<bslsk05>
ec-/q3asm - Fast assembler for Quake3 Arena QVM (6 forks/14 stargazers)
<bslsk05>
ec-/q3lcc - LCC-based compiler for Quake III Arena QVM (8 forks/15 stargazers)
<bslsk05>
jnz/q3vm - Q3VM - Single file (vm.c) bytecode virtual machine/interpreter for C-language input (64 forks/896 stargazers/GPL-2.0)
<kof673>
it is seemingly really a puts() syscall but it includes a printf() wrapper that calls it
itrsea has quit [Remote host closed the connection]
itrsea has joined #osdev
amj_ is now known as amj
amj has quit [Quit: WeeChat 3.8]
amj has joined #osdev
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Left_Turn has joined #osdev
<_Heat>
linux operating
<zid>
system?
<_Heat>
yes.
<nikolar>
_Heat: yes
<mjg>
linux kernal system
goliath has joined #osdev
<mjg>
i tried watching windows NT videos
<mjg>
fucking guy claimed the success of the OS was due to tech
<mjg>
as opposed to MS being pieces of shit
<mjg>
so i turned it off
<_Heat>
There's a hint of truth to that
urandom__ is now known as urandom
chiselfuse has joined #osdev
<nikolar>
what tech
<nikolar>
what was so much better about it lol
innegatives has joined #osdev
<_Heat>
it didn't crash all the time
<_Heat>
super compatible way back to win 3.1
<_Heat>
(and ms-dos)
<nikolar>
it was already winning when it was crashing all the time
<nikolar>
(win 95 and 98)
<_Heat>
sure
<_Heat>
but still in a limited way
<zid>
w95 was just a skin over the very successful win 3.1 and dos afterall
<nikolar>
i mean it was a bodged dos lol
<nikolar>
yeah exactly zid
<zid>
all the programs everybody was already running didn't just disappear
<zid>
there are *still* point of sale systems running DOS programs for their UI
<nikolar>
indeed
<_Heat>
meanwhile linux guarantees compatible up to... yesterday?
<zid>
Inertia is always the main reason
<nikolar>
_Heat: hey, linux keeps source compatibility
<nikolar>
somtimes
<_Heat>
yeah, except when it doesn't
<zid>
but none of the makefiles, tools, etc do
<nikolar>
exactly
<zid>
there's nothing holistically driving linux to be stable, sadly
<zid>
like on windows
<_Heat>
there is, you just need to pay for it :p
<nikolar>
to be fair, i did get some ancient software to build on modern linux
<_Heat>
BUY
<_Heat>
SLE
<nikolar>
no
<_Heat>
LICENSE
<_Heat>
yeah i mean
<_Heat>
it kind of works sometimes
<zid>
posix is somewhat stable, ish, enough that gcc with a modified makefile might build it still
<_Heat>
but e.g gcc bumping the C/C++ standard ruins it all
<_Heat>
glibc tries to keep the ABI at least
<nikolar>
as we've seen with gcc 15
<nikolar>
where they, for some retarded reason, bumped the standard to c23
<zid>
ehh I *get* it, but it's sucky, mainly the suck is the C23 standard tbh
<zid>
gcc probably *should* push the default standard forward, the problem is that the default standard is crappier than the previous one :p
<nikolar>
yeah
<nikolar>
c standards shouldn't be doing what c23 did
<zid>
or at least, not un-crappy enough to be worth all the breakage
<nikolar>
indeed
<zid>
if in 1992 they made breaking changes and so-on, and we got an actually improved language every now and then? Also fine
<zid>
but we got a breaking standard that added nothing useful, and fixed basically nothing
<nikolar>
zid: but hey, it was for c++ compat
<zid>
phew, thank god
<zid>
I'm constantly accidentally compilgin my C as C++
<nikolar>
i know right
<nikolar>
happens all the time
<zid>
I should start posting defect reports about the "C++ compatibilty"
<_Heat>
what was?
<zid>
int class; doesn't compile
<_Heat>
bool is not for C++ compatibility
<zid>
_Heat: wording in the nullptr rfc
<_Heat>
bool is for modern language welcome to 1998
<zid>
was "for C++ compatibility"
<zid>
and it makes sense right
<zid>
because if it was for fixing C, they'd have just fixed NULL
<zid>
instead of adding some C++
<_Heat>
that's even worse
<nikolar>
what's even worse
<nikolar>
nullptr didn't solve literally anything that fixing NULL wouldn't have
<_Heat>
now gcc bumps the standard and you have breakage all over
<nikolar>
how
<zid>
the breakage wasn't from nullptr
<zid>
it was mainly from bool
<_Heat>
because i guarantee you there are people out there relying on NULL being either 0 or (void *) 0
<nikolar>
what broke a lot of configure scripts was not allowing `main() {}`
<nikolar>
:P
<nikolar>
i think i had to patch a couple of things because of that
<_Heat>
and bool affects the 1% of projects that typedef bool (kernel, firmware, libc)
<zid>
*everyone* relies on it being (void *)0
<_Heat>
NULL affects everything that uses C
<zid>
because it is
<nikolar>
yeah, it's only the linux kernel right
<nikolar>
no biggy
<_Heat>
everyone, except the ones who don't
<zid>
There's literally no implementations anyone uses or cares about not breaking with newer standards
<zid>
that use
<zid>
0
<nikolar>
^^^
<zid>
They're running C89 -std= already
<nikolar>
that's the whole argument
<nikolar>
if something is going to implement c23, it's already doing (void *)0
<_Heat>
you should have made that argument
<zid>
I'd rather the four people relying on 0 null
<zid>
break
<nikolar>
i am not on the comittee
<zid>
than EVERYBODY ELSE
<zid>
I don't know enough C++ to be on the comittee
<_Heat>
no one broke with nullptr
<zid>
(Every single C committee member is a C++ member)
<_Heat>
yeah no shit
<nikolar>
lol
<_Heat>
there's no money is doing C compiler work
<nikolar>
ok
<zid>
so why don't they fuck off breaking things then
<zid>
if there's no money
<nikolar>
don't mess with the frontend then
<nikolar>
leave c be
k0valski18891621 has joined #osdev
<zid>
They apparently did it for shits and giggles, then
<nikolar>
instead of adding c++ junk
<_Heat>
but C sucks
<nikolar>
c++ sucks harder
<zid>
Yes, and it isn't fixed by adding MORE JUNK
<zid>
It's fixed by fixing it
<zid>
If you fix it by adding more junk, you end up with C++
<zid>
and your basic pointer type changes every 2 years
<_Heat>
you cannot fix it
<_Heat>
you know why C++ is the way it is?
<_Heat>
because you cannot fix it
<zid>
Because they add junk to fix junk
<_Heat>
you cannot break backwards compat
<zid>
Because they add junk to fix junk
<_Heat>
you can't fix junk
<zid>
You can, AND THEY DID
<_Heat>
you just can't do it
<_Heat>
who? where?
FreeFull has joined #osdev
<zid>
their rationale for not changing NULL was 'compatibility', then they changed bool
<zid>
which BROKE COMPATIBILITY
<nikolar>
adding nullptr fixed literally nothing
<_Heat>
no one relies on bool being a typedef
<nikolar>
it improved litearlly nothing
<nikolar>
_Heat: some people do
<zid>
except for all the code that relied on it not being a reserved keyword
<zid>
which it now is
<nikolar>
^
<_Heat>
they changed the C23 wording on realloc and it caused a bunch breakage and now the 2 big relevant C library implementations aren't going to follow it
<_Heat>
warra win!
<_Heat>
fixed junk!
<nikolar>
that's because they made realloc retarded
<zid>
not news: They don't fix things they just break things
<nikolar>
another c23 fail
<_Heat>
GOOD
<_Heat>
THE WHOLE OF C IS
<zid>
You're literally making our argument for us
<nikolar>
lol
<zid>
then claiming you won
<zid>
like a pigeon shitting on the chess board
<nikolar>
kek
<_Heat>
god.
<nikolar>
can't say i've heard of that expression before
<kof673>
i've heard it before, but it seems more common now actually :D slowly gaining
itrsea has quit [Remote host closed the connection]
itrsea has joined #osdev
<Ermine>
_Heat: i can run 90s software flawlessly on sle?
<_Heat>
no
<Ermine>
bummer
<_Heat>
you can run whatever ran flawlessly when $SLE_VERSION released
<Ermine>
that's kinda self-evident
<nikolar>
only kinda
<_Heat>
it is certainly not true for any !stable distro
<Ermine>
vmware crashed :(
<nikolar>
nice
<nikolar>
don't use it :P
<Ermine>
why
<nikolar>
because it's vmware
<Ermine>
that only makes me want to use it
Turn_Left has joined #osdev
averymt has joined #osdev
Left_Turn has quit [Ping timeout: 260 seconds]
<nikolar>
sure, go ahead using a shitty vm, it's your computer
averymt has quit [Quit: Quit]
averymt has joined #osdev
<_Heat>
Ermine: dalias hasn't cut a release for so long I had to backport 60 commits
<Ermine>
nikolar: it's not shitty
<Ermine>
_Heat: geez
<nikolar>
you just said it crashed
<nikolar>
qemu never crashed for mw
<nikolar>
*me
<nikolar>
nor did virtual box for that metter lol
<nikolar>
*matter
<nikolar>
i can't words
<_Heat>
good news, onyx is gcc 15 approved
<nikolar>
congrats?
<zid>
you mean gnu23 approved?
<_Heat>
no, i force the std
<_Heat>
i'm not an idiot
<Ermine>
qemu doesn't even run windows in the first place
<zid>
my code should all be gcc23 clean because I never use nullptr or bool as keywords, but I've not tested it