summaryrefslogtreecommitdiff
path: root/lib
Commit message (Collapse)AuthorAgeFilesLines
* Compatibility with frozen string literalsSamuel Giddins2016-01-3193-29/+130
|
* [RuboCop] Fix Style/MutableConstantSamuel Giddins2016-01-3111-27/+31
|
* [RuboCop] Fix Performance/TimesMapSamuel Giddins2016-01-311-1/+1
|
* [RuboCop] Fix Performance/StartWithSamuel Giddins2016-01-311-2/+2
|
* [RuboCop] Address Style/GuardClauseSamuel Giddins2016-01-3120-119/+94
|
* [PathPreserver] Use the passed-in envseg-preserve-pathSamuel Giddins2016-01-311-5/+5
|
* Refactor path preservation into generic codeSamuel Giddins2016-01-313-18/+14
|
* Preserve PATH in with_clean_envSamuel Giddins2016-01-312-0/+8
|
* Auto merge of #4250 - ↵Homu2016-01-313-3/+8
|\ | | | | | | | | | | | | | | | | | | RochesterinNYC:improved-error-message-for-invalid-gemspec, r=indirect Change behavior and improve clarity of error messages when local gems have invalid gemspecs Now, when a local gem's gemspec contains errors, an error is raised immediately and helpful message is provided. This message instructs the user to fix the gemspec before an attempt is made to fetch the local gem that has an incorrect gemspec. - Addresses #4248
| * Change behavior and improve clarity of error messages when local gems have ↵James Wen2016-01-303-3/+8
| | | | | | | | | | | | | | | | | | invalid gemspecs - Addresses #4248 - Implements behavior so that when a local gem's gemspec contains errors, an error is raised immediately and a helpful message is provided. This message instructs the user to fix the gemspec before further resolving steps are taken.
* | Auto merge of #4249 - cirdes:outdated-semantic, r=indirectHomu2016-01-312-0/+14
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | Bundle outdated with major and minor options This feature adds the ability to specify a outdated command to show only the gems with a major or minor updates. I was looking for this feature and found the issue https://github.com/bundler/bundler/issues/3805 @indirect, how can I improve this PR?
| * | bundle outdated with major and minor optionsCirdes Henrique2016-01-292-0/+14
| | | | | | | | | | | | | | | This feature adds the ability to specify on outdated command to show only the gems with a major or minor updates
* | | Auto merge of #4244 - dtognazzini:fix-path-to-gemfile, r=indirectHomu2016-01-312-7/+18
|\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix the path to the Gemfile during evaluation. This is a continuation of https://github.com/bundler/bundler/pull/3349 Following up per: https://github.com/bundler/bundler/pull/3349#issuecomment-88307579 The issue here is that paths used with `Bundler::Dsl#gemspec` will only work when the Gemfile being evaluated is `Bundler.default_gemfile`. Passing a Gemfile other than `Bundler.default_gemfile` to `Bundler::Dsl.evaluate` will break uses of `path:` options in the Gemfile. These changes update `Bundler::Dsl` to remember the Gemfile passed into `eval_gemfile` and use it to resolve relative paths.
| * | | Forwarding root path to Source::Path from Dsl.Donnie Tognazzini2016-01-272-13/+13
| | | |
| * | | Make all paths be relative to Gemfile during evaluation.Donnie Tognazzini2016-01-271-3/+11
| | | |
| * | | Fix the path to the Gemfile during evaluation.Donnie Tognazzini2016-01-271-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | `Bundler.default_gemfile` was being used instead of the explicitly passed Gemfile path. From: https://github.com/bundler/bundler/commit/ea3ded94a6f44dea237142b6b200e902cdd4d6ba
* | | | Auto merge of #4254 - bundler:seg-thread-names, r=indirectHomu2016-01-313-4/+14
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | [Worker] Add a worker name to ease debugging Also helps with https://github.com/bundler/bundler/issues/4246. @indirect r?
| * | | | [Worker] Add a worker name to ease debuggingseg-thread-namesSamuel Giddins2016-01-303-4/+14
| | |_|/ | |/| |
* | | | Auto merge of #4252 - bundler:seg-debug-once, r=segiddinsHomu2016-01-311-1/+2
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | [Resolver] Determine debug mode once Should be a tiny speed boost
| * | | | [Resolver] Determine debug mode onceseg-debug-onceSamuel Giddins2016-01-301-1/+2
| |/ / /
* | | | [Resolver] Update Molinillo to 0.4.2Samuel Giddins2016-01-308-24/+72
|/ / /
* | | Auto merge of #4231 - bundler:seg-exec-spaces, r=indirectHomu2016-01-281-1/+1
|\ \ \ | |/ / |/| | | | | | | | | | | [Exec] Fix exec-ing with 0 args in a directory with spaces Closes #4230 by forcing `Kernel.exec` to never use a shell
| * | [Exec] Fix exec-ing with 0 args in a directory with spacesseg-exec-spacesSamuel Giddins2016-01-251-1/+1
| | |
* | | Fail in an informative way when no YAML is available at allSamuel Giddins2016-01-251-1/+1
| | |
* | | Update dependency fetcher specs for refactoringSamuel Giddins2016-01-252-2/+3
| | |
* | | [Definition] Insure @locked_ruby_version is initializedSamuel Giddins2016-01-251-0/+1
| | |
* | | [Fetcher] Avoid method redefinition warningSamuel Giddins2016-01-251-0/+1
| | |
* | | [CompactIndexClient] Don't request when the gem does not existSamuel Giddins2016-01-251-1/+2
| | | | | | | | | | | | This will save us getting a known 304 when attempting to download a gem the server doesnt have any indexed versions for
* | | [Updater] Simplify content logicSamuel Giddins2016-01-251-3/+2
| | |
* | | [CompactIndex] Send the proper arity function to the WorkerSamuel Giddins2016-01-251-1/+2
| | |
* | | [EndpointSpec] Avoid creating unused hashesSamuel Giddins2016-01-251-1/+2
| | |
* | | [CompactIndex] Centralize error handlingSamuel Giddins2016-01-251-14/+20
| | |
* | | [CompactIndexClient] Don't raise Bundler::HTTPErrorSamuel Giddins2016-01-253-3/+7
| | |
* | | [CompactIndex] Fetch in parallel using Bundler::WorkerSamuel Giddins2016-01-251-0/+7
| | |
* | | [CompactIndexClient] Remove dependency upon Bundler::WorkerSamuel Giddins2016-01-251-8/+11
| | |
* | | Refactor cache slug logic into the RemoteSamuel Giddins2016-01-252-8/+17
| | |
* | | [EndpointSpecification] Remove redundant nil-checksSamuel Giddins2016-01-251-3/+3
| | |
* | | [CompactIndex] Use the cache_uri for the cache pathSamuel Giddins2016-01-251-2/+3
| | |
* | | [Remote] Add the concept of a cache_uriSamuel Giddins2016-01-251-1/+2
| | |
* | | [EndpointSpecification] Freeze constant stringSamuel Giddins2016-01-251-1/+1
| | |
* | | [CompactIndexClient] Require setSamuel Giddins2016-01-251-0/+1
| | |
* | | update digest to include host+port as well as pathAndre Arko2016-01-251-6/+10
| | | | | | | | | | | | also shortening to the first 6 hex chars to reduce path length
* | | find is better than .select.firstAndre Arko2016-01-251-2/+1
| | | | | | | | | …yeah, that one seems clear in retrospect
* | | uniq the completed list per @segiddinsAndre Arko2016-01-251-1/+1
| | |
* | | unify global config and compact index cache dirsAndre Arko2016-01-252-3/+6
| | |
* | | it’s supposed to return an arrayAndre Arko2016-01-251-3/+3
| | | | | | | | | Turns out that very old RubyGems returned a hash here, with each remote URI as the key and the array of specs as the value. We don’t deal with multiple remotes at the same time, though, so we just want the array of spec values.
* | | ArgumentErrors are not really retriableAndre Arko2016-01-251-1/+1
| | |
* | | flat_map doesn’t exist on 1.8.7 :/Andre Arko2016-01-251-1/+1
| | |
* | | resolve warning for extra proc parametersAndre Arko2016-01-251-1/+1
| | |
* | | turns out URI#hostname is 1.9+, URI#host is 1.8+Andre Arko2016-01-251-4/+5
| | |