Copy VS runtime DLLs to output directory

Copy VS runtime DLLs to output directory

When using the automatic toolchain for vs2013, they won't necessarily
be system-installed, so copy them next to the output binaries.

R=dpranke@chromium.org
BUG=326357
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=248790

[Dirk Pranke, 2014-02-04 20:40:30]

lgtm.

seems ok, but I wish there was some way to make this step be a build target in a
gyp file rather than in gyp_chromium. But, I'm not really seeing a way to do
that easily, given that the source files are in depot_tools and you may not need
it if you're using an installed toolchain.

[scottmg, 2014-02-04 20:54:03]

On 2014/02/04 20:40:30, Dirk Pranke wrote:
> lgtm.
> 
> seems ok, but I wish there was some way to make this step be a build target in
a
> gyp file rather than in gyp_chromium. But, I'm not really seeing a way to do
> that easily, given that the source files are in depot_tools and you may not
need
> it if you're using an installed toolchain.

Hmm. We really only need the paths, so one way could be to generate a .gypi file
at this point.

By default the list of files copied could be empty, and a gypi could replace
them. Trickiness:
- Different dlls per-arch and per-config
- I'm not sure where the 'copies' action dependency would go exactly. These DLLs
are really a dependency of all executable and shared_library targets. It should
really be pre-everything.

[Dirk Pranke, 2014-02-04 20:55:57]

Yeah, I'm not sure what the best way to do this would be; you'd need to somehow
detect it conditionally and add it as a target to base or something. I don't
think I'd be a fan of generating the .gypi, since AFAIK we don't do that
anywhere else.

(which is why I lgtm'ed it; I couldn't think of a great alternative).

[scottmg, 2014-02-04 20:57:18]

On 2014/02/04 20:55:57, Dirk Pranke wrote:
> Yeah, I'm not sure what the best way to do this would be; you'd need to
somehow
> detect it conditionally and add it as a target to base or something. I don't
> think I'd be a fan of generating the .gypi, since AFAIK we don't do that
> anywhere else.
> 
> (which is why I lgtm'ed it; I couldn't think of a great alternative).

OK, thanks.

[scottmg, 2014-02-04 20:57:25]

The CQ bit was checked by scottmg@chromium.org

[scottmg, 2014-02-04 21:53:13]

The CQ bit was unchecked by scottmg@chromium.org

[scottmg, 2014-02-04 22:12:33]

Committed patchset #3 manually as r248790 (presubmit successful).