isontheline / pro.webssh.net

iOS / iPadOS / macOS SSH Client
https://apps.apple.com/us/app/id497714887
MIT License
317 stars 46 forks source link

Better VoiceOver Support #1167

Open giulianozor opened 3 months ago

giulianozor commented 3 months ago

Remaining Tasks

isontheline commented 3 months ago

Hello @giulianozor 👋

Thanks for your feedback 🙏

Could you tell me more about your request?

VoiceOver is not implemented at all inside WebSSH and I would like to know where you want to start

isontheline commented 2 weeks ago

@giulianozor I need more information please

giulianozor commented 2 weeks ago

@isontheline Hello, the main issues I face are; – When using the keyboard, I get the voice feedback when inserting text, but not when deleting text. – I cannot get the terminal content reading. The only way is to enable screen recognition, but this will read everything from the beginning, so it is not very useful. The best would be to have each line selectable and readable. – The Corso to move up and down, left, and right in the upper row of the keyboard is not usable when VoiceOver is enabled. That is all I can think of right now.

Thank you. G G

isontheline commented 2 weeks ago

Thanks for your feedback @giulianozor !

– When using the keyboard, I get the voice feedback when inserting text, but not when deleting text.

I'm new to VoiceOver so I need some guidance. When I delete chars I hear a sound, is it not the intended behaviour?

– I cannot get the terminal content reading. The only way is to enable screen recognition, but this will read everything from the beginning, so it is not very useful. The best would be to have each line selectable and readable.

I'm not able to provide text selection on the terminal due to an underlying library, you need to use "Show Scrollback Buffer" I know that's not ideal so I will try to make a workaround

– The Corso to move up and down, left, and right in the upper row of the keyboard is not usable when VoiceOver is enabled.

Are you talking about this component?

image
giulianozor commented 2 weeks ago

@isontheline When VoiceOver is enabled you get the reading of each character you type or delete. For example, if you type 123, you can hear one, two, three. Then if you delete 32, you hear three, two. This allows you to know which character you typed or delete. Also, there are other mouths like words or lines where you get the reading for a whole for the line, but in this case only when you delete, move or select a different line or worth. If it is not possible to get the liner readings without entering the buffet screen, that is very unfortunate Punto it makes the whole software 99% useless because there is no easy way to receive the output of your comments. Maybe the underlying library could be updated to support voiceover? I do not use it on Mac, only on iPad, but I suppose the same issue is present in the desktop version as well. The comp you circled in the AMEX is exactly the one I am talking about.

Thanks.

G

PS: sorry for not quoting correctly, it is difficult for me to quote using dictation

isontheline commented 2 weeks ago

PS: sorry for not quoting correctly, it is difficult for me to quote using dictation

No problem at all

that is very unfortunate Punto it makes the whole software 99% useless because there is no easy way to receive the output of your comments

I will improve WebSSH in the upcoming releases to have a better accessibility

If you are still using WebSSH, I'll ping you when I've made some improvements to the accessibility elements

giulianozor commented 2 weeks ago

@isontheline Yes, I still use it, at least while I have some vision left. I will be glad to provide feedback and test.

I also tried other clients, but I have not found anything even close to web SSX, so I really hope you can improve the accessibility.

Thanks

G i

isontheline commented 1 week ago

Hello @giulianozor 👋

WebSSH 28.1.1438 beta release is available with better VoiceOver support : https://testflight.apple.com/join/QSrBK59z

giulianozor commented 1 week ago

@isontheline hello, Thank you very much for the update! The cursor keys now working properly, and the voice reading for command outputs is great!

If possible, it would be great if you can change the voice output for the left and right arrow keys so the current character is red instead of left and right. These may or may be not related to the issue where the deleted characters are not spoken when deleted. Also, it would be awesome if when using up and down you get the command reading, if any is retrieved from the history.

New issue: it seems it is not possible to y scroll up and down to see the screen buffer. this is usually done with a swipe up or down with three fingers.

again, thanks a lot for the update!

giulianozor commented 6 days ago

@isontheline hello I noticed that the issue about spoken deleted characters is fixed if I start a session with VoiceOver enabled, but if I start VoiceOver after connecting to a session, it is still present. In this case, the cursor key reverts back to the original button too.
I think that maybe an option in the settings to enable voiceover integration could work better, at least for this. I have seen this approach in other southwest, like IntelliJ idea. On the first start with VoiceOver enabled you get a pop-up that explains how to turn on the full integration in the settings. If you do not enable this option, you still get some integration, but with some limitations. if you enable this option, the changes are kept even if you turn off voiceover i.e. the keyboard shortcuts are different

Thank you.

G

isontheline commented 5 days ago

Hello @giulianozor 👋

Many thanks for your strong feedback 🙏

No problem I will make these improvements over the upcoming weeks and I will ping you each time a new enhancement is available.

Thank you to help me to enhance WebSSH accessibility!

giulianozor commented 1 day ago

@isontheline hello, Another useful addition could be to speak the auto completion text when you press the tab key. Thanks G

isontheline commented 23 hours ago

Hello @giulianozor 👋

No problem I'm planning it at same time with "When using up and down, speak the retrieved command from history"