Add net_unittests to the GN build.

Add net_unittests to the GN build.

BUG=
R=rsleevi@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=275186

[brettw, 2014-06-04 21:26:39]

Currently this crashes with a problem in V8.

[Ryan Sleevi, 2014-06-05 01:09:32]

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn
File net/BUILD.gn (left):

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#oldcode94
net/BUILD.gn:94:
"//net/base/registry_controlled_domains:registry_controlled_domains",
Could you explain a little more about this change, for my own understanding?

Is this equivalent to *? Or is this equivalent to "target whose name matches the
basename of the path" ?

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn
File net/BUILD.gn (right):

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#newcode1110
net/BUILD.gn:1110: deps += [ "//third_party/openssl" ]
Unnecessary - line 1049 covers this.

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#newcode1118
net/BUILD.gn:1118: deps += [ "//crypto:platform" ]
Unnecessary - line 1049 covers this.

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#newcode1145
net/BUILD.gn:1145: if (!enable_websockets) {
I feel like maybe we should just make this os_ios and get rid of this extra
flag. It's not like we want to support websocket enabling/disabling, it's just
an artifact of iOS.

I'm fine with keeping as-is for consistency, though same comments apply for
other pseudo-flags.

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#newcode1205
net/BUILD.gn:1205: # instead of setting use_v8_in_net to 0.
Comment needs updating ("to run gyp on Android" ?)

[brettw, 2014-06-05 17:29:55]

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn
File net/BUILD.gn (left):

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#oldcode94
net/BUILD.gn:94:
"//net/base/registry_controlled_domains:registry_controlled_domains",
On 2014/06/05 01:09:32, Ryan Sleevi wrote:
> Could you explain a little more about this change, for my own understanding?
> 
> Is this equivalent to *? Or is this equivalent to "target whose name matches
the
> basename of the path" ?

The latter

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn
File net/BUILD.gn (right):

https://codereview.chromium.org/311193004/diff/1/net/BUILD.gn#newcode1145
net/BUILD.gn:1145: if (!enable_websockets) {
On 2014/06/05 01:09:32, Ryan Sleevi wrote:
> I feel like maybe we should just make this os_ios and get rid of this extra
> flag. It's not like we want to support websocket enabling/disabling, it's just
> an artifact of iOS.
> 
> I'm fine with keeping as-is for consistency, though same comments apply for
> other pseudo-flags.

I kind of like the pseudo flags in some cases since it makes it more clear why
some things are being removed. I also want to keep this similar-looking to the
GYP, as you mentioned.

[Ryan Sleevi, 2014-06-05 17:58:05]

LGTM, mod the .gypi changes

https://codereview.chromium.org/311193004/diff/20001/net/net.gypi
File net/net.gypi (right):

https://codereview.chromium.org/311193004/diff/20001/net/net.gypi#newcode918
net/net.gypi:918: 'quic/quic_types.h',
Little curious about these changes. Out of date checkout?

[brettw, 2014-06-05 18:04:50]

https://codereview.chromium.org/311193004/diff/20001/net/net.gypi
File net/net.gypi (right):

https://codereview.chromium.org/311193004/diff/20001/net/net.gypi#newcode918
net/net.gypi:918: 'quic/quic_types.h',
On 2014/06/05 17:58:04, Ryan Sleevi wrote:
> Little curious about these changes. Out of date checkout?

I did a merge.

[brettw, 2014-06-05 18:05:20]

Committed patchset #2 manually as r275186 (presubmit successful).