On 2015-06-04 9:40 AM, David Rajchenbach-Teller wrote:
On 04/06/15 14:30, Ehsan Akhgari wrote:
There are very good reasons for warnings to not cause tests to fail. We
have a lot of tests that are testing failure conditions that are
expected to warn, because they are failure conditions.
Well, that's why the patch linked above offers a whitelist API, so that
tests can individually decide which families of warnings are either
expected, or ok-for-the-time-being.
Families of warnings? I don't really understand what you mean. Also,
looking at the patch in the bug, your changes seem to affect Assert.jsm
stuff, which is not related to NS_WARNING.
There are also warnings that were actually meant to be assertions, and
we should try to convert them to assertions, and that way they will fail
tests. That being said, assertion report stack traces in debug builds,
which is excruciatingly slow, so we have had to switch them to be
warnings in at least some cases (see bug 756045 for example.)
Are you talking about our weird non-fatal assertions or actual fatal
assertions?
They are both equally slow, but fatal assertions happen only once, by
definition. ;-)
_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform
