678120 - Browser-chrome mochitests should give a stacktrace on failures

Status: RESOLVED FIXED

(Testing :: Mochitest, defect)

Description

[Blair McBride [:Unfocused] (UNAVAILABLE)]

Trying to find out whats happening in a failing browser-chrome test can be annoying. A stacktrace can help.

Comment 1

[Blair McBride [:Unfocused] (UNAVAILABLE)]

Attachment: [checked-in] Patch v1

Had this sitting in my patchqueue for awhile, and have found it very useful.

No idea what this could potentially break.

Comment 2

[(not currently active) Ted Mielczarek]

Comment on attachment 552290 [details] [diff] [review]
[checked-in] Patch v1

Review of attachment 552290 [details] [diff] [review]:
-----------------------------------------------------------------

As long as it doesn't blow anything up it seems fine to me.

::: testing/mochitest/browser-test.js
@@ +341,5 @@
> +      this.msg += "\nStack trace:\n";
> +      var frame = aStack;
> +      while (frame) {
> +        this.msg += "    " + frame + "\n";
> +        frame = frame.caller;

In practice, do you need the whole stack, or could we just present the top few frames? Perhaps in case of failure it's better to err on the side of more information than less?

Comment 3

[Blair McBride [:Unfocused] (UNAVAILABLE)]

(In reply to Ted Mielczarek [:ted, :luser] from comment #2)
> In practice, do you need the whole stack, or could we just present the top
> few frames? Perhaps in case of failure it's better to err on the side of
> more information than less?

In practice, there's no way to tell ahead of time how much of the stack you'll need. So yes, much better to have more info than not enough, IMO.

Have ran this through Try several times with other patches that produced test failures, and everything has worked well.

https://hg.mozilla.org/integration/fx-team/rev/2470993f6287

Comment 4

[Rob Campbell [:rc] (:robcee)]

Comment on attachment 552290 [details] [diff] [review]
[checked-in] Patch v1

http://hg.mozilla.org/mozilla-central/rev/2470993f6287