Re: Stipples support in MS-Windows port

[Elijah G <eg642616@gmail.com>]

[-- Attachment #1.1: Type: text/plain, Size: 3450 bytes --]

On Fri, May 3, 2024 at 1:30 AM Po Lu <luangruo@yahoo.com> wrote:
> > The closest thing i got was using w32_draw_rectangle
> > inside w32_draw_glyph_string_background
> > but for some reason it freezes and crashes after evaling
> > the snippet, I'm not sure if it is related to the way I did it or is
> > anything related to GC.
>
> You appear to fundamentally misunderstand the nature of stipples and
> which component bears the burden of its implementation.  A stipple
> pattern is a repeating bitmap pattern that replaces the solid background
> color of a face, in which pattern "on" pixels are tinted the foreground
> color of the face, and off pixels the background, and whose repetition
> is such that its tiles (instances of its bitmap) extend from a
> conceptual "origin" defined by the position of the drawable.  In X,
> stipples are natively supported by graphics primitives, being simply
> properties of the GC (graphics context), to wit, the fill style, the
> stipple bitmap, and the tile/stipple origin offset, which three
> properties are direct determinants of the imagery produced by
> polygon-filling or stroking operations, and are configured some time
> before x_draw_glyph_string_background is invoked, relieving the latter
> of all responsibility, with respect to stipples, but enabling the
> prepared stipple pattern when appropriate.  This is the reason that the
> disabled code in the related W32 function is no more than a few calls to
> XDrawRectangle, but as the supporting infrastructure, both in Emacs and
> the display server, is not present on MS Windows, enabling just this
> code, even if it does not crash, serves little purpose.

Thank you for explain me, I was a little confused in how stipples works.

> Your task is not to uncomment this handful of lines, but to establish
> how best to express these X concepts in terms of primitives available on
> MS Windows, and implement them in a manner appropriate for _Windows_,
> without regard to the specifics of the X implementation.

Thanks, as far as i understood then should not i touch or even edit those
lines?
I was thinking in define them (since them are only available in X) inside
the same file, similar in how pgtkterm.c or androidterm.c does, i think i
should check again `image_create_bitmap_from_data' and
`prepare_face_for_display' definitions.

> > Below this mail can you get the patch for test it
> > (Also i'm sorry if the patch is too long i accidentally
> > formatted it with clang-format)
>
> I suggest reverting the portions of the patch that are the work of
> clang-format, as we cannot accept such drastic whitespace changes, and
> the substantive changes simply vanish in their midst.

Thank you, i've reverted the changes done by clang-format, reviewing the
test that i made i got notice it works a bit similar to `box' property and
not to a stipple, but anyways i'm sending it here if it can be useful.

> > Also can i ask what is the purpose of this line?
> > ```
> > # include "bitmaps/gray.xbm"
> > ```
>
> It includes the bitmap image, bitmap/gray.xbm, which is a C header
> defining an array and several parameters.  This is not relevant on
> Windows.

Thank you for the help, I wasn't familiar with X.org stuffs (since I'm in
Windows and most of them are not available) and with C (specially in a big
project such as emacs).
But I have faith that I can help in this.

[-- Attachment #1.2: Type: text/html, Size: 4181 bytes --]

[-- Attachment #2: 0001-Test-stipples-in-W32.patch --]
[-- Type: application/x-patch, Size: 1480 bytes --]