bug#46493: [feature/pgtk] Low contrast region face

["Basil L. Contovounesios" <contovob@tcd.ie>]

Dmitry Gutov <dgutov@yandex.ru> writes:

> On 13.02.2021 18:55, Basil L. Contovounesios wrote:
>> X-Debbugs-Cc: Yuuki Harano <masm+emacs@masm11.me>
>> Severity: wishlist
>> On master:
>> 0. emacs -Q
>> 1. M-2 M-b
>> 2. M-2 M-@
>
> That's not a GTK3 build, though, right?

configure.ac suggests otherwise:

  pgtk )
    term_header=pgtkterm.h
    with_gtk3=yes
    USE_X_TOOLKIT=none
  ;;

As does the GTK3 seen in the system-configuration-features part of my
signature in the OP.

>> Repeat the same on feature/pgtk:
>> I understand that each toolkit has its look & feel, and that colour
>> perception is subjective, but the default contrast on pgtk strikes me as
>> a bit too low for text editing.
>
> Seems like it uses the same background color as the GTK3 build (the current
> one)? And that is probably the color of the window background.
>
> My current GTK theme has a bit darker windows, so the background color looks
> like fine here, FWIW.

I don't use a desktop environment, and I'm not really familiar with GTK,
but here's my $XDG_CONFIG_HOME/gtk-3.0/settings.ini:

  [Settings]
  gtk-font-name                     = DejaVu Sans 10
  gtk-icon-theme-name               = Adwaita
  gtk-recent-files-enabled          = false
  gtk-recent-files-limit            = 0
  gtk-recent-files-max-age          = 0
  gtk-theme-name                    = Adwaita

> But the screenshot exhibits another (definite) bug: when Emacs is just started,
> the cursor shape is hollow. Switch away from its window and then back: the
> cursor is now filled.

I can't reproduce that here.  My cursor is always filled so long as the
frame is focused.  The hollow cursor in my screenshot seems to be the
result of invoking scrot via gmrun, during which Emacs seems to lose
focus.  I don't know why that doesn't happen with Lucid; I assumed it
was just toolkit-specific behaviour.  [BTW, disabling blink-cursor-mode
does not change anything.]

Is this rather some kind of mishandling of focus events on Emacs' side?

Thanks,

-- 
Basil