Topics

Possible issue with list column numbers in alpha

Noelia Ruiz
 

Hi Leonard and all:

Related to commit dbb1cf1e5e5f3f6b7895a12b037cc16b5f107f37:
Hide invisible columns in syslistview32 lists (#9873)
STR:

1. Hide some columns, for instance in eMule.
2. Navigate by columns in a list, for example servers.

Maybe expected: Report the column number according to the columns shown on the screen, not to the list item number of columns according to its object properties, so that the number reported agrees with the navigable columns.

Result: the reported number is based on the object properties.
Also, why should the moveToColumnNumber method be marked asprivate? This is just a question for which I'm curious. I use this method in eMule add-on and I am just working on it, so I have seen the mentioned issue.

Cheers

 

Hello Noelia,


Do you experience any differences in behaviour before and after this commit?

if an application is programmed correctly, it should really hide the columns (i.e. they shouldn't even be exposes by the syslistview32 control). However, this patch was written for situations where even though columns are visually hidden, they are exposed by the syslistview32 control. I hadn't thought about the numbering issue, but yes, I can see why that occurs. I have no idea what to do about it though. Walking every column to see whether it is hidden or not is very expensive.


Regards,

Leonard

Noelia Ruiz
 

Hello:

I'm testing with NVDA 2019.1.1 and the issue with column numbers doesn't happen, but of course hiden columns are reported by NVDA.
I understand the problem to fix this inconsistency.
At least I suppose this could be documented or configurable 馃
Regards

El 19/10/2019 a las 11:20, Leonard de Ruijter escribi贸:
Hello Noelia,
Do you experience any differences in behaviour before and after this commit?
if an application is programmed correctly, it should really hide the columns (i.e. they shouldn't even be exposes by the syslistview32 control). However, this patch was written for situations where even though columns are visually hidden, they are exposed by the syslistview32 control. I hadn't thought about the numbering issue, but yes, I can see why that occurs. I have no idea what to do about it though. Walking every column to see whether it is hidden or not is very expensive.
Regards,
Leonard

 

Feel free to file an issue on Github about this issue. I'm not sure whether to treat it as a regression though, probably not.


Regards,

Leonard

On 19-10-2019 11:39, Noelia Ruiz wrote:
Hello:

I'm testing with NVDA 2019.1.1 and the issue with column numbers doesn't happen, but of course hiden columns are reported by NVDA.
I understand the problem to fix this inconsistency.
At least I suppose this could be documented or configurable 馃
Regards

El 19/10/2019 a las 11:20, Leonard de Ruijter escribi贸:
Hello Noelia,


Do you experience any differences in behaviour before and after this commit?

if an application is programmed correctly, it should really hide the columns (i.e. they shouldn't even be exposes by the syslistview32 control). However, this patch was written for situations where even though columns are visually hidden, they are exposed by the syslistview32 control. I hadn't thought about the numbering issue, but yes, I can see why that occurs. I have no idea what to do about it though. Walking every column to see whether it is hidden or not is very expensive.


Regards,

Leonard





Noelia Ruiz
 

OK, I wouldn't treat it as a regresion since the column numbers are reported correctly according to original object properties, as previously. The problem is that the number of commands used in table navigation results inconsistent according to the column number.
And also if we implement commands to move to specific columns. For me the better solution is to make this configurable, the same as screen layout in browse mode, but maybe in object presentation.
I will fill an issue for this.

El 19/10/2019 a las 12:40, Leonard de Ruijter escribi贸:
Feel free to file an issue on Github about this issue. I'm not sure whether to treat it as a regression though, probably not.
Regards,
Leonard
On 19-10-2019 11:39, Noelia Ruiz wrote:
Hello:

I'm testing with NVDA 2019.1.1 and the issue with column numbers doesn't happen, but of course hiden columns are reported by NVDA.
I understand the problem to fix this inconsistency.
At least I suppose this could be documented or configurable 馃
Regards

El 19/10/2019 a las 11:20, Leonard de Ruijter escribi贸:
Hello Noelia,


Do you experience any differences in behaviour before and after this commit?

if an application is programmed correctly, it should really hide the columns (i.e. they shouldn't even be exposes by the syslistview32 control). However, this patch was written for situations where even though columns are visually hidden, they are exposed by the syslistview32 control. I hadn't thought about the numbering issue, but yes, I can see why that occurs. I have no idea what to do about it though. Walking every column to see whether it is hidden or not is very expensive.


Regards,

Leonard


Brian's Mail list account
 

Of course the now very old setting of guess the item number, for example in tabs or other lists also gives some false results and always has done as well.
Brian

bglists@...
Sent via blueyonder.
Please address personal email to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
This message sent from a Windows XP machine!

----- Original Message -----
From: "Noelia Ruiz" <nrm1977@...>
To: <nvda-devel@groups.io>
Sent: Saturday, October 19, 2019 11:58 AM
Subject: Re: [nvda-devel] Possible issue with list column numbers in alpha


OK, I wouldn't treat it as a regresion since the column numbers are reported correctly according to original object properties, as previously. The problem is that the number of commands used in table navigation results inconsistent according to the column number.
And also if we implement commands to move to specific columns. For me the better solution is to make this configurable, the same as screen layout in browse mode, but maybe in object presentation.
I will fill an issue for this.


El 19/10/2019 a las 12:40, Leonard de Ruijter escribi贸:
Feel free to file an issue on Github about this issue. I'm not sure whether to treat it as a regression though, probably not.


Regards,

Leonard

On 19-10-2019 11:39, Noelia Ruiz wrote:
Hello:

I'm testing with NVDA 2019.1.1 and the issue with column numbers doesn't happen, but of course hiden columns are reported by NVDA.
I understand the problem to fix this inconsistency.
At least I suppose this could be documented or configurable 馃
Regards

El 19/10/2019 a las 11:20, Leonard de Ruijter escribi贸:
Hello Noelia,


Do you experience any differences in behaviour before and after this commit?

if an application is programmed correctly, it should really hide the columns (i.e. they shouldn't even be exposes by the syslistview32 control). However, this patch was written for situations where even though columns are visually hidden, they are exposed by the syslistview32 control. I hadn't thought about the numbering issue, but yes, I can see why that occurs. I have no idea what to do about it though. Walking every column to see whether it is hidden or not is very expensive.


Regards,

Leonard