summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2009-07-26Check for "SunOS" in post-review and use gdiff instead of difftrowbrds
Solaris has BSD "diff" and GNU "gdiff". Since we use things which are only present in the GNU userland, we should account for that. This change checks for SunOS in the output of os.uname and uses gdiff if necessary.
2009-07-18Patch by Chris Clark to define a SCMCLIENTS list containing all knownchipx86
SCMClients, which makes it easier to update and maintain. Reviewed at http://reviews.review-board.org/r/882/
2009-07-18Install post-review as a Python module, with a generated post-review (orchipx86
post-review.exe) wrapping it. The official way to install post-review now is by installing RBTools. Reviewed at http://reviews.review-board.org/r/916/
2009-06-12Use 'json' module if it's available.chipx86
simplejson is now part of the Python standard library, but it's named 'json'. This change alters post-review to use it if it's available, which reduces the amount of stuff people have to install (if they have Python 2.6+) Committing on behalf of David. Reviewed http://reviews.review-board.org/r/887/
2009-06-10Fix a missing space when building a revision string for ClearCase.chipx86
2009-06-02Reintroduce forcing removal of the cookie when cookie-based validation fails ↵chipx86
the first time. This was broken in r1986.
2009-06-02Mark a file as read/write instead of write-only in PerforceClient._write_file.chipx86
Fixes bug #1143.
2009-05-28Patch by Steven Russell to fix working with Perforce file paths containingchipx86
spaces. Fixes bug #840. Reviewed at http://reviews.review-board.org/r/815/
2009-05-28Patch by Raghu Kaippully to normalize files from Perforce with \r\r\nchipx86
newlines, translating them to \r\n. This can happen in some Perforce setups when dealing with a file checked out from one platform and modified on another, for example. Fixes bug #1071. Reviewed at http://reviews.review-board.org/r/826/
2009-05-28Add Chris Clark to the authors list.chipx86
2009-05-28Patch by Chris Clark to set a max number of login retries, so as tochipx86
prevent entering an infinite loop. This can happen when the cookie is assumed to be valid but doesn't work properly on the server. While this won't fix logins in that situation, it will prevent the infinite loop followed by a "maximum recursion deptch" error. Reviewed at http://reviews.review-board.org/r/879/
2009-05-19Check for "hg: unknown command" as well as "abort:" when checking ifchipx86
"hg svn" works. Otherwise it assumes the repository is hg-svn.
2009-05-19Fix the ClearCaseClient to not break post-review if cleartool isn't ↵chipx86
installed. We now properly do a check_install on "clearcase help".
2009-04-23Patch by Jeremy Bettis to properly join a SVN repository base path and a ↵chipx86
file path using a "/" instead of "\" on Windows. This fixes bug #939.
2009-04-14Patch by Eric Huss to support posting paths of files to Review Boardchipx86
when using Perforce. This method doesn't take a change number, but rather a set of files and optional revisions. His documentation states: "We have some use cases where it would be nice to review entire files post-commit in reviewboard. Some examples: - Our documentation needs to be checked in before it is reviewed. If multiple people are working on it, or if it is created as a series of multiple checkins, reviewing specific change numbers is difficult. With this, we can do "post-review //docs/main/some/project/...". - If you work on a side branch for a while, possibly with multiple checkins or with multiple people, reviewing the changes on the side branch before integration to the main branch can be tricky. With this, you can now do "post-review //path/to/my/branch/...@100,@120" where "100" would be the initial revision that created the branch and "120" is the last revision on that branch. This allows you to easily review all changes on a branch in one review request. - If you want to review an old, existing file that has had little or no reviews, you can now do this with "post-review //path/to/some/file" to review the whole thing. The specific path types supported are: post-review //path/to/file # Upload file as a "new" file. post-review //path/to/dir/... # Upload all files as "new" files. post-review //path/to/file[@#]rev # Upload file from that rev as a "new" file. # (Not a very useful scenario, but it works.) post-review //path/to/file[@#]rev,[@#]rev # Upload a diff between revs. post-review //path/to/dir/...[@#]rev,[@#]rev # Upload a diff of all files between revs in that directory. You can specify multiple paths on the command line. I didn't use the range revision option because the Perforce path syntax is a little more flexible, especially if you post multiple paths. Reviewed at http://reviews.review-board.org/r/753/
2009-04-14Patch by Eric Huss to fix an infinite recursion problem when usingchipx86
post-review with invalid cookies. Fixes bug #808. Reviewed at http://reviews.review-board.org/r/813/
2009-03-29Bump the version to 0.2beta2.dev.chipx86
2009-03-29Release rbtools 0.2 beta 1.chipx86
2009-03-29Get the absolute path of the git directory before getting the parentchipx86
directory. Fixes bug #996.
2009-03-27This isn't a release, so don't mark it as such.chipx86
2009-03-27Patch by Raghu Kaippully to fix changesets with binary files in Perforce ↵chipx86
when using post-review. Reviewed at http://reviews.review-board.org/r/787/ Fixes bug #585
2009-03-27Move post-review into a new top-level "rbtools" packagechipx86
Since the dawn of time, post-review has lived in the Review Board tree, despite being a client script and not necessarily being needed on the server. This moves post-review into a new rbtools package, which users will be able to easy_install (easy_install -U RBTools). The simplejson dependency will be fetched, and users won't have to go hunting in SVN for a script. Over time, I'd like to add new scripts into here (such as one to apply a diff from a review request), and further down the road I'd like to create an "rbapi" (or something) package that contains a Python API for talking to a Review Board server, which rbtools (and post-review) will then use. For now, the goal is to just get post-review moved over. Reviewed at http://reviews.review-board.org/r/788/