772700 - OOP from a JS component or JSM doesn't work

Status: RESOLVED FIXED

(Core :: Storage: IndexedDB, defect)

Description

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Need to set up the actor machinery for windowless usage. Currently it crashes.

And then I guess I need to figure out some way to test this combination :-/

Comment 1

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Attachment: Patch, v1

This patch is pretty simple, but the testing is a bit messy.

Anyway, I figured out how to mirror a JSM/JSComponent global so that I don't have to rework all the ownership stuff in the parent. I'm using a BackstagePass with a Null principal (not "no" principal), so if it ever escapes somehow it can't actually do anything. We only need it for the event handling stuff that happens during the open sequence.

Everything seems to work!

Comment 2

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

Comment on attachment 641707 [details] [diff] [review]
Patch, v1

Review of attachment 641707 [details] [diff] [review]:
-----------------------------------------------------------------

I think mrbkap or bholley should review the XPConnect abuse here.

::: dom/indexedDB/IDBFactory.cpp
@@ +182,5 @@
> +IDBFactory::Create(IDBFactory** aFactory)
> +{
> +  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
> +  NS_ASSERTION(IndexedDatabaseManager::IsMainProcess(), "Wrong process!");
> +  NS_ASSERTION(nsContentUtils::IsCallerChrome(), "Only for chrome!");

I think instead of checking that the caller is chrome, you should check that the top JSContext on the stack is null.

Comment 3

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Attachment: Use a sandbox global, v1

Applies on top of the previous patch.

Comment 4

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 642158 [details] [diff] [review]
Use a sandbox global, v1

r=me with a comment about why we need to unwrap there.

Comment 5

[Ben Turner (not reading bugmail, use the needinfo flag!)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/592bfc94e628

Comment 6

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/1f9add03a089

Comment 7

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Somehow the last time we ran the xpcshell tests on mac even though I tried to disable them. This time I've restricted them to just windows (hopefully).

https://hg.mozilla.org/integration/mozilla-inbound/rev/06ba7dd9f841

Comment 8

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://tbpl.mozilla.org/php/getParsedLog.php?id=13527895&tree=Mozilla-Inbound

TEST-INFO | /Users/cltbld/talos-slave/test/build/xpcshell/tests/dom/indexedDB/ipc/unit/test_add_put.js | running test ...

command timed out: 1200 seconds without output, attempting to kill
process killed by signal 9
program finished with exit code -1
elapsedTime=2725.067969
TinderboxPrint: xpcshell<br/><em class="testfail">T-FAIL</em>
buildbot.slave.commands.TimeoutError: command timed out: 1200 seconds without output, attempting to kill
TinderboxPrint: xpcshell<br/><em class="testfail">timeout</em>

Backout, please.

Comment 9

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Yes, they don't work on mac. I've tried two different ways to disable them there, but I'm obviously not getting something right. Anyway, disabled entirely:

https://hg.mozilla.org/integration/mozilla-inbound/rev/6166a8e28786

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/06ba7dd9f841
https://hg.mozilla.org/mozilla-central/rev/6166a8e28786

Comment 11

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Followup to run the tests somewhere, please.