1132044 - Coloured and/or colored output from mochitests is broken

Status: RESOLVED WORKSFORME

(Testing :: Mochitest, defect)

Description

[:Gijs (he/him)]

STR:

1. current trunk, OSX (will be fixed by bug 1132031, if it has been, just find some other failing test)
2. in browser/app/profile/firefox.js, change browser.preferences.inContent to always be false
3. ./mach mochitest-browser browser/components/preferences/in-content/tests/browser_bug410900.js
4. Wait for the test to time out (pass --timeout 2 in step 3 if you're in a rush and don't want to)

ER:
see at a glance that the last 3 lines are failure messages

AR:
bunch of white text on my black-background console, especially after spam of an entire directory, hard to notice anything amiss.

I don't know when this broke but it's annoying. Can we please have it back?

Comment 1

[:Gijs (he/him)]

(In reply to :Gijs Kruitbosch from comment #0)
> STR:
> 
> 1. current trunk, OSX (will be fixed by bug 1132031, if it has been, just
> find some other failing test)
> 2. in browser/app/profile/firefox.js, change browser.preferences.inContent
> to always be false
> 3. ./mach mochitest-browser
> browser/components/preferences/in-content/tests/browser_bug410900.js

2b.: ./mach build browser/app

(sorry)

Comment 2

[Chris Manchester (limited bugmail, email directly)]

Does appending "--log-mach -" to the command line invoking mochitest-browser provide suitable colo(u)rization?

Comment 3

[Matthew N. [:MattN]]

(In reply to Chris Manchester [:chmanchester] from comment #2)
> Does appending "--log-mach -" to the command line invoking mochitest-browser
> provide suitable colo(u)rization?

Yes. I assume that's supposed to be the default when run via mach but something broke it?

Comment 4

[Matthew N. [:MattN]]

(In reply to Matthew N. [:MattN] from comment #3)
> (In reply to Chris Manchester [:chmanchester] from comment #2)
> > Does appending "--log-mach -" to the command line invoking mochitest-browser
> > provide suitable colo(u)rization?
> 
> Yes. I assume that's supposed to be the default when run via mach but
> something broke it?

Comment 5

[Chris Manchester (limited bugmail, email directly)]

I don't know. This may be able to be specified as a default in a .machrc file locally now that we have those. Changing the default would probably be fine.

Comment 6

[Andrew Halberstadt [:ahal]]

Matt, so this is still happening for you? The default should be changed already by bug 1421799. If you don't see coloured output by default for |mach test| or |mach mochitest|, please let me know. Chris is correct that there is a config option to set the default format in ~/.mozbuild/machrc:

[test]
format=mach

Though, setting this manually should be unnecessary. Note, |mach reftest| is still the old tbpl style log because the mach format isn't compatibile with the reftest-analyzer.

Comment 7

[Matthew N. [:MattN]]

(In reply to Andrew Halberstadt [:ahal] from comment #6)
> Matt, so this is still happening for you? The default should be changed
> already by bug 1421799. If you don't see coloured output by default for
> |mach test| or |mach mochitest|, please let me know. 

This wasn't fixed for me or Kit on macOS.

$ echo $TERM
xterm-256color

> Chris is correct that
> there is a config option to set the default format in ~/.mozbuild/machrc:
> 
> [test]
> format=mach

Yeah, I looked into this yesterday after my comment and gave it a try but unfortunately it break test verification (bug 1389500).

Comment 8

[Matthew N. [:MattN]]

(In reply to Matthew N. [:MattN] (PM if requests are blocking you) from comment #7)
> Yeah, I looked into this yesterday after my comment and gave it a try but
> unfortunately it break test verification (bug 1389500).

Oops, I meant bug 1432683.

Comment 9

[Andrew Halberstadt [:ahal]]

So I think the issues this bug was originally filed for were fixed somewhere along the way. The --verify issue MattN found will be fixed in bug 1432683, while the fact that 'mach' isn't the default format will be fixed in bug 1443557.

I think this can be resolved WORKSFORME. Please re-open if the initial problem still exists.