1197507 - Mochitest support for WebExtensions

Status: RESOLVED FIXED

(WebExtensions :: Untriaged, defect)

Description

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

I wanted to get be able to use xpcshell tests for WebExtensions, but that turned out to be a lot of work. It's probably best to start with Mochitests since these run on all platforms including mobile. These will test code that currently lives in toolkit/. For the stuff in browser/, we'll probably want to use browser-chrome tests. The code here could perhaps be adapted for that.

Comment 1

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: alias browser.* and chrome.*

I don't want to have to use chrome.* APIs in our tests, so I made this alias.

Comment 2

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: allow HTTP resources in moz-extension: substitutions

I would like to be able to load extensions off HTTP for testing, since all Mochitest files are loaded through HTTP as far as I know. I don't think adding HTTP here should be a problem since it doesn't grant any special privileges to the resource. We won't use this in production code.

Comment 3

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: add support for chrome.test API

This API is similar to an internal one that Chrome has. It allows extensions to communicate with test harness code. I used an event emitter to take care of the actual communication. The other endpoint is in a later patch.

Comment 4

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: allow mochitests to load WebExtensions via special powers

This does what it says.

Comment 5

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: simple tests

This patch has some test examples. I'm going to post a more substantial test in bug 1190330.

Comment 6

[Bobby Holley (:bholley)]

Comment on attachment 8651397 [details] [diff] [review]
allow HTTP resources in moz-extension: substitutions

These URIs claim URI_IS_LOCAL_FILE - isn't that a problem?

We could probably use nsIProtocolHandlerWithDynamicFlags to solve this, but this patch doesn't do that...

Comment 7

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: allow HTTP resources in moz-extension: substitutions v2

OK. I assume you meant URI_IS_LOCAL_RESOURCE since these handlers don't claim URI_IS_LOCAL_FILE. I hope I didn't go overboard here. It's a bit hard to understand all the things this flag is used for.

If you're worried about EnsureFile, I think we're covered there:
http://mxr.mozilla.org/mozilla-central/source/netwerk/base/nsStandardURL.cpp#1823

Comment 8

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: allow HTTP resources in moz-extension: substitutions v3

This makes the default be to not allow the chrome: scheme.

Comment 9

[Gabor Krizsanits (INACTIVE)]

Comment on attachment 8651399 [details] [diff] [review]
allow mochitests to load WebExtensions via special powers

Review of attachment 8651399 [details] [diff] [review]:
-----------------------------------------------------------------

If loading the extension fails, will the reason for the failure be visible somewhere (console, try log)? I guess it will be but I have seen some swallowed up exceptions over the years, so I was wondering. Another thing is that I'm missing some kind of load/startup failed hook, in case something goes wrong to avoid test hangs. I don't want to hold up landing this patch it's already great as it is, we can do a follow-up for that hook if you agree with it or let me know if there is a good reason why such additional hook is not needed.

Comment 10

[Gabor Krizsanits (INACTIVE)]

Comment on attachment 8651400 [details] [diff] [review]
simple tests

Review of attachment 8651400 [details] [diff] [review]:
-----------------------------------------------------------------

::: toolkit/components/extensions/test/mochitest/mochitest.ini
@@ +16,5 @@
> +  file_script_redirect.js
> +  file_script_xhr.js
> +
> +[test_simple_extensions.html]
> +[test_extension_webrequest.html]

this should go to the other patch

Comment 11

[Bobby Holley (:bholley)]

Comment on attachment 8651455 [details] [diff] [review]
allow HTTP resources in moz-extension: substitutions v3

Review of attachment 8651455 [details] [diff] [review]:
-----------------------------------------------------------------

Per discussion, I think we should find a way to solve this problem in the test infrastructure.

Comment 12

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Attachment: use TESTING_JS_MODULES for extensions

Comment 13

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/faa9493e9372
https://hg.mozilla.org/integration/mozilla-inbound/rev/1706813276e7
https://hg.mozilla.org/integration/mozilla-inbound/rev/e31c9124bd22
https://hg.mozilla.org/integration/mozilla-inbound/rev/2b875b229533
https://hg.mozilla.org/integration/mozilla-inbound/rev/91e5acc46643

Comment 14

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Oops. The last patch landed with the wrong bug number. 91e5acc46643 is actually bug 1190330.

Comment 15

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/7d17c55a8323

Comment 16

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/faa9493e9372
https://hg.mozilla.org/mozilla-central/rev/1706813276e7
https://hg.mozilla.org/mozilla-central/rev/e31c9124bd22
https://hg.mozilla.org/mozilla-central/rev/2b875b229533
https://hg.mozilla.org/mozilla-central/rev/91e5acc46643
https://hg.mozilla.org/mozilla-central/rev/7d17c55a8323