Calls: enable audio sharing flow on Linux

[paterkleomenis]

Enables the desktop share audio flow on Linux within Calls UI and plumbing.

Summary

• Source handling: added unique-source handling for the desktop capture chooser.
• UI wiring: updated calls and group panel wiring for the Linux audio sharing flow.
• Group call UX: added screencast-only fullscreen mode and clarified voice vs screen-share audio controls.

Depends on desktop-app/lib_webrtc#22
Closes #26642

[CLAassistant]

All committers have signed the CLA.

[ilya-fedin]

Something is wrong, that's what I see instead of the screen chooser with this PR:

That's what I see on the release version:

[ilya-fedin]

Ah, got it, the screen chooser appears after clicking SHARE SCREEN.

[ilya-fedin]

Wow, works wonderfully. I thought this would require something like a device selector for monitor sources as I thought it's not possible to know which monitor device is the right one.

[ilya-fedin]

Please remove lib_webrtc update commits, it will have to be updated to the upstream commit when the lib_webrtc PR is merged. For now, add Depends on <link to the webrtc PR>. Please also add Closes #26642.

[ilya-fedin]

Calls: clarify voice vs screen-share audio controls is a new feature, right? If yes, please move it to a separate PR

[paterkleomenis]

Calls: clarify voice vs screen-share audio controls is a new feature, right? If yes, please move it to a separate PR

it's done.

[ilya-fedin]

it's done.

Did you see my comment in the lib_webrtc PR?

[paterkleomenis]

it's done.

Did you see my comment in the lib_webrtc PR?

Sorry, i didn't see it, i will make the changes right now.

[ilya-fedin]

Depends on desktop-app/lib_webrtc#22

This now depends on a closed PR

[paterkleomenis]

Depends on desktop-app/lib_webrtc#22

This now depends on a closed PR

i reopened it

[ilya-fedin]

Doesn't build anymore

[build] In file included from /usr/src/tdesktop/Telegram/SourceFiles/calls/calls_top_bar.cpp:23:
[build] /usr/src/tdesktop/Telegram/SourceFiles/calls/calls_call.h:16:10: fatal error: webrtc/webrtc_system_audio_capture.h: No such file or directory
[build]    16 | #include "webrtc/webrtc_system_audio_capture.h"
[build]       |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

[paterkleomenis]

There was a mistake, now it has been resolved.

[john-preston]

@paterkleomenis Thank you for your effort. I've merged the submodule PR already, but this one I can't let in the release, sorry.

Maybe I'm missing something, but for me we shouldn't try to support the unique source case inside ChooseSourceProcess case with this weird look of "very wide window with small height with checkbox then couple of buttons". It feels that we should revert all changes to ChooseSourceProcess here and instead of calling:

Ui::DesktopCapture::ChooseSource(this);

To show it in this newly added mode, we should do something like:

uiShow()->show(Box([=](not_null<Ui::GenericBox*> box) {
  box->setTitle("Sharing options"); // from tr:: of course
  box->add(object_ptr<Ui::Checkbox>("Share system audio"));
  box->addButton("Choose source");
  box->addButton("Cancel");
}));

Connect everything and that's all. This will give us nice native inside-the-call-panel small box with this checkbox as body and more understandable button "Choose source" so we press it and start sharing with unique device id and see the choose source system interface.

Do I miss something?

[john-preston]

By the way, you can contact me at https://t.me/preston discussing serious changes in code (and especially anything touching UI) should be done on Telegram, we have English-only contributors group or we can discuss stuff directly, in case it doesn't concern anyone except me.

[ilya-fedin]

the group: https://t.me/tdesktop_contributions

[john-preston]

Thanks!