[Webkit-unassigned] [Bug 86796] Change a bunch of things in the test_expectations.txt format

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=86796





--- Comment #8 from Dirk Pranke <dpranke at chromium.org>  2012-05-17 20:04:28 PST ---
(In reply to comment #3)
> > For the record, though ...
> > > 
> > > 1. Make SLOW, SKIP and WONTFIX outcomes instead of modifiers.
> > 
> > Does SLOW imply PASS? We certainly have slow failures today, and if it doesn't imply pass, but only keeps its current semantics, it seems like it complicates the implementation (SLOW is now "special"). We could also just get rid of SLOW.
> 
> Yes. Slow needs to imply pass, unless you explicitly list another expected outcome as well. So, SLOW=>PASS, "SLOW TEXT"=>TEXT, "SLOW TEXT PASS"=>"TEXT PASS".
> 
> Specifically "SLOW TEXT" does not imply "SLOW PASS TEXT".
> 
> Personally, I prefer killing SLOW, but this has been controversial in the past and we've already had enough bikeshedding. :)
> 

Hm. I'm strongly tempted to kill SLOW instead of implement this :).

> > > 5. Use the same character as a separator before and after the test (":").
> > 
> > I'm not sure this is an improvement. If we were going to standardize on the same delimiter I would be open to using one that doesn't require a shift key as per request from rniwa. Possible candidates: "-", ";", ",", "=" (which I would prefer in that order, I think; frankly, I only really like "-" here, and I'm not sure I like any of them more than the existing ":"/"=" combo).
> 
> I don't really care what we use. Semi-colon seems fine to me. I've just seen it be a source of confusion that there are two different delimiters. 
> 

I quite like rniwa's brackets for this -- [] -- so that would be my vote. I'd probably want to stay away from commas since that starts to make the file look like CSV and that has all sorts of baggage. Frankly semicolons have some of the same issues. Avoiding any character that shows up in a layout test name already might be a good idea, which might rule out "-" and some of the others.

I also like the idea of allowing commas or more general glob-style expansion. It's a clean extension of the fact that we already allow directory prefixes, and can definitely cut down on the verbosity and maintenance for manuel entry. It might make automated updates more complicated, of course ...

I also agree w/ Ojan that I don't think we need to separate bug URLs out from the platform modifiers. URLs are distinctive enough.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.