Implement onEvent callback framework for UDP

Wire up onEvent and put it to use when UdpSocket::Write blocks.

This is a checkpoint commit. It does the following:

- if Write() blocks, calls onEvent() when the underlying socket finishes its work.
- Demonstrates a working framework for sending events back to JS.

It does NOT do any of the following:

- Prove that the above statements are true.

In order to cause a UDP socket to block, we pretty much have to mock it, and I determined that I'd have to refactor my Socket class to let me put in a ClientSocketFactory for testing, which will then let me work DelayedSocketData in net/socket/socket_test_util.h into the mix and simulate Write blockage. After discussion with Antony, we agreed to break up this work into this CL and then a refactor/mock CL, which I'm working on now.

BUG=106802
TEST=none (see above).

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=114544

[miket_OOO, 2011-12-13 22:44:49]

http://codereview.chromium.org/8896013/diff/4001/chrome/renderer/resources/ex...
File chrome/renderer/resources/extensions/schema_generated_bindings.js (right):

http://codereview.chromium.org/8896013/diff/4001/chrome/renderer/resources/ex...
chrome/renderer/resources/extensions/schema_generated_bindings.js:27: native
function SendResponseAck(requestId);
I alphabetized this list of functions and added my one new function
(GetNextSocketEventId). It's possible I violated a code of conduct by doing so
and will be happy to revert, but I do feel it makes this block much more
readable and maintainable.

[Aaron Boodman, 2011-12-14 00:28:23]

lgtm

http://codereview.chromium.org/8896013/diff/4001/chrome/renderer/resources/ex...
File chrome/renderer/resources/extensions/schema_generated_bindings.js (right):

http://codereview.chromium.org/8896013/diff/4001/chrome/renderer/resources/ex...
chrome/renderer/resources/extensions/schema_generated_bindings.js:27: native
function SendResponseAck(requestId);
On 2011/12/13 22:44:50, miket wrote:
> I alphabetized this list of functions and added my one new function
> (GetNextSocketEventId). It's possible I violated a code of conduct by doing so
> and will be happy to revert, but I do feel it makes this block much more
> readable and maintainable.

I think what you did is perfect. It's nice to cleanup code, but it's also nice
to note what change you actually made in the case that the cleanup makes it
unclear.

http://codereview.chromium.org/8896013/diff/4001/chrome/renderer/resources/ex...
chrome/renderer/resources/extensions/schema_generated_bindings.js:608: } catch
(e) {
Indent looks weird.

http://codereview.chromium.org/8896013/diff/4001/chrome/renderer/resources/ex...
chrome/renderer/resources/extensions/schema_generated_bindings.js:609: // Safe
to ignore.
That is almost never true :). How about console.error(e).

[asargent_no_longer_on_chrome, 2011-12-14 17:36:17]

LGTM

http://codereview.chromium.org/8896013/diff/8001/chrome/common/extensions/api...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/8896013/diff/8001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:8933: "description":
"writeComplete A 'data' event represents data received on the socket. An 'error'
event represents an error."
nit: looks like you need to update the description here to remove references to
'data' and 'error'.

http://codereview.chromium.org/8896013/diff/8001/chrome/renderer/resources/ex...
File chrome/renderer/resources/extensions/schema_generated_bindings.js (right):

http://codereview.chromium.org/8896013/diff/8001/chrome/renderer/resources/ex...
chrome/renderer/resources/extensions/schema_generated_bindings.js:595:
chrome.experimental.socket.onEvent.addListener(
I vaguely recall that when I added the listeners for context menu events, that I
was always adding them as you are here, and that erikkay noticed when doing some
profiling a little more overhead to this than we expected. So I ended up only
starting to listen when someone actually uses the API - see 
chromeHidden.contextMenus.ensureListenerSetup in this file for how that works.
Also I'll try to dig up the CL where I changed to doing that, to see if I
included a better description of why we decided the change was neeeded.

Anyhow, no big deal for now, but might be worth looking at doing in a future CL
(maybe we can unify the two code paths a bit, and see if any other places use
this same pattern).

[miket_OOO, 2011-12-14 17:58:22]

Fixed typo. I now recall that I saw a trampoline roll by in the middle of that
edit.

Agreed on speed issues, but I'm pretty sure we're going to have to re-examine
this for performance issues in general. I'd like to get something limping along
so people can start building on it.

[commit-bot: I haz the power, 2011-12-14 18:26:06]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/miket@chromium.org/8896013/14001

[commit-bot: I haz the power, 2011-12-14 18:26:09]

Can't process patch for file
chrome/renderer/extensions/schema_generated_bindings.cc.
File's status is None, patchset upload is incomplete.

[commit-bot: I haz the power, 2011-12-14 18:38:37]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/miket@chromium.org/8896013/9009

[commit-bot: I haz the power, 2011-12-14 22:27:55]

Try job failure for 8896013-9009 (retry) on mac_rel for steps "unit_tests,
browser_tests, ui_tests".
It's a second try, previously, steps "unit_tests, browser_tests, ui_tests"
failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[commit-bot: I haz the power, 2011-12-14 22:53:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/miket@chromium.org/8896013/9009

[commit-bot: I haz the power, 2011-12-14 23:16:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/miket@chromium.org/8896013/8012

[miket_OOO, 2011-12-15 00:08:38]

> chrome/renderer/resources/extensions/schema_generated_bindings.js:609: // Safe
> to ignore.
> That is almost never true :). How about console.error(e).

Hey Aaron, on the trybots many tests are now outputting this line:


[10654:10654:1214/154153:3935207438:INFO:CONSOLE(607)] "Error: You do not have
permission to use 'experimental.socket.onEvent'. Be sure to declare in your
manifest what permissions you need.", source:
extensions/schema_generated_bindings.js (607)

I think that this is why other APIs in that file, like TTS, silently ignore this
error. Thoughts?

[commit-bot: I haz the power, 2011-12-15 00:38:48]

Change committed as 114544