nick black
4f1de02710
replace random() with rand() for windows goodness
3 years ago
nick black
613d437b87
input: don't error out on all special keys, egads =[
3 years ago
nick black
2fb9db77e8
fbcon: update auxvecs on reblit
3 years ago
nick black
cd635316d9
fbcon_wipe: preserve auxvec
3 years ago
nick black
1113f7dd35
windows: patch out more constants
3 years ago
nick black
f28152d95e
ncplayer: eliminate %j conversions
3 years ago
nick black
1e2f3252bd
hark! a vagrant! get windows library linking
3 years ago
nick black
e5d77037dc
CMake: link in ws2_32 on windows
3 years ago
nick black
dd9e5bc312
CMake: exclude libm/librt on win32
3 years ago
nick black
03fbf57f24
fix up windows strndup compilation
3 years ago
nick black
a26b28e10f
prefer ansi c rand() to posix random(), doh
3 years ago
nick black
3f8e21fc61
huh, it's macOS, not OS X now
3 years ago
nick black
b867935708
termdesc: wrap linux-only parts up for linux
3 years ago
nick black
3172f4712a
s/SA_SIGNINFO/SA_SIGINFO/g =[
3 years ago
nick black
3d3561fe0e
windows: elide still more signal jank
3 years ago
nick black
59cb0b8bfa
plot: use PRIu64
3 years ago
nick black
0d0f30ca93
windows: don't try and use termios
3 years ago
nick black
fd13b70d27
windows: no sys/ioctl.h, move to compat.h
3 years ago
nick black
8332ff060c
windows: no langinfo.h
3 years ago
nick black
7ace56cee6
remove unused png mmap code
3 years ago
nick black
b2fb50fc39
png: cut out mmap portions on windows
3 years ago
nick black
c733293421
[rage at sky]
3 years ago
nick black
8df13818de
windows: stub out siginfo_t
3 years ago
nick black
15c3d88ddb
windows: more signal/termios mitigation #463
3 years ago
nick black
d90736f84f
define O_CLOEXEC as O_NOINHERIT on windows
3 years ago
nick black
698d1e59f5
replace j conversion with PRI for windows
3 years ago
nick black
917361d87d
windows: fd.c needs internal.h for logerror()
3 years ago
nick black
d3d0838c07
windows hates all of fd.c
3 years ago
nick black
ff2ef8571d
windows: no sys/wait.h, weird printf specifiers
3 years ago
nick black
dae8e340d2
still more windows garbage
3 years ago
nick black
0ae839c63d
windows craveth not posix signals
3 years ago
nick black
96fb4d75bf
add ncport.h
3 years ago
nick black
873ee6f227
windows cleanups
3 years ago
nick black
b97dc03a36
windows craves _byteswap_ulong
3 years ago
nick black
ed99ac670e
strike most poll.h, use winsock2.h on windows
3 years ago
nick black
432f9b0317
no netinet/in.h on windows
3 years ago
nick black
f79e41294c
notcurses_getc: don't count nonblocking exit as input_error for stats
3 years ago
nick black
8d46a0bd17
capture bytes in linux framebuffer graphics to sprixelbytes stat
3 years ago
nick black
b278d16f67
fbcon_wipe: build up auxvec
3 years ago
Michael Bradley, Jr
f008c75275
windows: use <winsock.h> instead of <sys/endian.h>
...
<sys/endian.h> doesn't exist on Windows + mingw-w64-ucrt-x86_64.
The definition of htole included in this commit is a guess and may be wrong;
and it may be wrong to use <winsock.h>.
3 years ago
Michael Bradley, Jr
3c5b8f1e10
ci: revise GHA worfklow for Windows
...
Run on all commits to `master` and PR branches.
Aim for consistency re: GHA workflows for macOS and Linux.
Also switch from mingw64 to ucrt64 and drop use of sudo, etc. that aren't
relevant on Windows + mingw-w64-ucrt-x86_64.
Add README badge for the GHA workflow for Windows - let's make the "red light"
turn into a "green light"!
3 years ago
nick black
6c45990abb
kitty: use zlib level 2, not 6 #1695
3 years ago
nick black
729290d81a
deflate for animate #1695
3 years ago
nick black
5da51e138c
intro: don't tolerate drawing failures
3 years ago
nick black
4fc1561d0f
blitterargs: no need to pass cell-pixel geom, it's in sprixel
3 years ago
nick black
a5ec720e5b
kitty: fold transcolor verdict into auxvec
3 years ago
nick black
383ce5c2b0
xray: always set marsh.dm
3 years ago
nick black
19305aa653
kitty: reuse auxvecs across frames
3 years ago
nick black
cc6bdf4b55
kitty: write transcolor pixel alpha as 0
...
when we're building out the kitty graphic, we check whether
alpha is low, or whether the color matches the transparent
color. in the latter case, we're writing the original alpha
to the auxvec. in rebuild, however, we don't know what the
transcolor was, and thus a pixel that matched it will become
visible. instead, rewrite alpha as 0 on transcolor match, so
that 0 is restored should we rebuild the cell.
3 years ago
nick black
fcd62a04ec
kitty: on redraw, move between annihilations
3 years ago