Add script to mass add a new #include.

Add script to mass add a new #include.

Useful for things like scoped_ptr<T> => std::unique_ptr<T>.

Usage:
  python add_header.py --header "<memory>" a/b.cc c/d.cc ...

BUG=557422

[dcheng, 2015-12-07 19:21:57]

Description was changed from

==========
Implement a script to mass-add a header file.

Useful for things like scoped_ptr<T> => std::unique_ptr<T>.

Usage:
  python add_header.py --header "<memory>" a/b.cc c/d.cc ...

BUG=557422
==========

to

==========
Add script to mass add a new #include.

Useful for things like scoped_ptr<T> => std::unique_ptr<T>.

Usage:
  python add_header.py --header "<memory>" a/b.cc c/d.cc ...

BUG=557422
==========

[dcheng, 2015-12-07 19:21:57]

dcheng@chromium.org changed reviewers:
+ danakj@chromium.org, thakis@chromium.org

[dcheng, 2015-12-07 19:28:54]

I wrote this... and then found sort-headers.py. There's some overlap (but my
version is tested!):

Things that it looks like I need to better handle:
- Forcing <windows.h> to be first, and other fun implicit dependencies.
- "config.h" in Blink (though the tool as is won't work on Blink anyway, since
Blink doesn't use full paths)
- Apparently you can have bare import with out the #?

On the other hand, this tool handles:
- Sorting the primary header first (this might be handled in sort-headers.py but
I can't really tell where)
- Preamble comments (this can be easily extended to #define, #undef, etc which
may be useful for some crazy X11 stuff):
  // moo
  #include "cow.h"
- Add newlines to separate primary from system from user headers.

[Nico, 2015-12-07 19:31:56]

On 2015/12/07 19:28:54, dcheng wrote:
> I wrote this... and then found sort-headers.py. There's some overlap (but my
> version is tested!):
> 
> Things that it looks like I need to better handle:
> - Forcing <windows.h> to be first, and other fun implicit dependencies.
> - "config.h" in Blink (though the tool as is won't work on Blink anyway, since
> Blink doesn't use full paths)
> - Apparently you can have bare import with out the #?

that's clang's module extension
(http://llvm.org/devmtg/2012-11/Gregor-Modules.pdf for import vs #import etc).
We hopefully don't use that anywhere.

clang-format will have header sorting after the next roll unless we disable
this. It probably also gets much of this wrong.

Could this tool use sort_headers in some form to become simpler?

> 
> On the other hand, this tool handles:
> - Sorting the primary header first (this might be handled in sort-headers.py
but
> I can't really tell where)
> - Preamble comments (this can be easily extended to #define, #undef, etc which
> may be useful for some crazy X11 stuff):
>   // moo
>   #include "cow.h"
> - Add newlines to separate primary from system from user headers.

[danakj, 2015-12-07 19:33:18]

On Mon, Dec 7, 2015 at 11:31 AM, <thakis@chromium.org> wrote:

> On 2015/12/07 19:28:54, dcheng wrote:
>
>> I wrote this... and then found sort-headers.py. There's some overlap (but
>> my
>> version is tested!):
>>
>
> Things that it looks like I need to better handle:
>> - Forcing <windows.h> to be first, and other fun implicit dependencies.
>> - "config.h" in Blink (though the tool as is won't work on Blink anyway,
>> since
>> Blink doesn't use full paths)
>> - Apparently you can have bare import with out the #?
>>
>
> that's clang's module extension
> (http://llvm.org/devmtg/2012-11/Gregor-Modules.pdf for import vs #import
> etc).
> We hopefully don't use that anywhere.
>
> clang-format will have header sorting after the next roll unless we disable
> this. It probably also gets much of this wrong.
>
> Could this tool use sort_headers in some form to become simpler?


That would implicitly test sort_headers too... :)


>
>
>
> On the other hand, this tool handles:
>> - Sorting the primary header first (this might be handled in
>> sort-headers.py
>>
> but
>
>> I can't really tell where)
>> - Preamble comments (this can be easily extended to #define, #undef, etc
>> which
>> may be useful for some crazy X11 stuff):
>>    // moo
>>    #include "cow.h"
>> - Add newlines to separate primary from system from user headers.
>>
>
>
> https://codereview.chromium.org/1505823003/
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[dcheng, 2015-12-07 19:56:30]

I think it would be easier to express sort-headers in terms of this module,
since this script breaks it down into more composable parts.

On Mon, Dec 7, 2015, 11:33 Dana Jansens <danakj@chromium.org> wrote:

> On Mon, Dec 7, 2015 at 11:31 AM, <thakis@chromium.org> wrote:
>
>> On 2015/12/07 19:28:54, dcheng wrote:
>>
>>> I wrote this... and then found sort-headers.py. There's some overlap
>>> (but my
>>> version is tested!):
>>>
>>
>> Things that it looks like I need to better handle:
>>> - Forcing <windows.h> to be first, and other fun implicit dependencies.
>>> - "config.h" in Blink (though the tool as is won't work on Blink anyway,
>>> since
>>> Blink doesn't use full paths)
>>> - Apparently you can have bare import with out the #?
>>>
>>
>> that's clang's module extension
>> (http://llvm.org/devmtg/2012-11/Gregor-Modules.pdf for import vs #import
>> etc).
>> We hopefully don't use that anywhere.
>>
>> clang-format will have header sorting after the next roll unless we
>> disable
>> this. It probably also gets much of this wrong.
>>
>> Could this tool use sort_headers in some form to become simpler?
>
>
> That would implicitly test sort_headers too... :)
>
>
>>
>>
>>
>> On the other hand, this tool handles:
>>> - Sorting the primary header first (this might be handled in
>>> sort-headers.py
>>>
>> but
>>
>>> I can't really tell where)
>>> - Preamble comments (this can be easily extended to #define, #undef, etc
>>> which
>>> may be useful for some crazy X11 stuff):
>>>    // moo
>>>    #include "cow.h"
>>> - Add newlines to separate primary from system from user headers.
>>>
>>
>>
>> https://codereview.chromium.org/1505823003/
>>
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Nico, 2015-12-07 19:57:01]

That's fine too; in any case we shouldn't have to very similar scripts
checked in

On Mon, Dec 7, 2015 at 2:56 PM, Daniel Cheng <dcheng@chromium.org> wrote:

> I think it would be easier to express sort-headers in terms of this
> module, since this script breaks it down into more composable parts.
>
> On Mon, Dec 7, 2015, 11:33 Dana Jansens <danakj@chromium.org> wrote:
>
>> On Mon, Dec 7, 2015 at 11:31 AM, <thakis@chromium.org> wrote:
>>
>>> On 2015/12/07 19:28:54, dcheng wrote:
>>>
>>>> I wrote this... and then found sort-headers.py. There's some overlap
>>>> (but my
>>>> version is tested!):
>>>>
>>>
>>> Things that it looks like I need to better handle:
>>>> - Forcing <windows.h> to be first, and other fun implicit dependencies.
>>>> - "config.h" in Blink (though the tool as is won't work on Blink
>>>> anyway, since
>>>> Blink doesn't use full paths)
>>>> - Apparently you can have bare import with out the #?
>>>>
>>>
>>> that's clang's module extension
>>> (http://llvm.org/devmtg/2012-11/Gregor-Modules.pdf for import vs
>>> #import etc).
>>> We hopefully don't use that anywhere.
>>>
>>> clang-format will have header sorting after the next roll unless we
>>> disable
>>> this. It probably also gets much of this wrong.
>>>
>>> Could this tool use sort_headers in some form to become simpler?
>>
>>
>> That would implicitly test sort_headers too... :)
>>
>>
>>>
>>>
>>>
>>> On the other hand, this tool handles:
>>>> - Sorting the primary header first (this might be handled in
>>>> sort-headers.py
>>>>
>>> but
>>>
>>>> I can't really tell where)
>>>> - Preamble comments (this can be easily extended to #define, #undef,
>>>> etc which
>>>> may be useful for some crazy X11 stuff):
>>>>    // moo
>>>>    #include "cow.h"
>>>> - Add newlines to separate primary from system from user headers.
>>>>
>>>
>>>
>>> https://codereview.chromium.org/1505823003/
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.