15:23 < RT|AO> mira: minoca os's bug has been fixed in their side. |
|
More...
|
over 2 years ago
|
pretty straightforward port to Minoca OS by RT • default CC to gcc • minor issues as two tmpfile tests fail, one even with an OS assertion |
|
More...
|
over 2 years ago
|
run the history-ed-* tests under set -e to error out early if ed(1) is missing to make the tests less illegible |
|
More...
|
over 2 years ago
|
act on getrusage errors |
|
More...
|
over 2 years ago
|
this was apparently wrong; use sysconf(_SC_CLK_TCK) even on SerenityOS |
|
More...
|
over 2 years ago
|
so, apparently some OSes have times(3) but not CLK_TCK, the unit in which it is defined (probably because POSIX changed this in 1996 to a sysconf(3)) |
|
More...
|
over 2 years ago
|
add some sanity checks for grep(1), sed(1), tr(1) |
|
More...
|
over 2 years ago
|
SerenityOS has no sigsuspend nor pause so emulate it with poll; from RT |
|
More...
|
over 2 years ago
|
allow building with HAVE_GETRUSAGE=0 when getrusage _is_ defined (rename the function to ksh_getrusage) |
|
More...
|
over 2 years ago
|
initial SerenityOS port by RT |
|
More...
|
over 2 years ago
|
fix compilation if both MKSH_NOPROSPECTOFWORK and MKSH_NO_SIGSUSPEND are defined; spotted by RT |
|
More...
|
almost 3 years ago
|
adapt tests for SkyOS | skyos skyhost 5.0.0 beta i586 unknown unknown skyos running on 'skyos, Perl version 5.008008' |
|
More...
|
almost 3 years ago
|
as we learnt in commitid 10060EF6B0067D2E34A BAFHror(eax,cl) is only conforming ISO C for cl in [1; 31] not [0; 31] (though this hash uses it with cl=8 only anyway) |
|
More...
|
almost 3 years ago
|
drop globally-unused BAFHHost*() macros, they complicate things too much |
|
More...
|
almost 3 years ago
|
fix sh-mode-1 after refactor to exit 0 |
|
More...
|
almost 3 years ago
|
drop use of *int8_t completely |
|
More...
|
almost 3 years ago
|
move mksh_ari_u to var.c to localise the badness; drop its use elsewhere |
|
More...
|
almost 3 years ago
|
• declare BeOS < 6 “Zeta” as having MAJOR issues, not minor ones • RT found that 5.0 “Maui” and older require MKSH_NO_SIGSUSPEND |
|
More...
|
almost 3 years ago
|
default (all) BeOS versions to CC=gcc |
|
More...
|
almost 3 years ago
|
add new selftest-exec-a to test “exec -a newargv0 prog args…” as not only some qemu-user-* failed this, also does BeOS; split sh-mode-1-exec-a off sh-mode-1 and skip on BeOS |
|
More...
|
almost 3 years ago
|
some adjustments by RT for BeOS 6.2 | BeOS zeta 6.2 20392 BePC x86 GenuineIntel Zeta running on 'beos, Perl version 5.008' |
|
More...
|
almost 3 years ago
|
skip two more tests on LynxOS 3.0 |
|
More...
|
almost 3 years ago
|
Syllable Desktop can be reduced to minor issues (fd-cloexec-{1,2}) now tested by RT|AO |
|
More...
|
almost 3 years ago
|
enhance self-tests; might make it easier to spot where the error is |
|
More...
|
almost 3 years ago
|
not ever used uninitialised but GCC 4.1.2 fails to recognise that on syllable syllable 7 0.6 i586 Syllable |
|
More...
|
almost 3 years ago
|
regen |
|
More...
|
almost 3 years ago
|
ISO C is a weird beast; it may be better to not prefix transparently… |
|
More...
|
almost 3 years ago
|
whitelist Xenix for having a usable signal() |
|
More...
|
almost 3 years ago
|
hook up some Xenix stuff, cross-compiling from scosysv at the moment |
|
More...
|
almost 3 years ago
|
do not define _IBCS2, go the <sys/ptem.h> route for struct winsize |
|
More...
|
almost 3 years ago
|