Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

Comparing changes

Choose two branches to see what's changed or to start a new pull request. If you need to, you can also compare across forks.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also compare across forks.
base fork: reviewboard/rbtools
...
head fork: reviewboard/rbtools
Checking mergeability… Don't worry, you can still create the pull request.
Commits on Nov 06, 2011
@vestel vestel Fixed typo in ADD_REPOSITORY_DOCS_URL to show current docs directory
Due to changed URL structure on reviewboard website, this variable
should be updated.

Reviewed at http://http://reviews.reviewboard.org/r/2665/
a2ba0e7
Commits on Jan 22, 2012
@fallergyula fallergyula Informative error message on not being able to locate GNU diff for Pe…
…rforce

Currently, if GNU diff is not in the path, post-review gives an unhandled
exception and a call stack. It scares the user, and left him without any clue
what is wrong. I found this is solved in case of SVN so I ported that solution
to Perforce too.

Fixes bug 2352
Reviewed at http://reviews.reviewboard.org/r/2685/
ab4ad64
@fallergyula fallergyula Informative error message on not being able to locate GNU diff for Pe…
…rforce

Currently, if GNU diff is not in the path, post-review gives an unhandled
exception and a call stack. It scares the user, and left him without any clue
what is wrong. I found this is solved in case of SVN so I ported that solution
to Perforce too.

Fixes bug 2352
Reviewed at http://reviews.reviewboard.org/r/2685/
b6130ec
@fallergyula fallergyula Fix handling of new files in perforce post-commit scenarios.
It is a erroneous asumption that an added file has no revisions yet. It is true
for pre-commit reviews, but far from true in post commit reviews. So if a file
is added it must be explicitly specified that we need its first revision.
Otherwise we will get its last revision by default.

Fixes bug 2369
Reviewed at http://reviews.reviewboard.org/r/2686/
947cc35
@fallergyula fallergyula Fix handling of new files in perforce post-commit scenarios.
It is a erroneous asumption that an added file has no revisions yet. It is true
for pre-commit reviews, but far from true in post commit reviews. So if a file
is added it must be explicitly specified that we need its first revision.
Otherwise we will get its last revision by default.

Fixes bug 2369
Reviewed at http://reviews.reviewboard.org/r/2686/
59a7ec5
Commits on Feb 06, 2012
@bcelary bcelary Repository URL option causes UnboundLocalError for git repository
Looks like a bug to me. The origin_url.rstrip is already present in the else
statement above it. In the if statement the origin_url is not instantiated and
error comes up. There is also a possibility of getting empty upstream_branch
(i.e. == '') but I am not sure how to fix it properly. The fix comes from my
problems with posting git review when my origin url differs from the one present
on the RB server. The following error comes up without this fix:

File "/data/Work/Development/rbtools/rbtools/clients/git.py", line 153, in
get_repository_info
    url = origin_url.rstrip('/')
    UnboundLocalError: local variable 'origin_url' referenced before assignment

Fixes bug 2410
49bb969
@bcelary bcelary Repository URL option causes UnboundLocalError for git repository
Looks like a bug to me. The origin_url.rstrip is already present in the else
statement above it. In the if statement the origin_url is not instantiated and
error comes up. There is also a possibility of getting empty upstream_branch
(i.e. == '') but I am not sure how to fix it properly. The fix comes from my
problems with posting git review when my origin url differs from the one
present on the RB server. The following error comes up without this fix:

File "/data/Work/Development/rbtools/rbtools/clients/git.py", line 153, in get_repository_info
        url = origin_url.rstrip('/')
        UnboundLocalError: local variable 'origin_url' referenced before assignment

Fixes bug 2410
143dfb7
Commits on Feb 07, 2012
@davidt davidt Show a nicer error when trying to update someone else's review request.
When trying to post a change to a review request owned by someone else (and
without permissions to make changes to it), post-review would print its usual
nonsense about "Your review request still exists, but the diff is not attached."

This change checks for the specific permission-denied error code to show a
useful error instead.

Fixes bug 2465.
7570943
Commits on Feb 10, 2012
@davidt davidt Fix a broken link.
When running post-review against a server with no configured repositories, it
would print out a link to the documentation for adding them. Unfortunately, when
the docs got rearranged, this link got broken. This had already been fixed on
release-0.3.x but never made it to master. Trivial fix.

Fixes bug 2482.
6d524dd
Commits on Feb 12, 2012
@davidt davidt Don't crash if home directory isn't writable.
Most of our CookieJar.save() calls were wrapped in a try/except to catch
IOErrors, but the one in http_put wasn't. This meant if the home directory
wasn't writable (for example, in some automated post-commit scenarios),
post-review would crash with an uncaught exception. Trivial fix.

Fixes bug 2344.
1356362
@davidt davidt Don't crash if home directory isn't writable.
Most of our CookieJar.save() calls were wrapped in a try/except to catch
IOErrors, but the one in http_put wasn't. This meant if the home directory
wasn't writable (for example, in some automated post-commit scenarios),
post-review would crash with an uncaught exception. Trivial fix.

Fixes bug 2344.
ade9fe0
Commits on Feb 13, 2012
@davidt davidt Allow using --diff-filename=- with a valid cookie.
When using --diff-filename=-, there was a check to make sure that users wouldn't
have to type in their username and password (since stdin was used up by the
pipe), but this check didn't take into account that there might be a cookie with
the authentication information. This made it okay for automation but kind of
lame for people who actually wanted to use pipes themselves.

Fixes bug 2052.
91ca2eb
Commits on Feb 26, 2012
@vslavik vslavik Fix --repository-url handling for Git.
There were several problems:

(1) If --repository-url was used, post-review failed with
    UnboundLocalError: local variable 'origin_url' reference before
    assignment

    Fixed by using this code only in the autodetection case.

(2) self.upstream branch wasn't always set, thus failing later.

(3) The parameter was ignored completely when using git-svn.

Reviewed at http://reviews.reviewboard.org/r/2647/
d8db3de
Commits on Feb 27, 2012
@chipx86 chipx86 Merge branch 'release-0.3.x' 77dd4d5
@chipx86 chipx86 Add support for default parameter values in .reviewboardrc.
There are many parameters to post-review that people would like to make
default, either through a repository's .reviewboardrc or through their
own user's. There's also some parameters people were previously able to
hard-code inside postreview.py, but that makes less sense now.

These can now be set in a .reviewboardrc. The following are accepted:
PUBLISH, OPEN_BROWSER, TARGET_GROUPS, TARGET_PEOPLE, GUESS_SUMMARY,
GUESS_DESCRIPTION, BRANCH, SUBMIT_AS, USERNAME, PASSWORD, PARENT_BRANCH,
TRACKING_BRANCH, P4_CLIENT, P4_PORT, P4_PASSWD, DEBUG, HTTP_USERNAME,
and HTTP_PASSWORD.
1347dd4
@chipx86 chipx86 Add a -g, --guess-fields parameter.
I find myself typing --guess-summary --guess-description a lot. This
adds a new parameter, --guess-fields (or just -g) that is equivalent to
those two parameters. It's settable by the GUESS_FIELDS value in
.reviewboardrc.
80f01ba
@chipx86 chipx86 Add a --disable-proxy option for disabling the proxy server.
A common problem people hit is that they end up going through a proxy
server when they don't want to. This change adds a --disable-proxy
option (DISABLE_PROXY in .reviewboardrc) that disables the HTTP(S) proxy
support.
adda2b2
@chipx86 chipx86 Fix an early return when using a server with the deprecated API.
When using the deprecated API, check_api_version() bails too early. This
seems okay in real use, but breaks in unit tests. Now it continues on.
e5515e3
@chipx86 chipx86 Make the release script consistent with Review Board's.
This also changes some incorrect product names to say "RBTools."
52bae2c
@chipx86 chipx86 Release RBTools 0.4.0. 008f267
Commits on Feb 28, 2012
@chipx86 chipx86 Fix the check for determining whether to post a release to PyPI.
The wrong index in a tuple was being used, resulting in us not posting
releases to PyPI.
4a58454
Commits on Feb 29, 2012
@chipx86 chipx86 Fix the cookie/auth logic for --diff-filename=-
The new cookie logic introduced a bug with --diff-filename=. It was
failing when passing --username and --password, due to not being in the
'not' clause. This is now fixed to work if there's a valid cookie or
auth credentials.

Fixes bug #2503
1fa3c44
@chipx86 chipx86 Remove unnecessary properties that broke Python 2.4.
There were some @property instances in SCMClient. Some of these used
@property.setter, which isn't on Python 2.4, and this broke 2.4 support.

These weren't needed. They did nothing but set/get a variable. So,
replace them with the variable itself.

Fixes bug #2504
dc7e214
@chipx86 chipx86 Merge branch 'master' into api 0770853
Something went wrong with that request. Please try again.