Re: Introducing emacs-webkit and more thoughts on Emacs rendering (was Rethinking the design of xwidgets)

[Eli Zaretskii <eliz@gnu.org>]

> From: Tomas Hlavaty <tom@logand.com>
> Date: Tue, 01 Dec 2020 09:01:20 +0100
> 
> It would be more convenient, if there was a way to specify elisp
> function to draw an image.  By default, it could just call the existing
> C code but this would also allow me to specify a different elisp
> function which would then for example call w3mimgdisplay.

You need to be aware of how this works in Emacs.  In a nutshell,
there's no function "to draw the image": we call system APIs to do
that.  By the time we need to draw the image, it is already
preprocessed into a bitmap/pixmap, so we basically ask the GUI
subsystem of the OS to bitblt the image to some coordinates on the
screen which we compute.

The preprocessing of the images into bitmap/pixmap happens when the
display engine sees a 'display' property that specifies an image.
At that time we call some image library to convert the image's data
into bitmap/pixmap; then we cache the result and produce a handle to
the cache that the drawing routines will use.

So when you ask for the above, you need to fit into this scheme.

> It would be more convenient, if an image was represented as elisp data
> instead of C data.

create-image already knows how to accept image data.

> iirc there is no way to add new image types without touching C.

No, because you need to know how to convert that into the internal
format used to draw the image, which is platform dependent.

Why is it a problem to use one of the known image types?  E.g., SVG
(via the svg.el library).

> Is there a way to turn off cursor for a buffer?

There's the cursor-type variable.

> Is there a way run a function after emacs changes something on the
> screen in a buffer?

That depends by what you mean by "after emacs changes something on the
screen", since in Emacs redisplay is independent/separate of data
changes.  Please elaborate.