summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Version 1.1.5 with changelogv1.1.51-1-stableTerence Lee2012-07-172-1/+7
|
* special case `ruby` directiveTerence Lee2012-07-172-0/+27
| | | | | | This is light of the upcoming 1.2.0 release which adds this feature. This way people still using 1.1.x can run 1.2.0 Gemfiles that employ this feature and get a notification about upgrading.
* Fix typos in error messagesMani Tadayon2012-06-052-2/+2
|
* Version 1.1.4 with changelogAndre Arko2012-05-272-1/+11
|
* Revert "suppress rubygems deprecations during artifice setup"Andre Arko2012-05-271-27/+14
| | | | | | | It turns out suppressing the errors like this changes the output and causes the tests to fail. So oh well. Back to deprecation warnings... This reverts commit 040430d2fb5f41baf7de3295909e5fd38c194090.
* make tests pass under travis.Tim Miller2012-05-271-1/+6
| | | | | | | sudo tests were failing under travis because they were not being run as root. this change runs the root tests as root, and chowns the tests' "tmp" directory back to the original user to allow the subsequent 'realworld' tests to pass.
* work around Ruby bug empty files in YAML.load_fileAndre Arko2012-05-271-2/+10
| | | | Ruby 1.9.3p194 includes a change to Psych to support BOM markers. Unfortunately, that change means that Psych will sometimes throw an exception when you try to call YAML.load_file on a file with zero bytes in it. This commit includes a workaround so that we don't call YAML.load_file if the config file is empty.
* print the backtrace of install errors in verbose modeAndre Arko2012-05-271-0/+1
|
* instructions for when OpenSSL is missingAndre Arko2012-05-272-0/+44
|
* grammar fix for spec nameAndre Arko2012-05-271-1/+1
|
* we're already inside the Bundler namespace hereAndre Arko2012-05-271-1/+1
|
* guard against double-adding vendor to the load pathAndre Arko2012-05-271-1/+2
|
* add vendored_persistentAndre Arko2012-05-272-1/+4
| | | | This clears up fetcher.rb's implicit dependency on being loaded after vendored_thor (which added lib/vendor/ to the load path).
* suppress rubygems deprecations during artifice setupAndre Arko2012-05-271-14/+27
|
* revamp ISSUESAndre Arko2012-05-271-20/+16
| | | | | - remove frequent issues that aren't so frequent anymore - move Heroku instructions first, rewrite for clarity - link directly to the new ticket page
* point users at ISSUES.md instead of a new ticketAndre Arko2012-05-271-5/+3
|
* fixes for capistrano integrationAndre Arko2012-05-272-5/+5
| | | | | | | | switch the bundler hook to before "deploy:finalize_update", so that finalize_update will not run if `bundle install` fails. (thanks @leehambley) change from `current_release` to `latest_release`, because both cap and vlad provide more robust directory-finding code behind `latest_release`. (thanks @dontangg) closes #1264
* Fixing "bundle clean --force" banner.Doug Barth2012-05-241-1/+1
| | | | The --force option was incorrectly stating that it forces a clean if --path IS set. It actually forces a clean if --path is NOT set.
* Bundler 1.1.3v1.1.3Terence Lee2012-03-231-1/+1
|
* update changelog for 1.1.3Terence Lee2012-03-231-0/+6
|
* bundler root path should be escapedAaron Patterson2012-03-231-1/+1
|
* #1789. test when regex chars are in bundler rootTerence Lee2012-03-231-0/+23
|
* helper for custom app roots in specsTerence Lee2012-03-231-0/+4
|
* Bundler 1.1.2v1.1.2Terence Lee2012-03-201-1/+1
|
* update changelog for 1.1.2Terence Lee2012-03-201-0/+6
|
* Fixes #1782. Don't remove duplicate PATH sourcesTerence Lee2012-03-202-1/+15
| | | | | | | | | | When using specifying version number, you can have multiple PATH sections for the same source. We were over eager in https://github.com/carlhuda/bundler/commit/e31a1c7b24d3dfbce7ed39b6a723dd20f7d08c95#L1R51. We should only remove the duplicate GIT sources, so let's explicitly check for that.
* Test case for bundle install not checking out gitTerence Lee2012-03-201-0/+126
| | | | | | | | During the 1.1 pre/RC releases, there were duplicate GIT source sections generated in the `Gemfile.lock`. These would confuse bundler and the wrong source would get attached to the spec. This test case breaks if we don't fix this here: https://github.com/carlhuda/bundler/commit/e31a1c7b24d3dfbce7ed39b6a723dd20f7d08c95#L1R51
* Bundler 1.1.1v1.1.1Terence Lee2012-03-141-1/+1
|
* update changelog for 1.1.1Terence Lee2012-03-131-0/+1
|
* fix bundle cache edge case for #1202Terence Lee2012-03-132-1/+12
|
* Include 'your Gemfile and Gemfile.lock' in fatal exception errorRyan Bigg2012-03-131-3/+3
|
* Add 'Please include the full output of the command' to fatal error messageRyan Bigg2012-03-131-1/+3
|
* Changelog for 1.1.1Andre Arko2012-03-131-0/+14
|
* Fix URI exception by encoding gem names for rubygems dependency API #1672rohit2012-03-102-1/+12
|
* avoid warning from setting encoding if possibleAndre Arko2012-03-101-1/+1
|
* fix void use of == warningAndre Arko2012-03-101-2/+2
|
* fix shadowed variable warningAndre Arko2012-03-101-3/+3
|
* Rescue EAGAIN too so it works for JRuby on WindowsAndre Arko2012-03-101-1/+1
|
* somewhat confusingly, Bundler.ui.error is STDOUTAndre Arko2012-03-101-1/+1
|
* raising Gem::InstallError is the worst possible thingAndre Arko2012-03-103-2/+3
| | | This code meant that _any_ gem install failure resulted in a message requesting the user to report this "bug" to the Bundler issue tracker. That is very, very not cool, and I had already fixed this issue once before the hook pull request regressed it. :(
* clarify that "no sources" means only local sourcesAndre Arko2012-03-102-3/+5
|
* Split `git ls-files` using $\, so gemspecs will work on Windows.Postmodern2012-03-071-1/+1
|
* Reduce invocation of `git ls-files`.Akinori MUSHA2012-03-071-2/+2
| | | | | This also removes shell brace expasion which is a ksh/bash extension that does not work with POSIX sh.
* Version 1.1.0 \o/v1.1.0Terence Lee2012-03-071-1/+1
|
* update changelogTerence Lee2012-03-071-1/+3
|
* Bump versiontomhuda2012-03-071-1/+1
|
* Clean up corrupted lockfilestomhuda2012-03-074-1/+101
| | | | | | | | | Bundler 1.1.pre.5 introduced a bug that could introduce duplicate GIT sections, which then caused duplicate copies of gems in the GIT sections. This commit makes the LockfileParser aware of the bug, and has it clean up any corrupted lockfiles.
* Merge pull request #1695 from geemus/bundlerTerence Lee2012-03-061-1/+5
|\ | | | | | | | | | | ---
| * prevent duplicate sourcesgeemus2012-03-061-1/+5
| |
* | #1694: add failing test for duplicate git sourcesTerence Lee2012-03-061-0/+16
| |
View Lorry Controller Status