bug#37473: 27.0.50; antialias setting is not preserved by, inheriting

[Colin Baxter]

>>>>> David Ponce <da_vid@orange.fr> writes:

    > Hello, The below related today's commit to xface.c breaks
    > antialiasing:

    >  From b2d11d69dd49864874f8fe53669b4049e83bfce9 Mon Sep 17 00:00:00
    > 2001 From: Po Lu <luangruo@yahoo.com> Date: Sun, 19 Jun 2022
    > 21:57:11 +0800 Subject: More conservative fix for bug#37473

    > * src/xfaces.c (realize_gui_face): Add more conservative fix,
    > since the last change makes C-x C-+ lead to weight weirdness on my
    > machine.  --- src/xfaces.c | 23 ++++++++++++----------- 1 file
    > changed, 12 insertions(+), 11 deletions(-)

    > diff --git a/src/xfaces.c b/src/xfaces.c index 4242205..25b5e4d
    > 100644 --- a/src/xfaces.c +++ b/src/xfaces.c @@ -5909,7 +5909,7 @@
    > realize_gui_face (struct face_cache *cache, Lisp_Object
    > attrs[LFACE_VECTOR_SIZE] #ifdef HAVE_WINDOW_SYSTEM struct face
    > *default_face; struct frame *f; - Lisp_Object stipple, underline,
    > overline, strike_through, box; + Lisp_Object stipple, underline,
    > overline, strike_through, box, temp_spec;

    >     eassert (FRAME_WINDOW_P (cache->f));

    > @@ -5953,17 +5953,18 @@ realize_gui_face (struct face_cache
    > *cache, Lisp_Object attrs[LFACE_VECTOR_SIZE] if (! FONT_OBJECT_P
    > (attrs[LFACE_FONT_INDEX])) { /* We want attrs to allow overriding
    > most elements in the - spec, but we don't want to start with an
    > all-nil font, - either, because then we lose attributes like -
    > antialiasing.  This should probably be fixed in a - different way,
    > see bug#17973 and bug#37473.  */ - Lisp_Object spec =
    > copy_font_spec (attrs[LFACE_FONT_INDEX]); - Ffont_put (spec,
    > QCfoundry, Qnil); - Ffont_put (spec, QCfamily, Qnil); - Ffont_put
    > (spec, QCregistry, Qnil); - Ffont_put (spec, QCadstyle, Qnil); +
    > spec (IOW, to start out as an empty font spec), but + preserve the
    > antialiasing attribute.  (bug#17973, + bug#37473).  */ + temp_spec
    > = Ffont_spec (0, NULL); + + if (FONTP (attrs[LFACE_FONT_INDEX])) +
    > Ffont_put (temp_spec, QCantialias, + Ffont_get
    > (attrs[LFACE_FONT_INDEX], + QCantialias)); +
    > attrs[LFACE_FONT_INDEX] - = font_load_for_lface (f, attrs, spec);
    > + = font_load_for_lface (f, attrs, temp_spec); } if (FONT_OBJECT_P
    > (attrs[LFACE_FONT_INDEX])) { --

    > cgit v1.1


    > The below marked line seems not correct. According to the previous
    > condition "if (! FONT_OBJECT_P (attrs[LFACE_FONT_INDEX]))", it
    > should always be false, so the antialising attibute is never
    > preserved:


    >     if (FONTP (attrs[LFACE_FONT_INDEX])) <=====================
    > Ffont_put (temp_spec, QCantialias, Ffont_get
    > (attrs[LFACE_FONT_INDEX], QCantialias));

    > Thanks, David Ponce


Yes, I see exactly the same breakage. I reported this to emacs-dev -
wrong site!

Best wishes,

Colin.