (Nick Quaranto) Date: Fri, 22 Jul 2011 14:11:25 -0400 Subject: [Rubygems-developers] Using rubygems for another language/platform In-Reply-To: <245CA821-77F7-4C2B-BE55-32F00E8D0F38@segment7.net> References: <245CA821-77F7-4C2B-BE55-32F00E8D0F38@segment7.net> Message-ID: As a heads up, some folks in the .NET community started http://nuget.org/first by going with rubygems as a delivery mechanism. http://dannydouglass.com/2010/08/whats-new-in-the-dotnet-world/ http://dannydouglass.com/2010/10/nupack-vs-nuproj-vs-openwrap/ On Fri, Jul 22, 2011 at 12:24 PM, Eric Hodel wrote: > On Jul 21, 2011, at 7:55 PM, Marcelo de Moraes Serpa wrote: > > I love rubygems, I think it's the best packaging system out there. I've > been > > spoiled by it throughout my 5 years of working with Ruby. > > > > I was wondering if it could be extended / modified to serve packages for > > other languages. Let's say I develop a new language X, and start creating > > new libs. Instead of creating a new packaging system, I'd like to adapt > > RubyGems to work with it. Doesn't matter if ruby is needed. > > Since RubyGems is written in Ruby using it for another programming language > seems a little odd. Requiring Ruby to install packages for other languages > seems odd. > > > I know that ultimately it would be possible, I just don't know how hard > > would it be and if other people already though about that. Perhaps > someone > > is already doing something similar out there :) > > As far as I know, nobody has ported RubyGems or its packages to another > language. It's certainly possible to port RubyGems to another language, but > it would probably be better to steal the ideas of RubyGems instead. > > A .gem is a tar file containing a chunk of gzipped metadata and a gzipped > tar file with the contents. The tar format is easy to implement in ruby and > libz bindings can be found most anywhere. > > If I were writing RubyGems from scratch I wouldn't make the specification > behave the same way. In particular, it's too easy for people building > packages to touch things they're not supposed to like specification_version. > > Also, RubyGems' commands were built then API was extracted from the > commands long after. I would aggressively refactor API out of the commands > as they were being built. > _______________________________________________ > Rubygems-developers mailing list > http://rubyforge.org/projects/rubygems > Rubygems-developers at rubyforge.org > http://rubyforge.org/mailman/listinfo/rubygems-developers > From drbrain at segment7.net Fri Jul 22 16:38:21 2011 From: drbrain at segment7.net (Eric Hodel) Date: Fri, 22 Jul 2011 13:38:21 -0700 Subject: =?utf-8?Q?=E2=86=90_Subject_prefix_is_now_gone?= Message-ID: <2F684D3D-5A9F-4494-85A3-079384E38CC5@segment7.net> If you were filtering based off of the subject you should switch to the List-Id header: List-Id: RubyGems developers mailing list From drbrain at segment7.net Mon Jul 25 20:52:18 2011 From: drbrain at segment7.net (Eric Hodel) Date: Mon, 25 Jul 2011 17:52:18 -0700 Subject: [ANN] rubygems-update 1.8.6 Released Message-ID: <2E35CDEE-8A95-44AB-B2D6-503C05E025C4@segment7.net> rubygems-update version 1.8.6 has been released! * http://rubygems.org * http://docs.rubygems.org * http://help.rubygems.org * http://github.com/rubygems * http://rubyforge.org/projects/rubygems RubyGems is a package management framework for Ruby. This gem is an update for the RubyGems software. You must have an installation of RubyGems before this update can be applied. See Gem for information on RubyGems (or `ri Gem`) To upgrade to the latest RubyGems, run: $ gem update --system # you might need to be an administrator or root See UPGRADING.rdoc for more details and alternative instructions. ----- If you don't have RubyGems installed, your can still do it manually: * Download from: https://rubygems.org/pages/download * Unpack into a directory and cd there * Install with: ruby setup.rb # you may need admin/root privilege For more details and other options, see: ruby setup.rb --help Changes: === 1.8.6 / 2011-07-25 * Major enhancements: * Restore behavior of Gem::Specification#loaded? Ruby Bug #5032 * Minor enhancements: * Add autorequires and delay startup of RubyGems until require is called. See Ruby bug #4962 * Bug fixes: * Clean up SourceIndex.add_specs to not be so damn noisy. (tadman) * Added missing APPLE_GEM_HOME in paths. * Extend YAML::Syck::DefaultKey fixing to `marshal_dump` as well. * Fix #29216: check correct bin_dir in check_that_user_bin_dir_is_in_path. * Revert Gem.latest_load_paths to working order (PathSupport revert). * Restore normalization of GEM_HOME. * Handle the Syck DefaultKey problem once and for all. * Fix SystemStackError occurring with "gem list -r -a" on 1.9. From nick at quaran.to Fri Jul 29 16:54:49 2011 From: nick at quaran.to (Nick Quaranto) Date: Fri, 29 Jul 2011 16:54:49 -0400 Subject: rubygems organization repo reorganization Message-ID: Just a heads up, I've done a bit of shuffling around on GitHub, please update your remotes! Gemcutter (the rails app) is now: https://github.com/rubygems/rubygems.org Gemcutter (the gem) is now: https://github.com/rubygems/gemcutter I've also promoted two more repos to the rubygems org: Erik Michaels-Ober's gems gem, which is an api implementation/layer for Rubygems.org: https://github.com/rubygems/gems My gemwhisperer repo, which is an example implementation of how to make an endpoint for `gem webhook`: https://github.com/rubygems/gemwhisperer gemwhisperer also powers http://twitter.com/rubygems, and has a decent iUI mobile interface: http://gem.heroku.com/ -Nick