Kovid Goyal
a7d9c63a4e
wcwidth() for not assigned code points should be -4
...
That way, wcswidth() computes their width as 1, which matches how they
are rendered
2018-06-14 10:27:42 +05:30
Kovid Goyal
61dd52b50f
Ignore the non-characters from the unicode standard in addition to ignoring the control characters
2018-06-14 10:20:13 +05:30
Kovid Goyal
ad1109b6fe
Hints kitten: Overhaul making it more intelligent
...
It now uses information about the geometry of the screen.
So URLs that stretch over multiple lines are detected even if there is a
hard line break in between (some programs like mutt generate these).
2018-05-19 14:31:00 +05:30
Kovid Goyal
0b93b85cf2
Dont use case range in names.h as it prevents compilation with Visual Studio
2018-05-01 11:27:10 +05:30
Kovid Goyal
f7001ea068
Fix character names for control characters not being read from unicode database
...
Also allow unicode_names.c to be compiled with python 2 so I can re-use
it in calibre.
2018-05-01 10:13:58 +05:30
Kovid Goyal
0b99bb534f
Unicode input: When searching by name search for prefix matches as well as whole word matches
...
So now hori matches both "hori" and "horizontal". Switched to a
prefix-trie internally.
2018-04-24 07:45:20 +05:30
Kovid Goyal
9d67198ff9
Update path in gen-wcwidth.py
2018-04-23 21:25:33 +05:30
Kovid Goyal
0c3257b6b9
Start work on url hints kitten
2018-02-14 11:01:17 +05:30
Kovid Goyal
8c18486836
Module with all the data for unicode entry by character name
2018-02-09 19:56:25 +05:30
Kovid Goyal
ff2e5b3966
Avoid unnecessary calls to mark_for_codepoint
2018-02-06 11:23:39 +05:30
Kovid Goyal
9c874f6685
Add support for the unicode emoji presentation variation selector
2018-02-06 09:52:55 +05:30
Kovid Goyal
fbe4d036d8
Have wcwidth() return 0 for marks instead of -1
...
Since kitty always treats marks as combinig chars, this allows us to
remove a few unnecessary branches
2018-02-05 10:06:05 +05:30
Kovid Goyal
b9857f9499
Only consider emoji characters with emoji presentation to have width two
...
Fixes #308
2018-02-05 09:21:13 +05:30
Kovid Goyal
fc7ec1d3f7
Get rid of the option to use the system wcwidth
...
The system wcwidth() is often wrong. Not to mention that if you SSH into
a different machine, then you have a potentially different wcwidth. The
only sane way to deal with this is to use the unicode standard.
2018-02-04 21:02:30 +05:30
Kovid Goyal
32632264ee
Mapping that can be used to store unicode mark symbols in only two bytes
2018-01-18 16:06:07 +05:30
Kovid Goyal
5faa649452
Drop the dependency on libunistring
2018-01-18 00:09:40 +05:30
Kovid Goyal
2ddc7e8c80
Have all marks return a wcwidth of -1
...
Mirrors recent change to onsidering all marks combining chars instead of
only the marks with a non-zero combining class
2018-01-17 23:25:59 +05:30
Kovid Goyal
c2ad0607b8
Fix trailing whitespace in generated unicde data files
2017-12-20 22:44:20 +05:30
Kovid Goyal
d1282b9f55
Add a script to generate wcwidth as well
...
Generated function is more efficient than the implementation from
wcwidth9 and also makes it easy to update when the unicode standard
changes.
2017-12-20 16:18:05 +05:30
