Mark types used in template instantiatious explicitly as visible

Closed without committing. Turns out this wasn't needed, see comments below.


Mark types used in template instantiatious explicitly as visible

As of r175326, clang hides instantiated templates if any of the instantiating
types are hidden at instantiation time (see
http://comments.gmane.org/gmane.comp.compilers.clang.scm/67619 ). This change
is needed to keep the components build working with clang when rolling to
clang r175326 or newer.

BUG=177235

[hans, 2013-02-27 14:46:59]

lgtm

[danakj, 2013-02-27 14:51:10]

owners LGTM

[commit-bot: I haz the power, 2013-02-27 14:52:17]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thakis@chromium.org/12328142/1

[commit-bot: I haz the power, 2013-02-27 15:15:43]

Sorry for I got bad news for ya.
Compile failed with a clobber build on linux_chromeos.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_chro...
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.

[Nico, 2013-02-28 15:35:02]

This needs some more tweaking to make it work with gcc too, see patch set 3/4.

Mark: base/OWNERS feedback please. It's a bit gnarly, but apparently the best we
can do.

[Mark Mentovai, 2013-02-28 15:39:21]

Since compliers are treating this differently, can we just not use forward
declarations? This evidently doesn’t happen all that often.

https://codereview.chromium.org/12328142/diff/20001/base/export_template_type.h
File base/export_template_type.h (right):

https://codereview.chromium.org/12328142/diff/20001/base/export_template_type...
base/export_template_type.h:1: // Copyright (c) 2012 The Chromium Authors. All
rights reserved.
2013

https://codereview.chromium.org/12328142/diff/20001/base/export_template_type...
base/export_template_type.h:6: #define EXPORT_TEMPLATE_TYPE_H_
This is not the same name that you check for on the line above.

[Nico, 2013-02-28 16:50:25]

On 2013/02/28 15:39:21, Mark Mentovai wrote:
> Since compliers are treating this differently, can we just not use forward
> declarations? This evidently doesn’t happen all that often.

Great point. As it turns out, this CL isn't necessary at all since clang
r176164, for the reason you state: By the time the actual explicit instantiation
in the cc file is reached (as opposed to the instantiation declaration in the
header), the header for Insets / InsetsF / Layer / LayerImpl etc is already
loaded, and due to clang r176164 everything works. I just didn't test building
without any local changes after that revision, due to so much back and forth on
that thread.

So it's all good as is, without this CL. Closing. Thank you for your comment :-)