report subscripts and superscripts #10919: exclude reporting on elements in browsers such as buttons etc.

[Adriani90]

Steps to reproduce:

1. Open NVDA and enable reporting of subscripts and superscripts in NVDA's document formating settings
2. Open a browser i.e. Firefox
3. While navigating on websites with arrow keys or quick commands in browse mode, notice what NVDA reports (i.e. on some buttons in the comment area of this issue)

Actual behavior:

NVDA reports very often "base line"

Expected behavior:

NVDA should report "base line" only in mathematical content or if this is part of a text block. Not on elements such as buttons etc.

System configuration

NVDA installed/portable/running from source:

Installed

NVDA version:

Version: alpha-20086,f436d4d7

Windows version:

Windows 10 1909 Update

Name and version of other software in use when reproducing the issue:

All commonly used browsers, newest versions.

Other information about your system:

Other questions

Does the issue still occur after restarting your computer?

yes

Have you tried any other versions of NVDA? If so, please report their behaviors.

yes, the version before merging #10919 but this had not have separate reporting for subscripts and superscripts.

If addons are disabled, is your problem still occuring?

yes

Did you try to run the COM registry fixing tool in NVDA menu / tools?

n/a

[feerrenrut]

Other than in MathML or other explicitly defined math, it could be difficult to determine "math content". I think a reasonable path forward is to only announce base line if the last text announced was superscript or subscript. Examples could help to clarify your issue description:

Given:
This is a regular sentence with ^{some superscript note} followed by more text

Expected:
Speech: "This is a regular sentence with superscript some superscript note baseline followed by more text"

[feerrenrut]

Interestingly, I can't get NVDA to recognise that "some superscript note" is actually formatted as superscript. It was written like this in the comment, and visually appears as superscript. Pressing NVDA+F on this text announces it as baseline.

This is a regular sentence with <sup>some superscript note</sup> followed by more text

[ultrasound1372]

Glad I'm not the first to report the issue just referenced, but check there for a good example and breakdown of the behavior. Discussion should continue here though.

[CyrilleB79]

@Adriani90 and @ultrasound1372, would #11897 allow to close this issue?
The issue with spurious reporting of "baseline" should be fixed but the implemented solution is not to exclude some web elements for reporting.

[CyrilleB79]

Here is a link to download the test build.

[CyrilleB79]

@ultrasound1372 and @Adriani90 could you try the test build and indicate if it meets your needs? Would #11897 solve the root cause of this issue?

Note however that it may not solve the issue raised by @feerrenrut regarding <sup> tags in GitHub comments described in #11078 (comment)

[CyrilleB79]

Ping @Adriani90, @ultrasound1372
Please provide feedback.

[ultrasound1372]

@CyrilleB79 appears to work for the use case I presented in my issue.

[CyrilleB79]

@CyrilleB79 appears to work for the use case I presented in my issue.

Yes @ultrasound1372, I have made a PR that may solve this issue.
And that's why I was asking you to test the test build to know if the issue is fixed in this build.
Thanks.

[ultrasound1372]

Yes, I was saying I tested the test build and it fixes the issue I reported.

[CyrilleB79]

@Adriani90 it is a last attempt to ask you to test this updated build of PR #11897.
Let me know if you consider that it resolves this issue, even if the fimplemented solution is not what you suggest in this issue.

[Adriani90]

This is still reproducible in NVDA alpha-23852,df92ef66.
Especially in browse mode. I have updated the description accordingly.
Also when you move to the address bar with f6 and press ctrl+left arrow, NVDA reports "base line".

I can reproduce this issue now only in Firefox, version 92 currently.

cc: @jcsteh

[CyrilleB79]

@Adriani90 my question was not to test last alpha but the build of PR #11897 (cf. my previous message). Could you give a try and see if it resolves this problem for you?