We used to set all four quadrant bits if a full block
came through pool_direct_load(). Since users of this
function are going to be explicitly setting the quadrant
bits themselves immediately after this call, there's no
point (except for manual writes of full blocks, but those
ought be uncommon enough not to tip the scales the other
way -- this optimization was directed at blitters).
About a 5% speedup in notcurses-demo. Originates in #1316.
Eliminate NCDIRECT_OPTION_NO_READLINE (it was only introduced in
Notcurses 2.1.6). Add a new function, ncdirect_readline(). Upon
first call to this function, initialize libreadline. Destroy
libreadline in ncdirect_stop() iff we initialized it. Add
hilodirect guessing game PoC from #1325. Rename
notcurses_directmode.3 to notcurses_direct.3. Closes#1326.
Implement channels_set_fg_palindex() and channels_set_bg_palindex()
in terms of channel_set_palindex(). Add missing entries to
notcurses_channels.3 man page. Use uint32_t in place of unsigned
when dealing with channels in notcurses.h.
This completes the work for #1068. This addressed a subtle issue.
When we're using pixel->semigraphic art, we want slightly different
rendering. Essentially, imagine that we have two images, each two
pixels tall and one pixel wide. The top image is a transparent pixel
above a white pixel. The bottom image is a white pixel above a black
pixel. We'd expect the result to be two white pixels, but we can
instead get a black pixel above a white pixel. This is because the
*background* color is being merged from the bottom plane, but really
we want the *top* color. Ncvisuals are now blitted along with
information regarding which quadrants they draw over, and when
appropriate, we invert the foreground and background. Closes#1068.
Reclaim the "blitted" and "full foreground" bits in the
channels. Instead, we now write four bits, encoding the
four quadrants we might occupy as the result of a blit.
These four imply the previous two, leaving us with four
free bits remaining in the channels. This opens a clear
path to O(1)-time, zero-space blitter stacking #1068.
w00t!
For properly stacking transparent blittings (#1068), we
need tag those cells which both (1) originated in an
ncvisual operation and (2) have some transparency. For
the three affected blitters (halfblock, quadrant, and
sexblitter), call cell_set_blitted().
Rather than checking to see if there is enough space
for the crender vec following a geometry change, stash
the previous allocation, and realloc() whenever it
changes (i.e. whenever we get a resize). This brings
the crender vec back down in size if the screen gets
smaller, and seems to eliminate the problem in #1302.
