[PATCH] Clang-format support for removing space before assignment operators

[Daniel Jasper]

Submitted as r191375.


On Wed, Sep 25, 2013 at 5:08 PM, Daniel Jasper <djasper at google.com> wrote:

> Sorry, got side-tracked and forget. Will do this now.
>
>
> On Tue, Sep 24, 2013 at 5:49 PM, Aaron Wishnick <
> aaron.s.wishnick at gmail.com> wrote:
>
>> Ping!
>>
>>
>> On Fri, Sep 13, 2013 at 11:23 AM, Aaron Wishnick <
>> aaron.s.wishnick at gmail.com> wrote:
>>
>>> Thanks. I didn't know about the auto-generation of the docs; that's
>>> pretty slick. Sorry about the tabs vs. spaces; I've corrected this in my
>>> editor for the future. I don't have commit access, so if you don't mind
>>> fixing the style nits, I really appreciate it.
>>>
>>>
>>> On Fri, Sep 13, 2013 at 11:19 AM, Daniel Jasper <djasper at google.com>wrote:
>>>
>>>> The few nits below aside, this looks good to go in. Do you have commit
>>>> access or should I commit this for you? In the latter case, you don't need
>>>> to worry about the style nits (I'll fix them before committing).
>>>>
>>>> +**SpaceBeforeAssignmentOperators** (``bool``)
>>>> +  If ``false``, spaces will be removed before '=', '+=', etc.
>>>> +
>>>>
>>>> Not, that these are actually auto-generated using a python script. It
>>>> doesn't hurt like this, but it will be eventually overwritten by what is in
>>>> the flags comment.
>>>>
>>>> +               R.SpaceAfterControlStatementKeyword &&
>>>> +   SpaceBeforeAssignmentOperators ==
>>>> +       R.SpaceBeforeAssignmentOperators;
>>>>
>>>> These seem to be indented with tabs. Use spaces instead.
>>>>
>>>> +  if (!Style.SpaceBeforeAssignmentOperators &&
>>>> +  Tok.getPrecedence() == prec::Assignment)
>>>> +    return false;
>>>>
>>>> Same as above.
>>>>
>>>> +  FormatStyle Spaces = getLLVMStyle();
>>>> +
>>>> +  verifyFormat("int a = 5;", Spaces);
>>>>
>>>> There seem to be unnecessary spaces in the empty line.
>>>>
>>>>
>>>> On Fri, Sep 13, 2013 at 4:12 PM, Aaron Wishnick <
>>>> aaron.s.wishnick at gmail.com> wrote:
>>>>
>>>>> Thanks for the feedback! I've addressed those three issues, and the
>>>>> updated patch is attached.
>>>>>
>>>>>
>>>>> On Fri, Sep 13, 2013 at 4:12 AM, Daniel Jasper <djasper at google.com>wrote:
>>>>>
>>>>>> You should be able to use "Tok.getPrecedence() == prec::Assignment"
>>>>>> instead of the new IsTokenAssignment function. Also, please put the tests
>>>>>> into "unittests/Format/FormatTest.cpp". We have plans to turn (some of)
>>>>>> those into file-based tests, but for now, I'd like to keep
>>>>>> formatting-specific tests together. Also, in addition to actually testing
>>>>>> whether the style-configuration works as expected, there are also tests
>>>>>> there which test whether the option can be successfully parsed from JSON
>>>>>> (see how it is done for the other flags close to the bottom).
>>>>>>
>>>>>>
>>>>>> On Fri, Sep 13, 2013 at 12:14 AM, Aaron Wishnick <
>>>>>> aaron.s.wishnick at gmail.com> wrote:
>>>>>>
>>>>>>> My organization's coding standards include not putting a space
>>>>>>> before assignment operators, e.g. "int a= 42;". In order for us to be able
>>>>>>> to use clang-format, we need support for this.
>>>>>>>
>>>>>>> I've attached a patch that adds a new format style option,
>>>>>>> "SpaceBeforeAssignmentOperators." All the existing hard-coded styles now
>>>>>>> default to having it set to true. I've also included a test.
>>>>>>>
>>>>>>> Please let me know if this looks good, or if there's anything else I
>>>>>>> can do.
>>>>>>>
>>>>>>> Thank you!
>>>>>>> Aaron
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> cfe-commits mailing list
>>>>>>> cfe-commits at cs.uiuc.edu
>>>>>>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130925/02801044/attachment.html>