Web Inspector: Canvas tab: Determine isFunction by looking at the prototype

• Component: Web Inspector
• Status: RESOLVED FIXED

[Devin Rousso]

Instead of maintaining a static list of function calls on the different canvas context types, we can just look at the prototype to determine if a given key resolves to a function.

[Devin Rousso]

Created attachment 338771 [details]
Patch

[EWS Watchlist]

Comment on attachment 338771 [details]
Patch

Attachment 338771 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.webkit.org/results/7530624

New failing tests:
inspector/canvas/recording-2d.html

[EWS Watchlist]

Created attachment 339248 [details]
Archive of layout-test-results from ews103 for mac-sierra

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: ews103  Port: mac-sierra  Platform: Mac OS X 10.12.6

[EWS Watchlist]

Comment on attachment 338771 [details]
Patch

Attachment 338771 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.webkit.org/results/7530960

New failing tests:
inspector/canvas/recording-2d.html

[EWS Watchlist]

Created attachment 339251 [details]
Archive of layout-test-results from ews106 for mac-sierra-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews106  Port: mac-sierra-wk2  Platform: Mac OS X 10.12.6

[EWS Watchlist]

Comment on attachment 338771 [details]
Patch

Attachment 338771 [details] did not pass mac-debug-ews (mac):
Output: http://webkit-queues.webkit.org/results/7531248

New failing tests:
inspector/canvas/recording-2d.html

[EWS Watchlist]

Created attachment 339258 [details]
Archive of layout-test-results from ews114 for mac-sierra

The attached test failures were seen while running run-webkit-tests on the mac-debug-ews.
Bot: ews114  Port: mac-sierra  Platform: Mac OS X 10.12.6

[Devin Rousso]

Created attachment 339276 [details]
Patch

[EWS Watchlist]

Comment on attachment 339276 [details]
Patch

Attachment 339276 [details] did not pass win-ews (win):
Output: http://webkit-queues.webkit.org/results/7534477

New failing tests:
http/tests/preload/onload_event.html

[EWS Watchlist]

Created attachment 339287 [details]
Archive of layout-test-results from ews200 for win-future

The attached test failures were seen while running run-webkit-tests on the win-ews.
Bot: ews200  Port: win-future  Platform: CYGWIN_NT-6.1-2.9.0-0.318-5-3-x86_64-64bit

[Devin Rousso]

Comment on attachment 339287 [details]
Archive of layout-test-results from ews200 for win-future

Unrelated failure

[BJ Burg]

Comment on attachment 339276 [details]
Patch

r=me

How often is this code called? Does it affect performance to dynamically look up for every call, rather than using a fixed Set lookup?

[Devin Rousso]

(In reply to Brian Burg from comment #12)
> How often is this code called? Does it affect performance to dynamically
> look up for every call, rather than using a fixed Set lookup?
`WI.RecoringAction.isFunctionForType` is called once per `WI.RecordingAction` (assuming it has been `swizzle()`d), and is called for each value in the `WI.Recording`'s `initialState.attributes`.  I think a good middle-ground would be to keep the `WI.RecordingAction._functionNames` object, but have it be filled with values as each action is processed (sort of like a memoization).

[BJ Burg]

(In reply to Devin Rousso from comment #13)
> (In reply to Brian Burg from comment #12)
> > How often is this code called? Does it affect performance to dynamically
> > look up for every call, rather than using a fixed Set lookup?
> `WI.RecoringAction.isFunctionForType` is called once per
> `WI.RecordingAction` (assuming it has been `swizzle()`d), and is called for
> each value in the `WI.Recording`'s `initialState.attributes`.  I think a
> good middle-ground would be to keep the `WI.RecordingAction._functionNames`
> object, but have it be filled with values as each action is processed (sort
> of like a memoization).

If you can capture a profile during the initial setup, and you can't see this code, then don't bother optimizing it.

[Devin Rousso]

(In reply to Brian Burg from comment #14)
> (In reply to Devin Rousso from comment #13)
> > (In reply to Brian Burg from comment #12)
> > > How often is this code called? Does it affect performance to dynamically
> > > look up for every call, rather than using a fixed Set lookup?
> > `WI.RecoringAction.isFunctionForType` is called once per
> > `WI.RecordingAction` (assuming it has been `swizzle()`d), and is called for
> > each value in the `WI.Recording`'s `initialState.attributes`.  I think a
> > good middle-ground would be to keep the `WI.RecordingAction._functionNames`
> > object, but have it be filled with values as each action is processed (sort
> > of like a memoization).
> 
> If you can capture a profile during the initial setup, and you can't see
> this code, then don't bother optimizing it.
I was able to get a profile of it while recording the canvas on <http://acko.net> and it only took 0.1ms.  The worst I've seen so far (100k actions) was 0.9ms.  I think it's fine as is.

[WebKit Commit Bot]

Comment on attachment 339276 [details]
Patch

Clearing flags on attachment: 339276

Committed r231368: <https://trac.webkit.org/changeset/231368>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/39984315>