Bug 122312 – Remove Qt features from build and test tools

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Brent Fulgham]

Remove the Qt-specific logic in our build scripts and test tools

[Brent Fulgham]

Created attachment 213328 [details]
Patch

[Brent Fulgham]

Comment on attachment 213328 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=213328&action=review

> Tools/Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:147
>              'LayoutTests/platform/mac': '5',

rniwa: should this be renumbered to '4'?

> Tools/Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:150
>              'LayoutTests/platform/mac': '5',

rniwa: ditto

[Alexey Proskuryakov]

> rniwa: should this be renumbered to '4'?

These are tests, did you try running them?

[Brent Fulgham]

(In reply to comment #3)
> > rniwa: should this be renumbered to '4'?
> 
> These are tests, did you try running them?

I thought they got run as part of run-webkit-tests, but running "test-webkitpy" I see a few errors.  I'll fix and upload a new patch.

Thanks for the suggestion!

[Brent Fulgham]

Created attachment 213370 [details]
Revised to pass all script tests

[Brent Fulgham]

The patch passes all tests, but I'm still not convinced that the code in "baselineoptimizer_unittest.py" is correct. I feel like the numbers should be reordered to be consecutive.

[Dirk Pranke]

The code in baselineoptimizer_unittest.py is, at best, a bit strange. I don't think it tests what Ojan thought it did (nor did he, once I got him to look at it again), and I eventually rewrote the Blink version almost completely to use a mock filesystem; it ended up shorter and much clearer. 

I don't think you can just copy the Blink version over, though, since the fallback graph is now pretty different between WebKit and Blink, but you might take a look at it at some point and adapt it instead. 

The numbers don't need to be consecutive, just different (to indicate which results should match and which shouldn't). You might find having them be consecutive makes things at least clearer, though.

[Brent Fulgham]

Created attachment 213375 [details]
Modified based on dpranke's comments.

[Hugo Parente Lima]

*** Bug 122330 has been marked as a duplicate of this bug. ***

[Hugo Parente Lima]

After this patch I think nobody will be using old-run-webkit-tests.

[Dirk Pranke]

(In reply to comment #10)
> After this patch I think nobody will be using old-run-webkit-tests.

I think Apple is still using ORWT internally for the iOS port, unless something has changed.

[WebKit Commit Bot]

Comment on attachment 213375 [details]
Modified based on dpranke's comments.

Clearing flags on attachment: 213375

Committed r156895: <http://trac.webkit.org/changeset/156895>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Ryosuke Niwa]

We should just get rid of the baseline optimizer.  It's already broken.