971279 - Add Java Test Plugin

Status: RESOLVED FIXED

(Core Graveyard :: Plug-ins, defect)

Description

[John Schoenick [:johns]]

Originally in bug 738396, but needed more work. Being able to test all the special-cased stuff for java is a plus.

Comment 1

[John Schoenick [:johns]]

Attachment: Add plugin.java.mime

Originally I had a hack in nsObjectLoadingContent to check for x-java-test for <applet> in debug builds, but that breaks on OS X which has a default java plugin. And since we store the pref for both addons as "plugin.state.java" or whathaveyou disabling one and not the other is broken (and not worth fixing)

We also have "application/x-java-vm" hardcoded in a ton of spots to specify magic java behavior, it seems saner to just store this as a pref. Which then lets tests flip the pref to the java test plugin!

Comment 2

[John Schoenick [:johns]]

Attachment: Add nptest java plugin

This was r+'d in 738396 ages ago, but was bit-rotted to hell, and the old hacks that plugin.java.mime obsoletes removed. Feel free to rubber stamp.

Comment 3

[John Schoenick [:johns]]

(In reply to John Schoenick [:johns] from comment #1)
> Originally I had a hack in nsObjectLoadingContent to check for x-java-test
> for <applet> in debug builds, but that breaks on OS X

It also doesn't allow testing other magic-cases like <object classid="java:">

Comment 4

[John Schoenick [:johns]]

Attachment: [Foldme] Always treat plugin.java.mime as a java type, rename npjavatest -> nptestjava

This hit issues on try due to us blacklisting DLLs beginning with "npjava". I don't want to touch that, so s/npjavatest/nptestjava/

I also just made IsJavaMIME() read and recognize java.plugin.mime, rather than the debug-specific application/x-java-test hack.

Uploading these two changes for additional r? due to messing-with-java-plumbing paranoia

Comment 5

[John Schoenick [:johns]]

Attachment: Add nptest java plugin. r=bsmedberg

Comment 6

[John Schoenick [:johns]]

Attachment: Add plugin.java.mime. r=bsmedberg

Folded for landing

Comment 7

[John Schoenick [:johns]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/34ecdab0cfc2
https://hg.mozilla.org/integration/mozilla-inbound/rev/5028e1290b45

Try:
https://tbpl.mozilla.org/?tree=Try&rev=fd55a944e195

Comment 8

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

I backed this (and everything else from that push to inbound) out in http://hg.mozilla.org/integration/mozilla-inbound/rev/128cbf1edc40 due to various Java/plugin related failures they caused:
Crashtest failure: https://tbpl.mozilla.org/php/getParsedLog.php?id=35003251&tree=Mozilla-Inbound
XPCShell failure: https://tbpl.mozilla.org/php/getParsedLog.php?id=35003355&tree=Mozilla-Inbound

Comment 9

[Georg Fritzsche [:gfritzsche]]

(In reply to Wes Kocher (:KWierso) from comment #8)
> XPCShell failure:
> https://tbpl.mozilla.org/php/getParsedLog.php?id=35003355&tree=Mozilla-
> Inbound

TEST-UNEXPECTED-FAIL | /builds/slave/test/build/tests/xpcshell/tests/services/healthreport/tests/xpcshell/test_provider_addons.js | 3 == 2 - See following stack:

Ouch, we should just check based on the plugin tag count here.

Comment 10

[Georg Fritzsche [:gfritzsche]]

Attachment: Fix FHR addons provider test.

Comment 11

[Georg Fritzsche [:gfritzsche]]

Attachment: Fix FHR addons provider test.

Comment 12

[John Schoenick [:johns]]

Attachment: Add plugin.java.mime. r=bsmedberg

So the assertion failure in a random crash test was CnP fail, the second assert should be nsPluginHost::IsJavaMIME(javaMIME)

Comment 13

[John Schoenick [:johns]]

Attachment: Fix FHR addon provider test for plugin data count.

With the fix we move on to failing slightly lower down. Even more fixage!

Comment 14

[John Schoenick [:johns]]

Attachment: Add nptest java plugin. r=bsmedberg

Another try push later... test_provider_addons.js still fails. Only on OS X. Because the capitalization in Info.plist didn't match nptest.rc. And this test wants the exact plugin description to match. Just in case.

Comment 15

[Gregory Szorc [:gps]]

Comment on attachment 8379964 [details] [diff] [review]
Fix FHR addon provider test for plugin data count.

Review of attachment 8379964 [details] [diff] [review]:
-----------------------------------------------------------------

LGTM.

Comment 16

[John Schoenick [:johns]]

Attempt 2:
https://hg.mozilla.org/integration/mozilla-inbound/rev/b6a4f6e8ef8c
https://hg.mozilla.org/integration/mozilla-inbound/rev/7254cd0a6d75
https://hg.mozilla.org/integration/mozilla-inbound/rev/12b636c82b24

Extensive try run since this push burned the tree before:
https://tbpl.mozilla.org/?tree=Try&rev=b656bdd77ce1

Comment 17

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/b6a4f6e8ef8c
https://hg.mozilla.org/mozilla-central/rev/7254cd0a6d75
https://hg.mozilla.org/mozilla-central/rev/12b636c82b24

Comment 18

[John Schoenick [:johns]]

Followup, this chunk was removed in the comment 4 patch and but dropped when I folded it:
https://hg.mozilla.org/integration/mozilla-inbound/rev/505119ed119e

Comment 19

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/505119ed119e