1040966 - framework failures running tests on unusual chars in the output

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Randell Jesup [:jesup] (needinfo me)]

Attachment: bar

I just merged forward (since yesterday?) and running a test for leaks I see this: UnicodeEncodeError: 'ascii' codec can't encode character u'\ufffd' in position 107: ordinal not in range(128)

Doing ./mach test  content/media/test/test_fake_h264.html

Testing bug 1037125 plus bug 1040345 (in the middle of trying to merge the changes and run the tests).

Note: even disabling the fixes code I'm seeing problems.

Comment 1

[(not currently active) Ted Mielczarek]

If I print the message it looks like:
520455936[7fd60c18b4a0]: [GSM Task|tnp] ui.c:654: SIPCC-UI_API: 1/2, ui_set_call_status: the stat string =�

...so I think some logging code is spewing out bad strings, but we should sanitize them instead of choking.

Comment 2

[(not currently active) Ted Mielczarek]

Attachment: Fix unicode encoding when passing bad unicode to stack fixer

This is sort of terrible but it works around the issue.

Comment 3

[Andrew Halberstadt [:ahal]]

Comment on attachment 8460428 [details] [diff] [review]
Fix unicode encoding when passing bad unicode to stack fixer

Review of attachment 8460428 [details] [diff] [review]:
-----------------------------------------------------------------

Yuck, if it works though then good enough for me.

Comment 4

[Gregory Szorc [:gps]]

Comment on attachment 8460428 [details] [diff] [review]
Fix unicode encoding when passing bad unicode to stack fixer

Review of attachment 8460428 [details] [diff] [review]:
-----------------------------------------------------------------

::: testing/mochitest/runtests.py
@@ +1917,5 @@
>        return message
>  
>      def fix_stack(self, message):
>        if message['action'] == 'log' and self.stackFixerFunction:
> +        message['message'] = self.stackFixerFunction(message['message'].encode('ascii', 'replace'))

y u no utf-8?

Comment 5

[(not currently active) Ted Mielczarek]

That just kicks the can down the road and we hit an error later. We can leave this open and fix it right, but this unbreaks developers right now.

Comment 7

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/305e0ca907a8

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/305e0ca907a8

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/48047fdbe126