Bug 119370 - Sidebar - Calc sidebar unresponsive when editing a chart
Summary: Sidebar - Calc sidebar unresponsive when editing a chart
Status: RESOLVED WORKSFORME
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Calc (show other bugs)
Version:
(earliest affected)
6.1.0.0.alpha0+
Hardware: All All
: medium normal
Assignee: Not Assigned
URL:
Whiteboard:
Keywords: bibisected, bisected, regression
: 119870 120110 (view as bug list)
Depends on:
Blocks: Sidebar-UI-UX
  Show dependency treegraph
 
Reported: 2018-08-19 19:25 UTC by Alex
Modified: 2020-05-30 08:32 UTC (History)
9 users (show)

See Also:
Crash report or crash signature:


Attachments
Screenshot side bar for chart OpenGL disable. (84.82 KB, image/png)
2018-08-19 20:56 UTC, m_a_riosv
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Alex 2018-08-19 19:25:05 UTC
Description:
When editing a chart using the sidebar, the sidebar is unresponsive and so cannot be used. Seems to be the case when creating new charts and opening old ones, both docx and odt. 

Observed on Win10 LO 6.1.0.3

Steps to Reproduce:
1. Open Calc
2. Create a chart (e.g. enter 1,2,3 in A1-A3 and B1-B3 then create a scatter graph)
3. Double click into the graph to go into edit mode
4. Go into the sidebar to edit.

Actual Results:
Nothing....

Expected Results:
The buttons do their function, e.g. change the colour of the scatter points.


Reproducible: Always


User Profile Reset: Yes



Additional Info:
Opening a pre-existing chart also doesn't work
Comment 1 Alex 2018-08-19 19:26:57 UTC
Bug also present when editing a chart pasted into a docx/odt file - same LO version.
Comment 2 m_a_riosv 2018-08-19 20:56:11 UTC
Created attachment 144313 [details]
Screenshot side bar for chart OpenGL disable.

Same behavior wit OpenGL enable.

With OpenGL disable instead nothing it's like select styles. Attached screenshot.
Comment 3 Xisco Faulí 2018-08-20 18:10:08 UTC
Regression introduced by:

https://cgit.freedesktop.org/libreoffice/core/commit/?id=10e4528e8cb9818e7b8fcef8b1e96985f8b28640

author	Katarina Behrens <Katarina.Behrens@cib.de>	2018-03-06 12:17:07 +0100
committer	Katarina Behrens <Katarina.Behrens@cib.de>	2018-03-19 12:09:53 +0100
commit 10e4528e8cb9818e7b8fcef8b1e96985f8b28640 (patch)
tree 5f2d336afbe21e5ca834618f4a974d79b7ab8602
parent d05ee87bdd779a8d799269af4bb45ae8a4ea3150 (diff)
tdf#115806: Remember last active deck also on context switch
within 1 application

Bisected with: bibisect-linux64-6.1

Adding Cc: to Katarina Behrens

In order to bisect this issue, I did it with a clear profile on every step. After 10e4528e8cb9818e7b8fcef8b1e96985f8b28640, the sidebar works the first time I select the chart, if I deselect it and select it again, it no longer works...
Comment 4 riesslibo 2018-09-05 10:11:50 UTC
Can also reproduce this with a given chart in an ods with Win7/Libo 5.2.0.4

Workaround for color, trancparency and line/border:

use the classic menue Format -> Format selected object

(first entry in the menue Format)
Comment 5 Xisco Faulí 2018-09-14 10:05:51 UTC
*** Bug 119870 has been marked as a duplicate of this bug. ***
Comment 6 Martin Sourada 2018-09-14 12:39:41 UTC
Since there's an additional info in the dupe (sorry about duplication, I hadn't found this one...):

* The regression is earlier than the one in the bibisection (Comment 3), actually it's at least two regressions. 
  * Initial regression is between 5.4.7.2 and 6.0.6.2 (edit works for first edit, but unresponsive after leave-enter, in 5.4 it works)
  * Second regression is the one that was bibisected.
* The problem noted in Comment 4 is probably Bug 100180
Comment 7 Martin Sourada 2018-09-16 18:14:42 UTC
Unless I made a mistake (it's my first bibisection), it seems like the first regression is caused by this commit (bibisect-linux-64-6.0):
 897245ca0ff3c8ce47b678214fb70326fdbed2ce is the first bad commit
commit 897245ca0ff3c8ce47b678214fb70326fdbed2ce
Author: Jenkins Build User <tdf@pollux.tdf>
Date:   Mon Apr 9 21:17:04 2018 +0200

    source 2c51260d91490a6fc512875d8befb38367bf3227
    
    source 2c51260d91490a6fc512875d8befb38367bf3227

:040000 040000 93920d0559b6566bd3e4b044b3263740026e2caa 83370c22b7eb32bff6ffa8cc4e0d83dc490a34a5 M	instdir
Comment 8 Martin Sourada 2018-09-16 18:39:04 UTC
(In reply to Martin Sourada from comment #7)
> Unless I made a mistake (it's my first bibisection), it seems like the first
> regression is caused by this commit (bibisect-linux-64-6.0):
> 
> 897245ca0ff3c8ce47b678214fb70326fdbed2ce is the first bad commit
> commit 897245ca0ff3c8ce47b678214fb70326fdbed2ce
> Author: Jenkins Build User <tdf@pollux.tdf>
> Date:   Mon Apr 9 21:17:04 2018 +0200
> 
>     source 2c51260d91490a6fc512875d8befb38367bf3227
>     
>     source 2c51260d91490a6fc512875d8befb38367bf3227
> 
> :040000 040000 93920d0559b6566bd3e4b044b3263740026e2caa
> 83370c22b7eb32bff6ffa8cc4e0d83dc490a34a5 M	instdir

Seems like I *made* a mistake, not sure where though :( Since the source commit seemed very unlinkely to blame, I tried hard resetting the repo and bibisecting again, with double checking every step and resetting user profile in each try. The "correct" bad commit seems to be: 790f6ffe130014691c5fc7473bebde812903eec9 is the first bad commit
commit 790f6ffe130014691c5fc7473bebde812903eec9
Author: Jenkins Build User <tdf@pollux.tdf>
Date:   Thu Apr 12 21:31:47 2018 +0200

    source afaa944633cb94205fb835c1d93775fa51ead749
    
    source afaa944633cb94205fb835c1d93775fa51ead749

:040000 040000 76dd577dbc02604c3d1d8b6b8c80a1b752030b5a a2f86bcd7804c1883ef9b4a25bfdee351f10043a M	instdir


aka

diff options
context:	
space:	
mode:	
author	Katarina Behrens <Katarina.Behrens@cib.de>	2018-03-06 12:17:07 +0100
committer	Caolán McNamara <caolanm@redhat.com>	2018-04-12 17:19:45 +0200
commit afaa944633cb94205fb835c1d93775fa51ead749 (patch)
tree b07c3a8aea487481760c074d3b3946ee4462e586
parent a7e8fa20d8c1a529ef3cd7f69d3afa2e06412800 (diff)
tdf#115806: Remember last active deck also on context switch


I'm sorry about the confusion.
Comment 9 Oliver Brinzing 2018-09-25 09:17:25 UTC
*** Bug 120110 has been marked as a duplicate of this bug. ***
Comment 10 Oliver Brinzing 2018-09-25 09:26:17 UTC
reproducible with lo 6.1.1.2 

but i can no longer confirm this issue with upcoming:

Version: 6.1.2.1 (x64)
Build-ID: 65905a128db06ba48db947242809d14d3f9a93fe
CPU-Threads: 4; BS: Windows 10.0; UI-Render: Standard; 
Gebietsschema: de-DE (de_DE); Calc:
Comment 11 Alex 2018-10-06 10:05:39 UTC
I can still reproduce this bug with:

Version: 6.1.2.1 (x64)
Build ID: 65905a128db06ba48db947242809d14d3f9a93fe
CPU threads: 4; OS: Windows 10.0; UI render: GL; 
Locale: en-GB (en_GB); Calc: CL

It's interesting that Oliver Brinzing cannot [Comment 10].
Comment 12 Katarina Behrens (Inactive) 2018-10-22 20:45:46 UTC
Im Caolan fixed this in https://cgit.freedesktop.org/libreoffice/core/commit/?id=542a9a340972da7739c33465620ffe48da8745dd

but please retest
Comment 13 Aron Budea 2018-11-25 04:04:32 UTC
For me it still doesn't work properly in 6.3.0.0.alpha0+ (67e062aa5e5946d4985921fe2b6f87766f363ddc) / Windows 7.

However, what I see doesn't seem to be a regression from the mentioned commit, as I can reproduce even with 5.1.0.3, when the chart sidebar was introduced.

My steps (based on description):
- Have sidebar open, but no decks open initially.
- Create a chart, and click into it.

(=> Bug 1: list of sidebar decks isn't updated, it should only show Properties)

- Select an element, eg. the label.
- Open Properties deck, and click anywhere on it.

=> It behaves as if you selected a cell in the sheet where you clicked.

This only happens if you entered the chart with no open deck/sidebar. However, if you entered it with an open deck/sidebar, and attempt to hide it, its empty space will remain on the screen. It isn't exactly what Miguel Angel's screenshot shows, but is likely related.
Comment 14 Alex 2019-06-23 13:11:05 UTC
I don't fully reproduce this in:

Version: 6.3.0.0.beta1 (x64)
Build ID: a187af327633f5f00363be5131bd21a13e0f1a7b
CPU threads: 4; OS: Windows 10.0; UI render: GL; VCL: win; 
Locale: en-GB (en_GB); UI-Language: en-GB
Calc: threaded

For example I can change the colour of things. I can turn on subtitle but I cant turn it off.

Performance is very jumpy. But it doesn't act as if you've clicked the cell below the sidebar as it did before.

Next time ive access to a Linux machine I'll retest on there as well.
Comment 15 Alex 2019-06-23 13:13:47 UTC
I cannot therefore reproduce the observations of Comment 13. I only get the properties icon in the sidebar.

Thanks for the updates and the commit!
Comment 16 Aron Budea 2020-05-30 08:27:47 UTC
There are at least three bugs on various issues with the chart sidebar appearing broken in some cases: bug 94322, bug 103109 and bug 115147. And it can seem as it's not working/broken, because of the broken appearance, eg. as shown in attachment 144313 [details].

I'm marking this bug as WORKSFORME, because the steps are now different, and not about the original issue. If there's an issue that doesn't fit in the above 3 bugs (ie. anything concerning actual behavior of the chart sidebar controls), please open a new bug report.
Comment 17 Aron Budea 2020-05-30 08:32:59 UTC
(In reply to Aron Budea from comment #16)
> appearing broken in some cases: bug 94322, bug 103109 and bug 115147. And it
> can seem as it's not working/broken, because of the broken appearance, eg.
> as shown in attachment 144313 [details].
I wasn't precise at referring to this attachment as this particular case is probably fixed by the commit mentioned by Katarina in comment 12.

However, if you first close the sidebar, double-click the chart, then open the properties sidebar, you will see underlying cells shown in some places, and other oddities, which fit the mentioned bug reports.