Commit graph

30 commits

Author SHA1 Message Date
fef
b6659a321d
string: add static NSTR_DEFINE initializer macro 2021-07-18 21:00:07 +02:00
fef
30954cd078
string: add tests 2021-07-16 21:47:18 +02:00
fef
0f3c11110c
string: implement nstrdup 2021-07-16 14:43:26 +02:00
fef
66032c95f2
string: sanitize nstrcmp return value on error 2021-07-16 14:43:02 +02:00
fef
7169a434bf
utf: add utf8_ncheck 2021-07-16 13:59:11 +02:00
fef
6ab34d0893
string: rename _capacity field to _size 2021-07-16 13:35:32 +02:00
fef
f27d00a4a2
utf: fix name ambiguity of utf8_chrsize 2021-07-16 00:12:10 +02:00
fef
aae039ae29
string: nstrcat refactor 2021-07-15 23:14:55 +02:00
fef
2fbbfc61c0
string: use memcpy instead of strncpy in nstrmul 2021-07-15 23:10:57 +02:00
fef
8c591796fb
string: always call strcmp() in nstrcmp()
Some people seem to depend on the unstandardized
behavior of some (notably glibc) implementations
of strcmp() where the arithemtic difference
between s1 and s2 is returned, not just any
positive or negative number.  This is explicitly
undocumented in libneo as well, but still doing
it won't hurt either.
2021-07-15 23:07:00 +02:00
fef
7d64438f70
string: fix off-by-one error in leftpad 2021-07-15 23:00:17 +02:00
fef
900bce9b8e
string: add nnstr string constructor 2021-07-15 22:58:15 +02:00
fef
c71e36e7c8
error: fix memory leaks and other bugs
It was really late when i wrote this okay
2021-07-15 22:50:33 +02:00
fef
b011d5feb5
string: remove unnecessary refcount update
Passing the argument to a pure function is
essentially a "borrow", and the caller has had to
make sure the refcount doesn't drop to zero anyway
2021-07-15 14:16:06 +02:00
fef
5aed426864
string: add more utility functions 2021-07-15 01:47:04 +02:00
fef
43da1220ee
error: fix brainfart 2021-07-14 22:58:49 +02:00
fef
b010638dd1
string: add creationg and char access functions 2021-07-14 22:24:07 +02:00
fef
c70068f1b8
utf: fix some more bugs
It's beginning to actually work, wow
2021-07-14 22:19:33 +02:00
fef
06d538bc4a
utf: use unsigned types
Turns out the people warning about doing signed bitshifts were right
2021-07-14 21:15:07 +02:00
fef
befc18f8ff
add C++ support
This is so not a good idea but i'll need it for Catch2
2021-07-14 20:45:22 +02:00
fef
54869dcd6c
utf: use correct format specifier 2021-07-14 16:42:15 +02:00
fef
30a29150c3
implement UTF-8 conversion
i have no idea whether this works, the only thing i have is PTSD
2021-07-14 16:30:19 +02:00
ec5ed1e352
error: refactor the whole thing
a.k.a. make it more complicated.
2021-07-12 03:06:51 +02:00
fef
7df80d8cd1
add malloc wrapper 2021-07-11 15:06:42 +02:00
4dbc77845a
add error handling system 2021-07-11 02:11:58 +02:00
957f5e2a35
add refcounts 2021-07-11 01:44:59 +02:00
fef
7eed96edcf
remove jemalloc and fall back to libc
completely replacing libc while still supporting all POSIX compliant
systems is probably gonna take a little more effort than just one
developer.  Which is a bummer, but you gotta work with the cards you're
dealt, i guess.
2021-07-10 20:55:44 +02:00
a73ea804b5
implement generic syscalls 2021-07-09 16:56:23 +02:00
11548e74ba
integrate jemalloc into cmake build
well that was painful
2021-07-09 14:07:13 +02:00
fef
673ed9bfb8
initial commit uwu 2021-07-09 00:04:44 +02:00