Move RFHM browser tests to the embedded test server.

Move RFHM browser tests to the embedded test server.

Using the spawned test server leads to flaky tests on Android, as test
server startup in a separate process takes up quite a bit of time. This
CL moves most tests away from this setup to use the embedded test server,
which runs in-process and doesn't have such a startup cost.

BUG=357747, 418236
Committed: https://crrev.com/555e2bb727ad24696f7c5eb8e037b86758c898a6
Cr-Commit-Position: refs/heads/master@{#351910}

[nasko, 2015-10-01 17:23:48]

nasko@chromium.org changed reviewers:
+ creis@chromium.org

[nasko, 2015-10-01 17:23:49]

Hey Charlie,
Android has finally broken me and I've rewritten most test to use the embedded
test server. Can you review this CL for me?
Thanks in advance!
Nasko

[Charlie Reis, 2015-10-01 17:43:47]

This will be great-- thanks!

https://codereview.chromium.org/1378403002/diff/1/content/test/data/click-nor...
File content/test/data/click-noreferrer-links.html (right):

https://codereview.chromium.org/1378403002/diff/1/content/test/data/click-nor...
content/test/data/click-noreferrer-links.html:5: function setBaseURL(baseUrl) {
I'm torn on this name.  It sounds like it's about setting the page's <base> URL,
which is for relative links.

This is more about replacing all but the path on 3 specific links.  Maybe
something like setOriginForLinks would convey that more clearly?

https://codereview.chromium.org/1378403002/diff/1/content/test/data/click-nor...
content/test/data/click-noreferrer-links.html:118: <a
href="http://REPLACE_WITH_HOST_AND_PORT/title2.html" id="noref_link"
Why keep the REPLACE_WITH_HOST_AND_PORT part if we can't use this with the
original server?  (Removing files/ would break that case, right?)  If you only
specify the path here, you can just append this value to the origin in the
function above.

Also, I noticed this file is used by WebContentsImplBrowserTest.NewNamedWindow. 
It looks like that test is squeaking by because it doesn't use these three
links, but that's risky.  Perhaps we need a warning comment at the top of this
file saying why/when the link rewriting is necessary.

[nasko, 2015-10-01 18:10:18]

https://codereview.chromium.org/1378403002/diff/1/content/test/data/click-nor...
File content/test/data/click-noreferrer-links.html (right):

https://codereview.chromium.org/1378403002/diff/1/content/test/data/click-nor...
content/test/data/click-noreferrer-links.html:5: function setBaseURL(baseUrl) {
On 2015/10/01 17:43:46, Charlie Reis wrote:
> I'm torn on this name.  It sounds like it's about setting the page's <base>
URL,
> which is for relative links.
> 
> This is more about replacing all but the path on 3 specific links.  Maybe
> something like setOriginForLinks would convey that more clearly?

Done.

https://codereview.chromium.org/1378403002/diff/1/content/test/data/click-nor...
content/test/data/click-noreferrer-links.html:118: <a
href="http://REPLACE_WITH_HOST_AND_PORT/title2.html" id="noref_link"
On 2015/10/01 17:43:46, Charlie Reis wrote:
> Why keep the REPLACE_WITH_HOST_AND_PORT part if we can't use this with the
> original server?  (Removing files/ would break that case, right?)  If you only
> specify the path here, you can just append this value to the origin in the
> function above.

It helps visually see which links are expected to be replaced.

> Also, I noticed this file is used by
WebContentsImplBrowserTest.NewNamedWindow. 
> It looks like that test is squeaking by because it doesn't use these three
> links, but that's risky.  Perhaps we need a warning comment at the top of this
> file saying why/when the link rewriting is necessary.

I've put a comment around those links and consolidated them together. Hopefully
this is better.

[Charlie Reis, 2015-10-01 18:12:14]

Thanks, works for me.  LGTM.

[nasko, 2015-10-01 18:12:38]

The CQ bit was checked by nasko@chromium.org

[commit-bot: I haz the power, 2015-10-01 18:13:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1378403002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1378403002/20001

[commit-bot: I haz the power, 2015-10-01 22:45:48]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-10-01 22:46:36]

Patchset 2 (id:??) landed as
https://crrev.com/555e2bb727ad24696f7c5eb8e037b86758c898a6
Cr-Commit-Position: refs/heads/master@{#351910}