summaryrefslogtreecommitdiff
path: root/lint.py
Commit message (Collapse)AuthorAgeFilesLines
...
* recent astroid cset trigger a maybe-no-member message that makes some tests failSylvain Th?nault2014-04-021-1/+1
|
* ensure --init-hooks is evaluated before other options. Closes issue #166Sylvain Th?nault2014-03-211-7/+9
|
* enhance option preprocessing error detection and related testsSylvain Th?nault2014-03-211-7/+12
|
* Updated FSF address.Arun Persaud2014-02-231-1/+1
|
* various pylint fixesSylvain Th?nault2013-12-221-43/+50
|
* Merged logilab/pylint into defaultDaniel Balparda2013-12-041-1/+19
|\
| * Take first steps to use the symbolic name internally instead of the numeric id.Torsten Marek2013-11-061-1/+1
| | | | | | | | | | | | | | This changeset changes utils.py to use the symbolic name as the primary key in _messages, and the numeric ID as an alternative name in a new dictionary. Also prepares the codebase for supporting renamed messages.
| * Allow linter subclasses to override should_analyze_file to skip allTorsten Marek2013-11-051-0/+18
| | | | | | | | processing of a file.
* | Adding a new message for pylint (I0022, deprecated-pragma) that warns on the ↵Daniel Balparda2013-12-041-4/+7
|/ | | | deprecated pragma disable-msg and enable-msg suppressions.
* Always call the reporter's on_close method, even if no file could be ↵Daniel Balparda2013-11-211-0/+2
| | | | analyzed. Some reporters still need to flush their state.
* [lint]?rename get_astroid into get_astSylvain Th?nault2013-10-251-3/+3
|
* fix typo in --msg-template description. Closes issue #67Sylvain Th?nault2013-09-041-1/+1
|
* some pylint and style fixesSylvain Th?nault2013-07-311-2/+2
|
* Templating mecanism for text reporters.Anthony Truchet2013-07-301-14/+10
| | | | | | | | | | | | This changeset introduces a reporters.Message class to represent messages with their attached data. Those messages object can then be formatted accordingto some template. A template can be specified on the command line through the msg-template option. This deprecates parseable and msvc output format, and kill --symbols and --include-ids option.
* update method renamed in astroidSylvain Th?nault2013-07-241-1/+1
|
* Turn reporters into proper plugins rather than classes that are loaded ↵Torsten Marek2013-06-191-25/+34
| | | | explicitly.
* move EmptyReport exception to pylint.utils and refactor imports so ↵Sylvain Th?nault2013-06-191-4/+6
| | | | pylint.utils stop importing checkers/reporters
* Extend the line length check.Torsten Marek2013-06-171-3/+1
|
* The tokenize module rejects certain source files that are accepted asTorsten Marek2013-06-171-1/+5
| | | | | valid by the Python interpreter, e.g. modules that end with a line continuation. Catch the exception and skip the files.
* astng has been renamed astroidDavid Douard2013-06-171-33/+33
|
* old-style classes are deprecated: refactor to new styleNicolas Chauvat2013-05-201-1/+1
|
* since PyLinter._dynamic_plugins is used as a set, make it a setNicolas Chauvat2013-05-201-2/+2
|
* add delta to comparison with note from previous executionSylvain Th?nault2013-05-071-2/+3
| | | | as suggested by S?bastien Celles
* Minor updates for the token checker interface, triggered by the review.Torsten Marek2013-05-031-7/+6
|
* Tokenize the input source only once and hand it to all checkers that need ↵Torsten Marek2013-05-011-11/+17
| | | | | | | | | the token stream. A lot of checkers need access to the token stream, but they all tokenize the source code again in BaseRawChecker.process_module. This change introduces a new checker type ITokenChecker, for which the token stream is created exactly once in PyLinter, and then injected into all registered checkers.
* closes #1 (bitbucket): fix "dictionary changed size during iteration" crashSylvain Th?nault2013-04-161-2/+2
|
* a few pylint fixes and copyright cleanupsSylvain Th?nault2013-03-291-5/+3
|
* Make sure that pragmas that apply to whole lines are interpreted literally, ↵Torsten Marek2013-03-291-6/+12
| | | | | | | | | so that their scope is not extended to the whole scope if they occur at the beginning of a scope. Closes #123285
* Two small fixes for suppression warnings I00{2,1}:tmarek2013-02-201-3/+5
| | | | | - make sure to include the full message name (with symbol, if requested) in the output - flush I002{0,1} warnings after each module, not after complete run
* backport stableSylvain Th?nault2012-11-121-10/+70
|\
| * Add hooks for import path setup and move pylint's sys.pathtmarek2012-11-071-5/+13
| | | | | | | | modifications into them. These hooks can be overridden by other implementations.
| * Add messages I0020 and I0021 for reporting of suppressed messages and ↵tmarek2012-11-121-0/+32
| | | | | | | | | | | | | | useless suppression pragmas. Closes #110840 Both messages are disabled by default, and only emitted after all other checkers have been processed.
| * Change the regular expression for inline options so that a preceeding # is ↵tmarek2012-11-071-1/+1
| | | | | | | | not optional.
| * add support for --disable=all option (closes: #105327)Alexandre Fayolle2012-10-031-4/+24
| | | | | | | | | | updated the documentation accordingly deprecate 'disable-all' as an inline directive in favor of 'skip-file'
* | Fix some typos and grammatical errors in the help textDavid Pursehouse2012-10-251-4/+4
|/
* explicit trailing whitespace, avoid temptation to drop those (thx Martin)Sylvain Th?nault2012-10-051-26/+26
|
* don't want to run a checker only because of a Fatal errorSylvain Th?nault2012-09-201-1/+2
|
* minor tweaks to previous changeset (custom reporter). Closes #105337Sylvain Th?nault2012-09-201-4/+5
|
* Added the ability to specify reporter class name as report's 'output-format'.Kevin Jing Qiu2012-09-201-3/+10
| | | | | | This allows a custom reporter to be used to format reports. Backwards-compatibility is maintained, so a user can still specify the named reporters such as 'text' or 'parseable'.
* lint fixesSylvain Th?nault2012-09-191-7/+7
|
* Closes #104572: symbolic warning names in output (by Martin Pool)Martin Pool2012-09-191-1/+20
| | | | triggered whatever the format using a command line option
* use .iter* version of dict methods when possibleSylvain Th?nault2012-09-191-7/+7
|
* Call reporter.set_output before reporter.on_set_current_module in order to ↵FELD Boris2012-07-171-3/+3
| | | | ease reporter work
* use lgc.compat to access builtins for cross-interpreters compatSylvain Th?nault2012-07-241-4/+3
|
* fix indentationSylvain Th?nault2012-07-131-4/+4
|
* Make pythonpath behaviour closer to python's (closes #88218)JT Olds2012-05-151-1/+16
| | | | | | When pylint is called with a single filename as argument, walk its path backwards to find the first directory that doesn't have a __init__.py, and add that to sys.path instead of the cwd.
* Add 'on_close' event trigerring and event callback for reporters.FELD Boris2012-05-031-0/+1
|
* Refactoring reporting.Sylvain Th?nault2012-05-031-12/+18
|
* Add 'on_set_current_module' event trigerring and event callback for reporters.FELD Boris2012-03-091-0/+1
|
* This patch is an attempt to completely remove the need for all ↵Sylvain Th?nault2011-10-261-62/+67
| | | | | | | | | | | | | | | | | | | | | | Google-internal patches done to pylint. Currently, we have to modify pylint in two places: * Default plugin loading: Due to the way Python applications are deployed, the import magic in checkers.package_load does not work and needs to be replaced. Placing the calls in an overridable methods allows us to modify the behavior in our own linter class * Special reporter: We use a custom reporter. It can be passed to Run using the reporter argument, but creating in the initializer allows for other reporters to be added to REPORTER_OPT_MAP, which is a more general solution and might also benefit others. These changes do not add any value to upstream pylint per se, but are not intrusive and would help us minimize the maintenance burden when upgrading to a new pylint upstream version, freeing up resources for more important work on pylint.
View Lorry Controller Status