Re: Renaming non-X x_* procedures in xdisp.c (and elsewhere)

[Alex]

Eli Zaretskii <address@hidden> writes:

>> From: Alex <address@hidden>
>> Cc: address@hidden
>> Date: Mon, 25 Mar 2019 13:21:22 -0600
>> 
>> There are three left in xdisp.c:
>> 
>> ** x_clear_under_internal_border
>> ** x_change_tool_bar_height
>> ** x_implicitly_set_name
>> 
>> Here are those called in frame.{c, h}
>> 
>> ** x_set_scroll_bar_default_{width, height}
>> ** x_set_frame_alpha
>> ** x_bitmap_icon
>> ** x_new_font
>> ** x_set_offset
>> ** x_set_window_size
>> ** x_focus_frame
>> ** x_make_frame_(in)visible
>> ** x_iconify_frame
>> 
>> This one is called in image.c:
>> ** x_query_color(s)
>
> So why do you prefer not to add this to redisplay_interface?

I just would be a bit confused by the association of some of those
procedures with redisplay. I would expect the procedures there to be
"closer" to the redisplay code, but I suppose I was imagining that there
was more of a difference here. There's also a (perhaps outdated) comment
in termhooks.h that says that redisplay_interface is window-based while
terminal hooks are frame-based.

I think that x_clear_under_internal_border is a good fit for the RIF in
any case.

> We could, of course, invent a new struct and a new macro, but we'd be
> reinventing the FRAME_RIF stuff anyway: the result will most probably
> look identical to FRAME_RIF, except for names.  Is that worth our
> while?

Probably not. I was thinking about the rest being terminal hooks,
though. WDYT?

>> This one is called in keyboard.c:
>> ** x_get_keysym_name
>
> This one should simply be renamed without the x_ prefix, I think.

That would still leave the issue of it making the code
single-backend-only. Though, I guess it's not going anywhere.


Should I take it that you are in agreement with the gui- prefix for the
Lisp side?