1132950 - Sorting direction markers are staying when another column is sorted

Status: RESOLVED FIXED

(Firefox :: General, defect)

Description

[Alice0775 White]

Steps To Reproduce:
1. Open about:config
2. Click "Preference name" tree header to change sort direction
   --- Sorting direction marker appears on "Preference name" column as expected
3. Click another tree header(e.g. "Status") to change sort direction
   --- Sorting direction marker appears on "Status" column  as expected
       However, Sorting direction marker is still shown on "Preference name" column

Actual Results:
Sorting direction markers are staying when another column is sorted

Expected Results:
Sorting direction markers should only be shown on column header that sorted

Comment 1

[Alice0775 White]

Attachment: screenshot, comparison between Nightly38.0a1 and Aurora37.0a2

Comment 2

[Richard Marti (:Paenglab)]

It's not a bug directly introduced by bug 1125636.

The bug comes from https://dxr.mozilla.org/mozilla-central/source/toolkit/themes/shared/in-content/common.inc.css#686

There are two solutions:
1. Change the CSS code to not show the arrow when the empty attribute "sortDirection" is set.
2. Check where the "sortDirection" is set in JS and not removed when a other column is the active sort column.

For 1. I could fix this, but I think 2. would be cleaner but I'm not a JS specialist.

Comment 3

[Tim Nguyen :ntim]

Attachment: Patch

Comment 4

[Jared Wein [:jaws] (please needinfo? me)]

Comment on attachment 8564319 [details] [diff] [review]
Patch

Review of attachment 8564319 [details] [diff] [review]:
-----------------------------------------------------------------

Looks like we could do with just the config.js changes. Any reason that we need to make the changes for common.inc.css? Presence of an attribute in XUL-land is pretty common to mean that it should be enabled.

Comment 5

[Tim Nguyen :ntim]

(In reply to Jared Wein [:jaws] (please needinfo? me) from comment #4)
> Comment on attachment 8564319 [details] [diff] [review]
> Patch
> 
> Review of attachment 8564319 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> Looks like we could do with just the config.js changes. Any reason that we
> need to make the changes for common.inc.css? Presence of an attribute in
> XUL-land is pretty common to mean that it should be enabled.

I guess so, I just want to prevent this in the future.

Comment 6

[Jared Wein [:jaws] (please needinfo? me)]

Comment on attachment 8564319 [details] [diff] [review]
Patch

Review of attachment 8564319 [details] [diff] [review]:
-----------------------------------------------------------------

Yeah, not necessary though. Let's keep this change small and only do the config.js change.

Also, we shouldn't use :-moz-any() as the right-most selector: https://developer.mozilla.org/en-US/docs/Web/CSS/:any#Issues_with_performance_and_specificity

r=me for only the config.js changes.

Comment 7

[Tim Nguyen :ntim]

Attachment: Patch v2 (r=jaws)

Removed the CSS changes.

Comment 8

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/integration/fx-team/rev/2c5db2a83e34

Comment 9

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/2c5db2a83e34

Comment 10

[David Lawrence [:dkl]]

Mistakenly filed against Firefox 38 and should be instead 38 Branch. Sorry for the spam. dkl