Writing a partial escape can easily lock up a terminal. This
is especially relevant when working with bitmaps, as they're
thousands or even millions of times longer than a typical
escape. Immediately before writing, block SIGINT, SIGQUIT, and
SIGTERM in the writing thread, and unblock them upon emerging
(at which point we'll immediately see any queued signal
get delivered). Don't block signals like SIGSEGV that would
seem indicative of actual problems. For this to actually work,
all other threads must also have the signals masked; we thus
now add them to the signal mask of notcurses_getc(), rather
than deleting them. Closes#1416.
* Paint sprixels in order, bottom-to-top
We don't want to have to track sprixel order whenever someone
moves an ncplane, so just keep a list growing backwards as we
pass top-to-bottom in notcurses_render_internal(). Each time
we hit a sprixel plane, splice it out of the sprixel list, and
add it to the front of our temporary list. When we hit the
bottom, stick this temporary list on the end of our existing
list (any such planes are to be deleted, which comes before
drawing). Closes#1575.
* reorder collected sprixellist; solves kitty but breaks sixel =/ #1575
* remove debugging cruft
* [rust] fix up mergedown mutability
* [bitmap] on move, clear old OPAQUE cells #1527
* update ncvisual_blitter_geom() documentation #1547
* replace some ncvisual_blitter_geom with NCVISUAL_OPTION_VERALIGNED
* [normal] replace ncvisual_blitter_geom with NCVISUAL_OPTION_VERALIGNED
* [intro] use ncvisual_blitter_geom() correctly #1547
* ncvisual_blitter_geom: only set *blitter on success
cell_release() and cell_duplicate() were deprecated in 2.2.6,
but replaced with static inlines. this breaks the abi, forcing
recompiles from client programs. restore them as exported
functions.
deprecate ncvisual_geom() in favor of new ncvisual_blitter_geom(), which allows caller to get the blitter used. replace the checks and calculations in ncvisual_render() with a call to ncvisual_blitter_geom(), unifying the two paths (and eliminating several bugs in the unloved ncvisual_geom()).
Deprecate ncplane_align(), after rewriting it as a passthrough
to new function ncplane_halign(). Add ncplane_valign(). Update
all callers. Closes#1468.
Move hardcursorpos into the renderstate carried
across rasterizations. This eliminates undesirable vertical
movement following pixel image emission #1400.
Add `notcurses_check_pixel_support()` and
`ncdirect_check_pixel_support()` per #1367. Removes
NCOPTION_VERIFY_SIXEL, again per #1367. Adds
`free_terminfo_cache()`, and calls it from both
`notcurses_stop_minimal()` and `ncdirect_stop()`.
Update all documentation. Closes#1371 and #1367.
notcurses_stats() and notcurses_stats_reset() now take the
new statlock member, as do stat modifications from render,
raster, writeout, resize, plane creation, and plane
destruction. Add nonnull attributes to stats API. Initialize
and destroy statlock as part of notcurses struct. Update
documentation. Free pilelock on error paths. Closes#1139.
After a long delay, apologies :)
Added:
* NotCurses: `can_sextant` (`notcurses_cansextant`)
* Notcurses: `linesigs_enable` (`notcurses_linesigs_enable`)
* Notcurses: `linesigs_disable` (`notcurses_linesigs_disable`)
* Pile: `top_with` (`ncpile_top`)
* Pile: `bottom_with` (`ncpile_bottom`)
* Plane: `resize_maximize` (`ncplane_resize_maximize`)
* Plane: `get_abs_x` (`ncplane_abs_x`)
* Plane: `get_abs_y` (`ncplane_abs_y`)
* Plane: `get_abs_yx` (`ncplane_abs_yx`)
* Plane: `load_egc32` (`cell_load_egc32`)
* Plane: `is_descendant_of` (`ncplane_descendant_p`)
* Progbar: new class, wraps `ncprogbar_*`
Changed:
* Plane (ABI break): `at_cursor` overloads now return `int` where
before they returned `bool` because the underlying Notcurses API
only signalled the operation status with the return value while now
it returns actual information.
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.
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 trying to force things in via library constructors
and weak symbols and --whole-archive and pkg-config tricks, just
add new functions ncdirect_core_init() and notcurses_core_init().
libnotcurses has ncdirect_init() and notcurses_init(), which pass
through to these. apps linking against notcurses-core ought use
the _core_ variants directly. This gets exactly the linkage we
want, everywhere. Convert many PoCs to _core_ variants. #1301
Signed-off-by: nick black <dankamongmen@gmail.com>
ncplane_destroy() needs to call ncplane_reparent_family(), not
ncplane_reparent() as it was doing (closes#1291). ->absy and
->absx actually are absolute; return them directly for an O(1)
ncplane_abs_yx() (down from O(N), huzzah). Add some unit tests
related to destroying and reparenting (#1286). Add ncplane_abs_y()
and ncplane_abs_x(), document them, etc.
Don't bias the nccell width by 1, or 0-length EGCs become
255 columns. We weren't actually using the width to drive
much anything until now, so this wasn't a problem, but it
is exposed as an error once we got rid of CELL_WIDE_ASIAN
and start looping through the actual egc column width.
Closes#1278, closes#1277
Improve the documentation for ncplane_set_base_cell() and
ncplane_set_base(). Go by render.c's authority (all three
channels are independently checked for transparency, and
replaced with the base cell if so--egc, fg, and bg). Add a
BASE CELL section to notcurses_plane.3. Check for, and reject,
an nccell which is secondary among a multicolumn EGC.
Closes#1238.
Add a new member 'sextants' to the terminfo cache (both
notcurses and ncvisual contain one of these, and both
initialize it the same way -- interrogate_terminfo()).
Add a new function, 'notcurses_media_defblitter()', and
deprecate 'ncvisual_default_blitter()' (the latter didn't
receive enough information to return NCBLIT_3x2). Update
all callers. Add new *internal* function rgba_default_blitter(),
so this logic can be freely changed in the future. If
sextants are available, and we're scaling, return NCBLIT_3x2.
Once we detect sextant availability, we'll have sexblitter
as a default -- stay tuned! #1114
Thread out ncls to perform the media decode in different threads, in parallel. Only the display needs be locked. On a directory of 200 files on my 39070X, this speeds ncls from ~5s to ~1s. On 75 files, we go from ~.5s to ~.2s. On a single file, we lose about 5%. To facilitate this, ncdirect_render_image() has been split into two helpers, ncdirect_render_frame() and ncdirect_raster_frame().
