Shift-space not registering

[titaniumbones]

Hi, first of all I'm not sure you guys are still taking bug reports here. I am using a recent emacs git with pgtk and gcc enabled. (the Arch Linux package).

For some time now I've realized that S-Space is being interpreted as just a plain old Space. For instance, C-h k S-Space will give me the help for SPC, no matter what mode I'm in, and (read-key) followed by shift and then space will give me keycode 32, the same code I get when I don't press the spacebar.

I can't figure out what the issue is. As far as I can tell, there is nothing relevant in my config, and emacs -Q shows the same issue. Other gtk programs can register the difference between space and shift-space.

thanks for your help as always!

[masm11]

If you don't need input methods, try:

(pgtk-use-im-context nil)

Other gtk programs can register the difference between space and shift-space.

Please tell me what those programs are.

[titaniumbones]

Hi @masm11 , I am just returning to this issue after some time, and hadn't realized I owed you a comment -- I had mixed this conversation up with another one on Reddit: https://www.reddit.com/r/emacs/comments/osscfd/pgtk_emacswaylandgnome_no_shiftspace/

To respond properly:

• turning off input methods does fix this issue for me, but it also breaks the compose key setup I use in other programs, so it is not ideal in my situation.
• other programs that seem to recognize Shift-space (usually to "scroll back up" as the inverse of scrolling down) include:
  • evince
  • firefox
  • google-chrome
  • web (the built-in gnome browser)
  • the Signal desktop app (built on Electron) I hadn't notice it before, but the WhatsApp Nativefier app (desktop client for whatsapp) seems to have the same issue as Emacs, though it is also built on Electron; it's possible that looking at the configuration differences between Signal and Whatsapp will help to diagnose the source of the problem.

I should also say that I'm running an up to date Arch Linux, Gnome 41.3 and recent versions of wayland & emacs git built form master using pgtk and native-comp.

It does seem that the keystroke is somehow accessible in some circumstances. I did just now try to use wev to inspect the keypress events, but it is not currently running on y system due to some kind of library compatibility issue (cf this comment: https://aur.archlinux.org/packages/wev/#comment-836682)

Thank you for all your work on this huge feature.

[masm11]

@titaniumbones Thanks for a lot of information.

I tried some of the noticed apps: WhatsApp Nativefier, Signal desktop, and Firefox.

I could do back-scroll with shift-space unless a text area is focused. I couldn't if a text area is focused.

I think that those apps switch disable/enable of the input method. If a text area is focused, the input method is enabled and shift-space doesn't work. If a non text area is focused, the input method is disabled and shift-space works as back-scroll.

Emacs doesn't have such a mode.

I'll take a look at the firefox source code to see if my think is correct.

[titaniumbones]

Emacs doesn't have such a mode.

ah, right, interesting. hmm. two thoughts so far:

• first, I don't really understand what work Shift-space is doing in ibus input method. It doesn't seem to be documented anywhere, but I am only using input methods in order to get the compose key working properly, which otherwise I don't seem to be able to do. So maybe it is commonly used for non-European character input.
• the main place I use Shift-Space in Emacs is with PDF's (bound to pdf-view-scroll-down-or-previous-page). I have bound to another key but as I often switch from one PDF reader to another, I find that's not quite perfect for me. Would it be possible to use a buffer-local variable to somehow affect the relationship of emacs to input methods? I would guess not, since pgtk-use-im-context is written in C, but thought I should ask.

many thanks.

[masm11]

Would it be possible to use a buffer-local variable to somehow affect the relationship of emacs to input methods?

OK, I'll try, though I don't know whether I can implement it.

[masm11]

Try this patch!

pgtk-im-per-buffer.diff.gz

EDIT: I changed pgtk-use-im-context from a function to a buffer-local variable.

[titaniumbones]

I missed this 2 weeks ago! sorry for the delay responding.

I will figure out how to apply this patch and recompile! Thanks so much for the help. Really looking forward to trying this.