Bug 182471 – Web Inspector: Styles: Control-Space should show completion popover

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: Web Inspector

[Nikita Vasilyev]

Control-Space shows completion in just about every modern code editor, including Sublime Text, Chrome DevTools, and Xcode.

The new styles sidebar doesn't have a way to show all available completion values for an empty prefix.

[Radar WebKit Bug Importer]

<rdar://problem/37222435>

[Nikita Vasilyev]

Created attachment 333059 [details]
Patch

[Nikita Vasilyev]

Created attachment 333060 [details]
[Animated GIF] With patch applied

[Blaze Burg]

Comment on attachment 333059 [details]
Patch

I don't like your changes. Why thread acceptEmptyPrefix everywhere? It would be better to make a different factory method on WI.CSSKeywordCompletions that returns an object with all completions, and use it.

This is not possible right now because completionProvider is a function, not a delegate/object. I don't understand the completionProvider design. It seems like this should just be a different delegate that provides completion objects to various widgets. As it's written now, it's just a callback so it can't have different methods (ie., completionsForProperty(), completionsForValue()). Not only does this make the code hard to read (i.e., what does this._completionProvider() actually do?) but it scatters the responsibility for figuring out what type completions to provide. It should be obvious at the calcite.

[Blaze Burg]

s/calcite/call site/

[Nikita Vasilyev]

(In reply to Brian Burg from comment #4)
> Comment on attachment 333059 [details]
> Patch
> 
> I don't like your changes. Why thread acceptEmptyPrefix everywhere? It would
> be better to make a different factory method on WI.CSSKeywordCompletions
> that returns an object with all completions, and use it.

How would the new factory method on WI.CSSKeywordCompletions be different from the existing ones?

[Blaze Burg]

(In reply to Nikita Vasilyev from comment #6)
> (In reply to Brian Burg from comment #4)
> > Comment on attachment 333059 [details]
> > Patch
> > 
> > I don't like your changes. Why thread acceptEmptyPrefix everywhere? It would
> > be better to make a different factory method on WI.CSSKeywordCompletions
> > that returns an object with all completions, and use it.
> 
> How would the new factory method on WI.CSSKeywordCompletions be different
> from the existing ones?

Perhaps not a different factory, but a different map chain function that doesn't filter stuff out.

[Nikita Vasilyev]

*** This bug has been marked as a duplicate of bug 194796 ***