f05e40a30a
Don't interpret an n*<Esc> as (n-1)*<Alt>-<Esc>
2015-08-09 14:52:02 +02:00
0b5d8ecd0d
Fix xterm cursor key modifiers
2015-07-25 19:46:42 +02:00
128524cd2b
Fix compiler warnings in test 05read
2015-07-25 19:08:11 +02:00
cd87fee243
Mark some issues
2015-07-25 19:05:23 +02:00
dc7b9efd71
Make cursor keys work better
2015-07-25 19:05:00 +02:00
477f732bc3
Fix SS3 keysyms
2015-07-25 19:02:57 +02:00
d49fbadd67
Fix off-by-one bug
2015-05-28 20:16:55 +02:00
bc0a71f082
Finally fix all tests
...
While still supporting rxvt shifted keys.
2015-05-28 20:02:14 +02:00
be1cccd252
Mark an issue that breaks tests
2015-05-27 21:42:21 +02:00
c628a434f6
Fix the 31position test
2015-05-27 21:36:50 +02:00
3456e53a7d
Fix the 30mouse test
2015-05-27 21:31:04 +02:00
1023dd445a
CMakeLists.txt: build tests
2015-05-27 21:30:13 +02:00
11b4d7cb4f
Fix rxvt shifted keys and OOB array access
2015-05-26 21:59:44 +02:00
298bc01ee2
Backport a change from libtermkey
...
"The Ctrl-H byte (0x08) shouldn't automatically become SYM_BACKSPACE"
2015-04-11 18:58:24 +02:00
1ef250cd7e
Miscellaneous little changes
2015-04-11 18:58:14 +02:00
b1a0bca294
Travis CI: Try enabling Coverity scan
2015-02-14 22:54:42 +01:00
28bb257eeb
Add Travis CI configuration
2015-02-14 22:47:52 +01:00
a9c0ea0e07
Fix static linking into a shared library
2015-02-14 22:47:20 +01:00
af193ef483
Go with ncursesw if found
...
It'd be better if the user could choose the exact library, however this
change is actually necessary to support compilation on systems with
only the wide version's headers installed (hello Debian).
2014-11-30 00:37:17 +01:00
e8b9c654af
Mark a TODO
2014-11-30 00:36:18 +01:00
6352428e85
Qt Creator screwed up formatting
2014-11-19 03:37:54 +01:00
788b252407
Update the README
2014-11-19 03:19:13 +01:00
2d777dd67e
Rewrite the mouse API
...
I wasn't aware of the fact that 1000, 1002 and 1003 are mutually
exclusive and turn each other off.
Also now it's not needed to set the protocol, it gets set by default.
2014-11-19 03:19:13 +01:00
5a80bceec9
Make mouse data signed
...
Because urxvt can output zero coordinates, meaning -1.
2014-11-19 00:45:35 +01:00
b8337b5b18
demo-draw.c: app_data_t -> app_context_t
2014-11-19 00:44:53 +01:00
2f8a39ed64
Better support for rxvt input
2014-11-19 00:42:59 +01:00
64188f1fcf
Revert "Don't bother with endianity"
...
It makes iconv() output the byte order mark first...
2014-11-18 22:30:19 +01:00
f1f9a00cda
Add support for setting the mouse protocol
2014-10-22 22:23:01 +02:00
70bcfde021
Add a demo for drawing in curses
2014-10-22 22:22:26 +02:00
86025fcf6f
Rewrite termo_interpret_mouse()
2014-10-18 20:32:45 +02:00
5ca6b7fb42
Don't bother with endianity
...
I hope this is correct... works for little endian (which is the odd one here).
2014-10-18 19:11:44 +02:00
d9b20f7586
Alignment
2014-10-18 19:11:08 +02:00
d70dd571d3
Fix build of the GLib demo
2014-10-14 23:20:55 +02:00
828f03a063
Get rid of most block comments
...
They're rather difficult to maintain.
2014-10-14 21:19:20 +02:00
77878abe90
Renaming continued
2014-10-14 00:24:47 +02:00
e98d9c0fd1
Rename to termo
2014-10-14 00:08:15 +02:00
e330d751a4
Make mode 1015 and 1006 work, prepare for 1005
2014-10-13 23:39:15 +02:00
456093e4ed
Work on mouse support, it's a mess
2014-10-13 01:03:13 +02:00
5692f32bcf
CMake-ify, rename to termkey2 for the time being
2014-10-09 23:47:24 +02:00
3465d5553f
Make the tests work again
...
Our UTF-8 parsing now behaves differently, though.
2014-09-28 13:03:04 +02:00
36bc6cd095
Move to iconv
...
That is the major change in this commit.
Also Ctrl-modified keys now should work in URxvt,
which was surprisingly trivial to achieve.
Coming up next:
- making sure the tests still work
- introducing CMake
2014-09-28 03:59:12 +02:00
cac1f8373b
Less code duplication
2014-09-23 03:05:01 +02:00
c0cc27c555
Introduce isatty(3)
2014-09-23 03:05:01 +02:00
75d3388a35
Introduce bsearch(3)
2014-09-23 03:05:01 +02:00
56f7847ce3
Add -Wextra
2014-09-23 03:05:01 +02:00
b630bf7a5f
WIP: Is mine now (^3^)
...
Seriously though, I've got some issues with how this thing is designed,
as well as with its formatting, and when you add the fact that the
original author wants to merge this thing into his bigger library that
also handles terminal output, which I'll kindly leave to ncurses,
it kind of makes sense for me to do this.
Manpages have been removed as they are going to become obsolete and
they're rather difficult to maintain. If anything, there will be
Doxygen-generated documentation.
The plan is to throw away any direct UTF-8 support and support all uni-
and multibyte character encodings. However some unrelated refactoring
is about to come first.
2014-09-23 03:05:01 +02:00
Paul LeoNerd Evans
7909067ac0
Be a little more lenient in locale env.var. testing for UTF-8ness
2014-03-05 01:23:16 +00:00
Paul LeoNerd Evans
8aa5b7acb5
Unit tests should use .code.codepoint/.sym instead of not .code.number when appropriate
2014-03-05 01:11:45 +00:00
Paul LeoNerd Evans
97e0e9cde2
Try tinfo before ncursesw (partially from c2dd8f247a
) [thanks o11c]
2014-03-04 14:04:11 +00:00
Paul LeoNerd Evans
d775d2a5ae
Make header compatible with C++ (from ddd7b0c5aa
) [thanks o11c]
2014-03-04 13:59:53 +00:00