Generate param traits even when shared by mutliple message files.

Generate param traits even when shared by mutliple message files.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=144082

[Tom Sepez, 2012-06-20 19:42:37]

Hey John,

  Please check this out.  I think this works to solve the generation of common
traits issue we keep running into.

[Tom Sepez, 2012-06-20 21:20:48]

PS#2 is same idea, but removes burden from message file includers.

[Tom Sepez, 2012-06-20 21:43:05]

PS#3 takes the same idea, and provides a "starter kit" for content/

[jam, 2012-06-20 21:55:48]

http://codereview.chromium.org/10577040/diff/11001/content/public/common/comm...
File content/public/common/common_param_traits.h (right):

http://codereview.chromium.org/10577040/diff/11001/content/public/common/comm...
content/public/common/common_param_traits.h:31: // Explicitly guarded include of
_defs.h files in header files.
even though I see why, it's odd reading that ifdef in  traits/message files.

how many traits does this cover? i guess you'd want to get rid of
common_param_traits in content/chrome completely?

[Tom Sepez, 2012-06-20 22:03:38]

I'd like to get rid of the case of manual generation just because two or more
message files depend upon it.  Looking at the _traits.cc file, there are only a
few of these so the effort is questionable.

[Tom Sepez, 2012-06-20 22:06:36]

I noticed this while reviewing a CL where a new enum value was added, and the
author dutifully tracked down all the places the enum was used, including the
hand-rolled logging function.  Would be great if it didn't even exist in the
first place to muck with it.

[Tom Sepez, 2012-06-22 22:40:02]

> even though I see why, it's odd reading that ifdef in  traits/message files.

Ok, how about guards but no "#pragma once" in _defs.h followed by an
#undef MUBMLE_MUMBLE_DEFS_H_ in the .cc generator?

[jam, 2012-06-25 19:52:24]

cool, this looks nice. i always didn't like having to do manual serialization of
a struct because it was used in multiple files

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
File chrome/common/common_param_traits.h (right):

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
chrome/common/common_param_traits.h:5: #ifndef
CHROME_COMMON_COMMON_PARAM_TRAITS_H_
nit: in this file and the one in content, please add a comment explaining what
the purpose of this file is

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
File chrome/common/common_param_traits_macros.h (right):

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
chrome/common/common_param_traits_macros.h:5: // Singly or multiply-included
shared traits file depending upon circumstances.
ditto re comment, please explain that this file allows reuse of IPC
serialization macros by multiple message files

http://codereview.chromium.org/10577040/diff/17001/content/public/common/comm...
File content/public/common/common_param_traits.h (right):

http://codereview.chromium.org/10577040/diff/17001/content/public/common/comm...
content/public/common/common_param_traits.h:228: struct
SimilarTypeTraits<WebKit::WebReferrerPolicy> {
nit: these ones too?

[Tom Sepez, 2012-06-25 20:30:24]

John, please review.

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
File chrome/common/common_param_traits.h (right):

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
chrome/common/common_param_traits.h:5: #ifndef
CHROME_COMMON_COMMON_PARAM_TRAITS_H_
Added comment to chrome/ file; one in content/ already has a comment.

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
File chrome/common/common_param_traits_macros.h (right):

http://codereview.chromium.org/10577040/diff/17001/chrome/common/common_param...
chrome/common/common_param_traits_macros.h:5: // Singly or multiply-included
shared traits file depending upon circumstances.
On 2012/06/25 19:52:24, John Abd-El-Malek wrote:
> ditto re comment, please explain that this file allows reuse of IPC
> serialization macros by multiple message files

Done.

http://codereview.chromium.org/10577040/diff/17001/content/public/common/comm...
File content/public/common/common_param_traits.h (right):

http://codereview.chromium.org/10577040/diff/17001/content/public/common/comm...
content/public/common/common_param_traits.h:228: struct
SimilarTypeTraits<WebKit::WebReferrerPolicy> {
On 2012/06/25 19:52:24, John Abd-El-Malek wrote:
> nit: these ones too?

Converted these 3 to IPC_ENUM_TRAITS macros.  Note this kicked loose a redefn
conflict in drag_messages.

[jam, 2012-06-25 20:32:55]

lgtm

[commit-bot: I haz the power, 2012-06-25 22:49:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tsepez@chromium.org/10577040/26004

[commit-bot: I haz the power, 2012-06-26 01:01:44]

Change committed as 144082