Add CrOS's dbus-c++ to Chromium Linux. This is needed to build mtpd.

Add CrOS's dbus-c++ to Chromium Linux. This is needed to build mtpd.

BUG=151685

[Lei Zhang, 2012-10-16 02:55:11]

Pawel: PTAL at tools/checklicenses/checklicenses.py

rsleevi: This carries over from https://gerrit.chromium.org/gerrit/#/c/34598/.
Turns out I can't get away from this.

There's some minor changes to the "dbus_cplusplus" target:
- shared_library since it's LGPL. Don't want anyone to accidentally link it into
a non-license-compatible binary.
- Since it's a .so, I need to not use -fvisibility=hidden.
- Got rid of the "dbusxx_xml2cpp" target. I'm going to check in a pre-generated
copy of the header. Otherwise I'd have to deal with Ubuntu Lucid having a much
older version of libctemplate.

[Paweł Hajdan Jr., 2012-10-16 03:34:57]

+satorux, I think I remember you working on the dbus part. Wasn't the idea to
drop dependency of libdbus, and use our in-house dbus code? Just checking, FYI.

LGTM with comments, thanks for doing the license check.

nit: Please file upstream bugs about the files that don't have license headers
and link to those bugs in your change.

Also, could you please add an option to use system dbus-c++ ? Feel free to defer
that one to a future CL, but in that case add a TODO in the gyp file.

[satorux1, 2012-10-16 03:52:38]

On 2012/10/16 03:34:57, Paweł Hajdan Jr. wrote:
> +satorux, I think I remember you working on the dbus part. Wasn't the idea to
> drop dependency of libdbus, and use our in-house dbus code? Just checking,
FYI.

The idea was to drop the dependency to libdbus-glib (a wrapper library for
libdbus).

Adding a dependency to chromeos dbus-c++ is rather unfortunate, but unavoidable
for building MTPD, hence I'm fine with that.


> 
> LGTM with comments, thanks for doing the license check.
> 
> nit: Please file upstream bugs about the files that don't have license headers
> and link to those bugs in your change.
> 
> Also, could you please add an option to use system dbus-c++ ? Feel free to
defer
> that one to a future CL, but in that case add a TODO in the gyp file.

[Lei Zhang, 2012-10-16 04:10:27]

On 2012/10/16 03:34:57, Paweł Hajdan Jr. wrote:
> +satorux, I think I remember you working on the dbus part. Wasn't the idea to
> drop dependency of libdbus, and use our in-house dbus code? Just checking,
FYI.
> 
> LGTM with comments, thanks for doing the license check.
> 
> nit: Please file upstream bugs about the files that don't have license headers
> and link to those bugs in your change.
> 
> Also, could you please add an option to use system dbus-c++ ? Feel free to
defer
> that one to a future CL, but in that case add a TODO in the gyp file.

The Ubuntu Lucid dbus-c++ and the forked dbus-c++ creates different class
signatures in the generated header. Trying to maintain mtpd to work with both
dbus-c++ libraries is a pain.

[Lei Zhang, 2012-10-16 04:17:59]

On 2012/10/16 03:34:57, Paweł Hajdan Jr. wrote:
> nit: Please file upstream bugs about the files that don't have license headers
> and link to those bugs in your change.

Done

> Also, could you please add an option to use system dbus-c++ ? Feel free to
defer
> that one to a future CL, but in that case add a TODO in the gyp file.

Added a TODO for the future. (Writing the .gyp part is easy. It's the dbus-c++
incompatibilities that's hard to deal with.)

[Paweł Hajdan Jr., 2012-10-16 17:03:49]

On 2012/10/16 04:10:27, Lei Zhang wrote:
> The Ubuntu Lucid dbus-c++ and the forked dbus-c++ creates different class
> signatures in the generated header. Trying to maintain mtpd to work with both
> dbus-c++ libraries is a pain.

Right, if possible those changes should be upstreamed. I think the cost of
forking is underestimated in the log term. Hey, I can even help in upstreaming,
feel free to contact me.

[Ryan Sleevi, 2012-10-16 17:49:44]

Is depending on glib going to cause problems for Aura/Ash on Linux? I thought
the move was to get away from those dependencies.

Since I already reviewed this on the CrOS side, the GYP changes LGTM, but +erg
to double check my memory on the expected glib dependencies.

https://codereview.chromium.org/11148035/diff/3002/DEPS
File DEPS (right):

https://codereview.chromium.org/11148035/diff/3002/DEPS#newcode528
DEPS:528: "src/third_party/cros_dbus_cplusplus/source":
Perhaps clarify the intended platforms (see line 532)

[Elliot Glaysher, 2012-10-16 18:05:28]

On 2012/10/16 17:49:44, Ryan Sleevi wrote:
> Is depending on glib going to cause problems for Aura/Ash on Linux? I thought
> the move was to get away from those dependencies.

Realistically, while I want to break the hard, link dependency on gtk so
linux_aura can work without it, we'll never be rid of a hard glib dependency in
the linux port.

The base::MessagePump implementation in aura builds use low level glib message
loop. I still want to be able to dynamically open the libgtk.so so we can get
native file dialogs plus some themeing information (which needs a glib message
loop). I want to be able to dynamically open libunity.so (which needs gobject
which lives in glib). Etc.

[Lei Zhang, 2012-10-16 18:34:46]

On 2012/10/16 17:49:44, Ryan Sleevi wrote:
> Is depending on glib going to cause problems for Aura/Ash on Linux? I thought
> the move was to get away from those dependencies.

This is for mtpd. It's a separate executable from the chrome binary.

[Lei Zhang, 2012-10-16 18:53:17]

https://chromiumcodereview.appspot.com/11148035/diff/3002/DEPS
File DEPS (right):

https://chromiumcodereview.appspot.com/11148035/diff/3002/DEPS#newcode528
DEPS:528: "src/third_party/cros_dbus_cplusplus/source":
On 2012/10/16 17:49:44, Ryan Sleevi wrote:
> Perhaps clarify the intended platforms (see line 532)

Done.

[commit-bot: I haz the power, 2012-10-16 19:58:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thestig@chromium.org/11148035/12001

[Lei Zhang, 2012-10-16 20:03:34]

On 2012/10/16 17:03:49, Paweł Hajdan Jr. wrote:
> On 2012/10/16 04:10:27, Lei Zhang wrote:
> > The Ubuntu Lucid dbus-c++ and the forked dbus-c++ creates different class
> > signatures in the generated header. Trying to maintain mtpd to work with
both
> > dbus-c++ libraries is a pain.
> 
> Right, if possible those changes should be upstreamed. I think the cost of
> forking is underestimated in the log term. Hey, I can even help in
upstreaming,
> feel free to contact me.

Someone working on CrOS forked this several years ago. I don't know the entire
story behind it. Just trying to use it in Chromium right now.

[commit-bot: I haz the power, 2012-10-16 20:18:41]

Sorry for I got bad news for ya.
Compile failed with a clobber build.
Your code is likely broken or HEAD is junk. Please ensure your
code is not broken then alert the build sheriffs.
Look at the try server FAQ for more details.

[Lei Zhang, 2012-10-16 20:58:23]

On 2012/10/16 20:18:41, I haz the power (commit-bot) wrote:
> Sorry for I got bad news for ya.
> Compile failed with a clobber build.
> Your code is likely broken or HEAD is junk. Please ensure your
> code is not broken then alert the build sheriffs.
> Look at the try server FAQ for more details.

I was missing a couple #defines that would have normally been in the config.h
file.

[commit-bot: I haz the power, 2012-10-16 20:58:31]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thestig@chromium.org/11148035/10002

[commit-bot: I haz the power, 2012-10-17 11:16:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thestig@chromium.org/11148035/10002

[commit-bot: I haz the power, 2012-10-17 13:09:38]

Change committed as 162373