336428 - [regression] Caret problems (turds and positioning)

Status: RESOLVED FIXED

(Core :: DOM: Editor, defect)

Description

[Steve England [:stevee]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20060502 Minefield/3.0a1 ID:2006050216

1. Go to gmail
2. Click compose
3. Enter a load of spaces
4. Hold the delete key and observe the caret isn't undrawn.

Actual: Lots of caret images left on the text area ("|||||||||||")
Expected: No caret images left on the text area

also

1. Go to gmail
2. Click compose
3. Enter "firefox " (without the quotes, but with the trailing space) and hit return.
4. Press left arrow once.

Expected: Caret should appear after the "firefox " (note space) (eg: "firefox |")
Actual: Firefox appears to ignore the space and draws the caret after "firefox" (eg: "firefox|")

I've filed these two bugs as one since I suspect they are caused by the same problem.

Comment 1

[Martijn Wargers (dead)]

Is this in a rich editor or the textarea? This sounds related to bug 335858.

Comment 2

[Steve England [:stevee]]

Don't know if google use a rich text editor, but this only started happening to me today.

Comment 3

[Blake Kaplan (:mrbkap) (inactive)]

I check a patch in today that could fix this sort of bug. Could you re-test with tomorrow's builds? Thanks.

Comment 4

[Steve England [:stevee]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20060504 Minefield/3.0a1 ID:2006050401

OK the turd problem seems to have gone (I assume bug 335834 fixed this) but the positioning bug is still there. I'll resolve this fixed and refile the caret positioning portion of the original report as a new bug.

(I would make this bug depend/block bug 335834 but I never know which filed to use)

Comment 5

[Martijn Wargers (dead)]

In general, when a bug is fixed by a patch in some other bug, it depends on that bug.
When a patch in some other bug caused the bug, the bug is blocking the other bug with the patch.

Comment 6

[Jason Bassford]

> When a patch in some other bug caused the bug, the bug
> is blocking the other bug with the patch.

Which has never made any logical sense to me at all.  (That's the tail wagging the dog.)  It's the patch that's blocking the correct behaviour.  If the patch is fixed (backed out / modified), then the bug will be fixed. :)

Comment 7

[Steve England [:stevee]]

Thanks martin! I've saved that little nugget somewhere accessable for future reference :)

And the other positioning bug I was seeing was bug 336408.