[Rubygems-developers] Test failures for 0.9.4.4 beta on Windows

Eric Hodel drbrain at segment7.net
Wed Sep 19 22:57:31 EDT 2007


On Aug 31, 2007, at 07:27, Berger, Daniel wrote:
>> -----Original Message-----
>> From: rubygems-developers-bounces at rubyforge.org
>> [mailto:rubygems-developers-bounces at rubyforge.org] On Behalf
>> Of Daniel Berger
>> Sent: Wednesday, August 29, 2007 8:37 PM
>> To: rubygems-developers at rubyforge.org
>> Subject: [Rubygems-developers] Test failures for 0.9.4.4 beta
>> on Windows
>>
>>
>> Hi all,
>>
>> Ruby 1.8.5 (one click)
>> Windows XP SP 2
>> Rubygems tag REL_0_9_4_4
>>
>> Most of these look like path handling or teardown issues in the test
>> suite itself, but I'll investigate a couple of these.
>>
>> Look for a patch this weekend...
>
> <snip>
>
> I've submitted a series of patches on the project page that cut it  
> down
> to 2 failures and 2 errors.
>
> Please see:
>
> http://rubyforge.org/tracker/index.php? 
> func=detail&aid=13528&group_id=12
> 6&atid=577
> http://rubyforge.org/tracker/index.php? 
> func=detail&aid=13529&group_id=12
> 6&atid=577
> http://rubyforge.org/tracker/index.php? 
> func=detail&aid=13533&group_id=12
> 6&atid=577
> http://rubyforge.org/tracker/index.php? 
> func=detail&aid=13535&group_id=12
> 6&atid=577
> http://rubyforge.org/tracker/index.php? 
> func=detail&aid=13537&group_id=12
> 6&atid=577

All applied!

> Here are the remaining issues:
>
>   1) Error:
> test_execute(TestGemCommandsPristineCommand):
> SecurityError: Insecure operation - initialize
>     ./lib/rubygems/installer.rb:272:in `initialize'
>     ./lib/rubygems/installer.rb:272:in `open'
>     ./lib/rubygems/installer.rb:272:in `shebang'
>     ./lib/rubygems/installer.rb:290:in `app_script_text'
>     ./lib/rubygems/installer.rb:229:in `generate_bin_script'
>     ./lib/rubygems/installer.rb:228:in `open'
>     ./lib/rubygems/installer.rb:228:in `generate_bin_script'
>     ./lib/rubygems/installer.rb:241:in `generate_bin_symlink'
>     ./lib/rubygems/installer.rb:215:in `generate_bin'
>     ./lib/rubygems/installer.rb:206:in `each'
>     ./lib/rubygems/installer.rb:206:in `generate_bin'
>     ./lib/rubygems/installer.rb:112:in `install'
>     ./test/gemutilities.rb:117:in `install_gem'
>     ./test/test_gem_commands_pristine_command.rb:19:in `test_execute'
>
> I may have introduced this bug in my attempt to fix the filehandle
> closing issue in patch #13533. I haven't investigated further.

Not sure, I'm not getting the SecurityError with your patches.

>   2) Failure:
> test_download_gem_local_read_only(TestGemDependencyInstaller)
> [./test/test_gem_dependency_installer.rb:356]:
> <"C:/WINNT/temp/test_rubygems_2032/a-1.gem"> expected but was
> <"C:/WINNT/temp/test_rubygems_2032/gemhome/cache/a-1.gem">.
>
> I'm not sure if the patch is bad or the test is bad. See my comment in
> patch #13537.

This test uses chmod which doesn't do anything on mswin.

>   3) Error:
> test_generate_index(TestGemIndexerIndexer):
> Zlib::BufError: buffer error
>     ./lib/rubygems/indexer/compressor.rb:22:in `inflate'
>     ./lib/rubygems/indexer/compressor.rb:22:in `unzip'
>     ./lib/rubygems/indexer/master_index_builder.rb:38:in `paranoid'
>     ./lib/rubygems/indexer/master_index_builder.rb:20:in `cleanup'
>     ./lib/rubygems/indexer/abstract_index_builder.rb:42:in `build'
>     ./lib/rubygems/indexer/indexer.rb:34:in `build_index'
>     ./lib/rubygems/indexer/indexer.rb:91:in `generate_index'
>     ./test/test_gem_indexer_indexer.rb:34:in `test_generate_index'
>     ./lib/rubygems/user_interaction.rb:45:in `use_ui'
>     ./lib/rubygems/user_interaction.rb:27:in `use_ui'
>     ./test/test_gem_indexer_indexer.rb:33:in `test_generate_index'
>
> No clue. Will require diving into some ugly C code with a wet suit  
> and a
> spear gun.

Is this repeatable?  I had one of these go by once in the last day.

>   4) Failure:
> test_self_match_version(TestGemPlatform)
> [./test/test_gem_platform.rb:74]:
> versionless == any.
> <false> is not true.
>
> This failure is caused by changing the architecture in patch  
> #13529, but
> the lack of a release number in Config::CONFIG['arch'] for MS Windows
> made me unsure how the tests should be refactored.

I've reworked Gem::Platform to be a real object now with methods like  
#=== and #=~ and so-forth, so this may have moved.

--
Poor workers blame their tools. Good workers build better tools. The
best workers get their tools to do the work for them. -- Syndicate Wars




More information about the Rubygems-developers mailing list