948893 - Hovering certain bounds in a HTML source while paused in the debugger can sometimes show a variable inspection popup

Status: RESOLVED FIXED

(DevTools :: Debugger, defect)

Comment 1

[Victor Porof [:vporof][:vp]]

Attachment: test.html

STR:
1. Open debugger on test.html
2. Click me
3. Hover trololol

Comment 2

[Victor Porof [:vporof][:vp]]

Attachment: dbg-vview-popup-html.patch

Comment 3

[Victor Porof [:vporof][:vp]]

Try looks good: https://tbpl.mozilla.org/?tree=Try&rev=0939e29f9289

Comment 4

[Panos Astithas (he/him) [:past] (please ni?)]

Comment on attachment 8345831 [details] [diff] [review]
dbg-vview-popup-html.patch

Review of attachment 8345831 [details] [diff] [review]:
-----------------------------------------------------------------

::: browser/devtools/debugger/test/browser_dbg_variables-view-popup-09.js
@@ +25,5 @@
> +      ok(false, "The variable inspection popup shouldn't have been opened.");
> +    });
> +
> +    reopenVarPopup(panel, { line: 18, ch: 10 });
> +    yield waitForTime(1000);

I assume for this to work we need 1000 >> EDITOR_VARIABLE_HOVER_DELAY, so maybe we should enforce it before someone decides to change the latter value causing oranges? Not that I'm confident we will avoid them anyway in our slow build machines :-)

Comment 5

[Victor Porof [:vporof][:vp]]

(In reply to Panos Astithas [:past] from comment #4)
> 
> I assume for this to work we need 1000 >> EDITOR_VARIABLE_HOVER_DELAY, so
> maybe we should enforce it before someone decides to change the latter value
> causing oranges? Not that I'm confident we will avoid them anyway in our
> slow build machines :-)

It's not necessarily related to that. Even if I remove the yield and there was, in fact, a problem with the debugger, the test would fail immediately.

However, the thing is, we want to make sure a popup never opens (in the reverse scenario), and we don't have an event for that :) We could add one, but maybe emitting an event whenever the mouse moves and an identifier isn't found might slow down things? I don't know.

We can't possible have time-dependant oranges here, except when the Parser becomes dumb again for whatever reason, in which case the test will fail immediately anyway. So I'm not too worried.

Comment 6

[Victor Porof [:vporof][:vp]]

https://hg.mozilla.org/integration/fx-team/rev/fe3e00a67749

Comment 7

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/fe3e00a67749