Kugel Fuhr
efa2020d93
Improved/fixed the time() function:
...
- When the underlying clock_gettime function returned an error, the value
returned via *timep was wrong.
- Reduced code size by 7 bytes.
- Don't suck in ldeaxi.
2024-09-12 09:14:57 +02:00
Bob Andrews
31a0d5cc40
Merge pull request #2503 from colinleroy/master
...
Optimize stpcpy's size and speed
2024-09-08 16:17:14 +02:00
coronax
838c8b48b7
Set the clock id to CLOCK_REALTIME when calling clock_gettime.
...
Previously, time() allocated stack space for the clock id argument, but didn't actually set a value.
2024-09-07 23:27:54 -05:00
Colin Leroy-Mira
55d3a6ea39
Optimize stpcpy's size and speed
2024-09-07 21:24:53 +02:00
Sven Michael Klose
aed94d2dae
Fix code style.
...
Have type, function name and argument declaration on a single line.
2024-07-16 01:33:48 +02:00
Sven Michael Klose
581b79e0b9
Add stpcpy().
...
Like strcpy() but returning pointer to ending zero of copied string.
2024-07-07 14:04:49 +02:00
Colin Leroy-Mira
a823d90082
Separated versions
2024-04-19 07:57:47 +02:00
Colin Leroy-Mira
71d82ab5d9
Use common naming scheme for tolowerdirect
2024-03-22 17:19:26 +01:00
Colin Leroy-Mira
0c681b42ef
Factorize to save 20 bytes
2024-03-19 18:07:17 +01:00
Colin Leroy-Mira
b5d259bafb
Implement strcasestr as part of strstr
2024-03-18 19:52:04 +01:00
Colin Leroy-Mira
82165c1a77
Implement strcasestr
2024-03-18 19:51:50 +01:00
Colin Leroy-Mira
1fab179cb4
a BIT of 65C02 optimisations
...
Use BIT immediate instead of AND when reloading A is required
afterwards.
Add an fread unit test as the optimisation touches fread.
Sprinkle a few zero page indexed while we're at it.
2024-02-02 19:45:49 +01:00
Colin Leroy-Mira
afd8ee627e
Remove useless branching code in fgets
2024-01-30 19:52:48 +01:00
Colin Leroy-Mira
2b97735d5d
Factorize file pointer check
2024-01-30 19:40:28 +01:00
Colin Leroy-Mira
1a5a7d67a7
Rewrite fputc in assembly
...
-36 bytes, -12% cycles
2024-01-30 17:19:13 +01:00
Colin Leroy-Mira
ba36071022
Rewrite fputs in assembly
...
-28 bytes, -1% cycles
2024-01-30 17:12:52 +01:00
Bob Andrews
8682095f9f
Merge pull request #2384 from colinleroy/asm-realloc
...
Rewrite realloc in asm
2024-01-27 00:28:17 +01:00
Colin Leroy-Mira
aa6f850b8d
Rewrite gets in assembler
...
+19 bytes if used alone, because it pulls in fgets, but as code is
factorized, -128 bytes in programs using both fgets and gets.
2024-01-26 20:29:17 +01:00
Colin Leroy-Mira
a8b870555e
Rewrite realloc in asm
...
-80 bytes, -39% cycles
2024-01-26 20:25:59 +01:00
Bob Andrews
65937684a0
Merge pull request #2380 from colinleroy/asm-fgetc
...
Rewrite fgetc in asm
2024-01-26 20:12:41 +01:00
Bob Andrews
ba6f9f5d3f
Merge pull request #2381 from colinleroy/asm-fgets
...
Rewrite fgets in asm
2024-01-26 19:36:57 +01:00
Colin Leroy-Mira
476591e8b7
Rewrite fgetc in asm
...
-82 bytes, -20% cycles
2024-01-25 09:13:39 +01:00
Bob Andrews
ee3a542539
Merge pull request #2383 from colinleroy/pmemalign-size-opt
...
Optimize -36 bytes out of posix_memalign
2024-01-24 21:26:50 +01:00
Bob Andrews
9ba270d97b
Merge pull request #2378 from colinleroy/fix-overflow-in-malloc-and-realloc
...
Fix malloc and realloc overflow
2024-01-24 21:15:31 +01:00
Colin Leroy-Mira
b87df9e1c6
Optimize -36 bytes out of posix_memalign
...
And add a unit test
2024-01-24 10:23:53 +01:00
Colin Leroy-Mira
1f820d0ae8
Rewrite fgets in asm
...
-104 bytes, -1% cycles
2024-01-24 10:16:26 +01:00
Colin Leroy-Mira
5d49fde788
add a return -1 helper
2024-01-24 09:58:09 +01:00
Colin Leroy-Mira
3e01ac9b04
Fix malloc and realloc overflow
...
If user requests a size >= 65532, adding the heap admin size
overflows size. Fixes #2358 .
2024-01-24 09:54:54 +01:00
Colin Leroy-Mira
10282a9b74
Rewrite asctime() in assembler (-7 bytes)
2024-01-18 09:09:38 +01:00
Colin Leroy-Mira
03d5e5fba0
Rewrite mktime in assembly
...
-415 bytes,
-39% cycles,
Unit test expanded to cover more cases (there was a bug in 2100 before!)
2024-01-18 09:09:38 +01:00
Colin Leroy-Mira
348a9048b7
Convert _time_t_to_tm to asm
...
46 bytes size gain, -8% cycles on the unit tests
2024-01-18 09:09:31 +01:00
Bob Andrews
2127778239
Merge pull request #2336 from colinleroy/optimize-strdup
...
Optimize strdup
2024-01-08 22:13:43 +01:00
Colin Leroy-Mira
df193c0947
Rework time functions a bit
...
- mktime: Work unsigned as time_t's type implies (shifting Y2K38 bug to 2106)
- mktime: Add unit tests
- gmtime/localtime: factorize
- gmtime/localtime: Add unit tests
- mktime/gmtime/localtime: Size optimisation (-130 bytes wrt master)
- mktime: Speed optimisation (from 23M cycles on the unit test to 2M)
2024-01-08 18:24:13 +01:00
Colin Leroy-Mira
3a439e0e1b
Little strdup optimisation
...
-12 bytes on disk
-20 cycles per strdup
-6 cycles per strlen called from strdup
2024-01-07 14:25:06 +01:00
Colin Leroy-Mira
a90aa57510
Address Oliver's comments
2023-09-07 16:48:56 +02:00
Colin Leroy-Mira
dfe7562f76
Add ntohs, htons, ntohl, htons.
2023-09-07 07:03:34 +02:00
Christian Groessler
fedfc3443d
fix for issue #1973 (printf("%c", '\0') doesn't _write '\0' to fd 0)
2023-02-24 02:12:19 +01:00
mrdudz
5bda57de87
add extra underscore to _bzero, add a test for bzero/memset
2022-08-29 23:20:48 +02:00
mrdudz
3b03a96375
add underscores to _poserror and _stroserror
2022-08-29 21:06:48 +02:00
mrdudz
89c08dc6d4
extra underscores for _randomize and _swap
2022-08-29 20:52:35 +02:00
mrdudz
b09024aa32
add extra underscore to heap stuff
2022-08-29 19:55:48 +02:00
mrdudz
df4b6f9d14
remove extra spaces
2022-08-28 22:37:33 +02:00
mrdudz
54aff47513
fix some errno related stuff, "make all" works again
2022-08-28 22:09:44 +02:00
mrdudz
2dabb65ee0
fix errno related underscores in all libsrc/*.s files
2022-08-28 21:52:53 +02:00
mrdudz
f70020a2b8
use __sig_ign and __sig_dft instead of _sig_ign and _sig_dft
2022-08-28 20:18:34 +02:00
mrdudz
9246775ebe
use __afailed instead of _afailed
2022-08-28 20:08:13 +02:00
mrdudz
8e5f6b822d
fix setjmp.h
2022-08-28 20:05:11 +02:00
polluks2
7a93d3c707
Shorter isascii
2022-08-04 00:25:04 +02:00
Bob Andrews
6239fbe18c
Revert "rename '_poserror' to '__poserror'"
2022-07-24 18:14:21 +02:00
Christian Groessler
758d9466c0
rename '_poserror' to '__poserror'
...
The old name could be non-conforming to the C standard. The new name
is definitely in the "implementation name space".
See issue #1796 .
2022-07-24 00:33:56 +02:00
