<KIMI>
hi, is someone have example for "reader interface" i don't understand how to use to read a file line by line
<ifreund>
KIMI: sounds like you might want Reader.readUntilDelimiterAlloc() or similar using '\n' as the delimiter
<KIMI>
i cann it directly?
<KIMI>
call
<KIMI>
i don't understand there is no Reader.readUntilDelimitedAlloc
<KIMI>
it is hard to read a file
<KIMI>
xD
<ifreund>
look in std/io/reader.zig where the Reader interface is defined
<ifreund>
line #125
<KIMI>
but i don't understand it is inside a function inside struct
<KIMI>
is there an example of code to use it?
<ifreund>
if you're asking how to get a Reader from a File, just use my_file.reader()
<KIMI>
i need read file line by line
<KIMI>
what is my_file?
<KIMI>
i need code example to understand how to do it
<ifreund>
whatever file you want to read?
<KIMI>
what is type of my_file in your example?
<ifreund>
std.fs.File
<KIMI>
can you show me full example i am lost xD
<ifreund>
const f = try std.fs.cwd().openFile("foo.txt", .{}); const line = try f.reader().readUntilDelimiterAlloc(allocator, '\n', 1024);
<KIMI>
thanks
<KIMI>
how to get next line?
<ifreund>
call readUntilDelimiterAlloc() again, you probably want to use a while loop
<KIMI>
why it can't find file in parent folder?
<ifreund>
cwd() gives you the current working directory when you run the program, that's where it looks
<KIMI>
how to read in a parent folder?
<KIMI>
this is hard :(
<andrewrk>
it's a legitimately hard problem and the API is not yet stable. the functions that exist were only added in order to unblock a different project. the API hasn't seen any intentional effort spent on it yet
<ifreund>
KIMI: you can pass a relative path, i.e. "../foo.txt"
<KIMI>
that is what i write ../data/font.fnt but i get: Unable to dump stack trace: FileNotFound
cole-h has quit [Ping timeout: 240 seconds]
<g-w1>
iirc it (".." in file path) doesn't work on windows
<KIMI>
yes i am on windows
<KIMI>
what should i do?
<g-w1>
you could do openFileAbsolute if you know the absolute path
<KIMI>
but if i move exe it wont work anymore
<KIMI>
if path change
a93 has joined #zig
<KIMI>
i can use fopen?
<KIMI>
i found C example
<KIMI>
i will keep on C i think for the moment i need to learn it properly
<g-w1>
I think std.fs.realpath gives you the absolute path, im not on windows so I can't really test it. sry. fopen will probably work.
<KIMI>
but i don't like headers :(
<andrewrk>
marler8997, I eliminated one thing as a possible cause of failure
<andrewrk>
https://clbin.com/brmaX working binary is at the top, crashing binary is next. you can see the working binary has a version number on the dlopen function. but in the failing stack trace, it does go into the correct version after all
<andrewrk>
this plus the fact that musl-based void linux is having a crash (somewhere else) seemingly due to lack of running start code, leads me to believe that is the next place to look
<KIMI>
why i cant use string on fopen? i get an error msg error: expected type '[*:0]const u8', found '[]const u8'
<ziguana[m]>
the TLS part was surprising to me in particular, didn't know that it needs special constructors and destructors
<KIMI>
i will stick with C i think sorry :(
<andrewrk>
KIMI, try again closer to 1.0 :)
<KIMI>
when is it released?
<andrewrk>
years into the future
<KIMI>
i will try again later someone on telegram told me to try zig for easier C, but i'm not very good even in C
kristoff_it has joined #zig
msingle has joined #zig
<andrewrk>
marler8997, sorry to keep pinging you, I'm using you as a rubber duck and to give little progress updates
<andrewrk>
next thing I discovered: the crash is in glibc main thread initialization code. it's doing pd->tid = INTERNAL_SYSCALL_CALL (set_tid_address, &pd->tid);
<andrewrk>
but the address it is using comes from THREAD_SELF and it ends up being 0x2d1 which is an invalid pointer. So something is wrong with THREAD_SELF
<andrewrk>
ok so the fs register is not set up properly
<andrewrk>
hmm I bet LemonBoy would know the answer here
<andrewrk>
zig sets the fs register in start code, but it looks like glibc did not like the value we set it to
xentec has quit [Ping timeout: 240 seconds]
radgeRayden has joined #zig
<andrewrk>
I see the problem. here are the last few syscalls the glibc dynamic loader does before entering zig's _start code, followed by the 1st syscall that zig's _start code does: https://clbin.com/9z8ok
<andrewrk>
basically zig's start code is clobbering the fs register value from glibc's start code
<andrewrk>
since we didn't actually link with libc, zig thought we were in charge of the start code
<andrewrk>
s/the start code/initializing TLS/
CommunistWolf has quit [Quit: No Ping reply in 180 seconds.]
CommunistWolf has joined #zig
<andrewrk>
yooo I got a window
<ziguana[m]>
git push when
<andrewrk>
it's with this patch to zig, which I can't just push, I need to think about how to make this work for normal use cases too: https://clbin.com/P0DND
<ziguana[m]>
cool! is there an issue where I can catch up with what this is for?
<andrewrk>
I'm planning to do a nice blog post on this soon
<ziguana[m]>
sweet, thanks
<andrewrk>
what distro do you use?
<g-w1>
are there any other types of errors besides @compileLog that don't stop analysis?
<ziguana[m]>
the glorious nixos
<andrewrk>
you wanna try my binary?
<ziguana[m]>
yeah, let me get to a pc
<g-w1>
could I try it? I use arch.
<ziguana[m]>
ok, got a nix-shell or something?
<g-w1>
I also have nix though
<ziguana[m]>
man, nvm i'm on llvm10 still for zig, sorry
<andrewrk>
ziguana[m], that's the thing, you don't need anything, you can run it without installing anything
<ziguana[m]>
i'm kind of scared about this project tbh, aren't you going to end up reimplementing a good portion of dynamic linking? that'd be a can of worms and a half...
<andrewrk>
run that in a terminal so you can ctrl+c it and tell me what it does instead of opening a window like it is supposed to
<ziguana[m]>
it just werkz holy hell
<andrewrk>
:-O
<g-w1>
is it supposed to work on any linux?
<andrewrk>
that's the goal but there are some TODOs
<andrewrk>
I'm about to try a musl-based linux
<g-w1>
yeah i get illegal instruction
<ziguana[m]>
you on x86-64, right?
<g-w1>
yes
msingle has quit [Ping timeout: 240 seconds]
<andrewrk>
g-w1, I need to get master branch patched up so that this doesn't depend on my local dirty tree, and then if you're interested to join the effort, you could try building from that zig-window repo and you could help figure out what patches might be needed to make it additionally work on your system
hlolli_ has joined #zig
<ziguana[m]>
looking forward to that blog post, i'm really curious about what exactly you did
<ziguana[m]>
looks like manually dlopen'ing x11 libs and doing the vulkan loading too
<andrewrk>
I am curious about where that illegal instruction is. My best guess is UBSAN found a bug in Vulkan-Loader
<g-w1>
I am interested! This seems like a very good learning oppurtunity for me to learn about lower level stuff. I am using a 2013 cpu so maybe ur cpu just has more instructions
<andrewrk>
oh!
<andrewrk>
hang on yeah that's a native build - let me change to baseline
<andrewrk>
:D
<ziguana[m]>
too much mtune?
<andrewrk>
when you compile for the native CPU, zig takes advantage of the native CPU features. you have to intentionally cross compile to do something different
hlolli__ has quit [Ping timeout: 272 seconds]
<andrewrk>
all I'm doing differently now is adding `-target x86_64-linux` to the build, which will then select the "baseline" CPU features
<andrewrk>
we need to detect what the proper soname for libdl.so is. right now it is hard coded to libdl.so.2
<andrewrk>
ok so basically the patch to zig we need is to be able to disable the call to initStaticTLS() in start.zig. or we need to have it runtime detect that it does not need to run that code
<ziguana[m]>
does the static TLS collide with something?
hlolli_ has quit [Ping timeout: 272 seconds]
<ziguana[m]>
or is it just that register?
<ziguana[m]>
ok, i can wait for the blog post
msingle has joined #zig
<Snektron>
wowie, ziguana has come to life
<waleee-cl>
a friend running slackware were 'meh' about testing the demo binary
raggi has joined #zig
<andrewrk>
yeah that's fair
<andrewrk>
I mean the source is available if they don't trust it
<Snektron>
andrewrk: your binary disconnects my monitor
<andrewrk>
o.O
<Snektron>
Its not your fault, i noticed this on my own example as well, so at least i know it didn't do anything wrong especially
<Snektron>
or we made the same mistake, which is entirely possible
<Snektron>
I thought the issue was already fixed since it didn't trigger on my own example
<Snektron>
basically my monitor turns black and then tells me there is no signal
msingle has quit [Ping timeout: 246 seconds]
<Snektron>
not really sure what to do about it...
<Snektron>
Now it works fine (glibc void)
<andrewrk>
what was it before?
<Snektron>
Same OS, i just rebooted
<Snektron>
I suspect some other application interacts with it somehow, or maybe my driver is somehow unstable
<Snektron>
im not sure how to trigger it
<waleee-cl>
soneone else running musl void also got a segfault
<andrewrk>
I'm able to repro that
<andrewrk>
will debug it soon
<andrewrk>
gotta go, see ya
radgeRayden has quit [Ping timeout: 272 seconds]
xackus__ has quit [Read error: Connection reset by peer]
xackus has joined #zig
xackus_ has joined #zig
xackus has quit [Read error: Connection reset by peer]
xackus__ has joined #zig
xackus_ has quit [Ping timeout: 240 seconds]
frmdstryr has quit [Ping timeout: 240 seconds]
xackus has joined #zig
xackus__ has quit [Ping timeout: 256 seconds]
layneson has quit [Ping timeout: 246 seconds]
<marler8997>
so if I understand correctly, we need some way to know whether or not we were loaded by ld, so we can know whether or not to initialize tls?
KIMI has joined #zig
<KIMI>
hi, i want to try again, how to convert [*]const u8 in to [*:0]const u8?
<KIMI>
i got [*]const u8 by doing path.ptr (path is a []const u8)
<KIMI>
i can't give up this easy i need finish it lol
<g-w1>
try std.cstr
<KIMI>
how to use it?
<KIMI>
how it is with llocator
<KIMI>
oh
<g-w1>
std.cstr.addNullByte
<KIMI>
one question why it is different than on C? on C i can just "the path" without 0 at the end
<KIMI>
is on C automatic?
<KIMI>
wit haddNullByte it work thanks
<KIMI>
it is possible to not have to use allocator and make it transparent like in C?
<KIMI>
it work fine too if i put the path directly like in the example
<KIMI>
i have my code on a function like thjis fn load(a: *Allocator, path: []const u8) !FontData
<KIMI>
i think my message is a confusing because i don't explain all, i call load(allocator, "the path") <= how to make it so this is directly passed into fopen(path) like i would do with fopen("the path")
<KIMI>
i ask this because i don't think i should allocate just to add a 0
<KIMI>
ok i read api of fopen require null terminated string since there is no size parameter, that is why it ask [:0], so i need make my function too that require [:0] anyways
<KIMI>
[:0] is what i needed all from the start xD
<pixelherodev>
KIMI: the null byte isn't automatic in Zig; it's only needed when interacting with C code. The default "string" type in Zig uses a length tracker instead of a null byte
kristoff_it has quit [Ping timeout: 264 seconds]
<KIMI>
thanks marler8997
<KIMI>
i could have never figured that code by myself
<marler8997>
you'll get there eventually if you keep at it
<pixelherodev>
^
<KIMI>
what i put as depth?
<pixelherodev>
(Positive) experience is the key determiner for code quality
<marler8997>
that's the number of ".." you want
<pixelherodev>
Of course, if most of the code you've worked with is bad, you'll learn to write bad code :(
<marler8997>
so long as you are vigilant with self evaluation and keep yourself to high quality standards you will get better
<pixelherodev>
Yeah, though knowing what quality is can be difficult if you haven't seen it :)
Jeanne-Kamikaze has quit [Quit: Leaving]
marnix has joined #zig
marnix has quit [Read error: Connection reset by peer]
marnix has joined #zig
ur5us_ has quit [Ping timeout: 264 seconds]
waleee-cl has quit [Quit: Connection closed for inactivity]
lucid_0x80 has joined #zig
supercoven has joined #zig
marnix has quit [Ping timeout: 246 seconds]
marnix has joined #zig
xackus has quit [Quit: Leaving]
xackus has joined #zig
a_chou has joined #zig
a_chou has quit [Client Quit]
decentpenguin has quit [Read error: Connection reset by peer]
decentpenguin has joined #zig
marnix has quit [Ping timeout: 256 seconds]
marnix has joined #zig
cole-h has joined #zig
osa1_ has joined #zig
KIMI has quit [Remote host closed the connection]
ur5us_ has joined #zig
<andrewrk>
marler8997, the tls initialization problem is solved in the most recent commit I pushed to zig master branch (assuming it passes CI)
<marler8997>
checking
<andrewrk>
"All checks have failed" => oh.
<andrewrk>
ugh that's just the stack traces test needing line numbers of start.zig to be in the test cases
<andrewrk>
marler8997, hmm I think it will be interesting to see what failed. perhaps something to do with X11?
gpanders has joined #zig
lucid_0x80 is now known as suskun
nyaayaya has quit [Remote host closed the connection]
nyaayaya has joined #zig
nyaayaya has quit [Remote host closed the connection]
nyaayaya has joined #zig
cole-h has quit [Quit: Goodbye]
mitchk has joined #zig
hlolli_ has joined #zig
mattmurr___[m] has joined #zig
<mitchk>
HI All, I'm having some trouble. Building on a brand new fedora 33 install. I've installed clang-devel,llvm-devel,lld-devel all 11.x. When I run "cmake .." I get an error "Could NOT find clang (missing: CLANG_LIBRARIES) " Is there a env variable I should set to define that?
<mitchk>
Build the zig project that is.
<dutchie>
andrewrk: i also see VulkanInitializationFailed
marnix has quit [Read error: Connection reset by peer]
marnix has joined #zig
<g-w1>
regarding my bug on debian where it couldn't find the .so file for static-window, /lib64 is just ld-linux-x86-64.so.2 (only thing in it) so anything it ties to load from there (libdl.so.2) just doesn't exist. libdl is here for me on debian /usr/lib/x86_64-linux-gnu/libdl.so.2
marnix has quit [Ping timeout: 272 seconds]
marnix has joined #zig
marnix has quit [Read error: Connection reset by peer]
hlolli_ has quit [Remote host closed the connection]
zugzwang_ has quit [Remote host closed the connection]
zugzwang_ has joined #zig
msingle has quit [Ping timeout: 256 seconds]
shynoob has quit [Ping timeout: 240 seconds]
zugzwang__ has joined #zig
zugzwang__ has quit [Client Quit]
shynoob has joined #zig
zugzwang_ has quit [Ping timeout: 260 seconds]
<shynoob>
when will zig be production ready?
<shynoob>
sorry I mean when will zig 1.0 be released?
swills has quit [Quit: swills]
swills has joined #zig
supercoven has joined #zig
supercoven has quit [Max SendQ exceeded]
supercoven has joined #zig
supercoven has quit [Max SendQ exceeded]
<karchnu>
shynoob: I guess we will have to wait until the self-hosted compiler will be ready, and when API will be stable and generic for all architectures.
<karchnu>
This could be years from now.
supercoven has joined #zig
supercoven has quit [Max SendQ exceeded]
swills has quit [Client Quit]
supercoven has joined #zig
supercoven has quit [Max SendQ exceeded]
supercoven has joined #zig
supercoven has quit [Max SendQ exceeded]
<shynoob>
uh oh
<karchnu>
Maybe this will be sooner than that, self-hosted compiler wasn't really a thing a few months ago and now it is still experimental but in the 0.7.0 release.
<karchnu>
I bet that the zig language and most of the standard library will be (kinda) stable in less than a year.
Akuli has joined #zig
donniewest has joined #zig
<wilsonk>
quit
<fengb>
Don't quit us :o
radgeRayden has quit [Read error: Connection reset by peer]
<shynoob>
lol
waleee-cl has joined #zig
xackus has quit [Read error: Connection reset by peer]
xackus has joined #zig
<marler8997>
andrewrk, I ran with "xtrace" which said Authentication Rejected: None of the authentication protocols specified are supported and host-based authentication failed
<mq32>
karchnu, shynoob: i guess that zig can be "production ready" in roughly 2 years from now on for the major platforms (x86_64, arm64, linux,windows, mac)
dumenci has joined #zig
suskun has quit [Ping timeout: 256 seconds]
chowder has joined #zig
<marler8997>
andrewrk, scratch my last response (was running with xtrace from glibc). actual X11 xtrace doesn't show much, I see setup/CreateWindows and a couple InternAtom comands
chowder has quit [Quit: Leaving]
TheLemonMan has joined #zig
mq32 is now known as ikskuh
filpAM has quit [Remote host closed the connection]
Guest59 has joined #zig
nrdmn0 has quit [Quit: Ping timeout (120 seconds)]
nrdmn0 has joined #zig
shynoob has quit [Quit: Leaving]
msingle has joined #zig
hnOsmium0001 has joined #zig
Akuli has quit [Remote host closed the connection]
<marler8997>
it also fails...which is good I guess :)
<marler8997>
vkEnumeratePhysicaDevices fails
<marler8997>
could be virtualbox doesn't have a GPU
<marler8997>
I have an ubuntu machine with a monitor attached I can try as well
<marler8997>
oh...looks like the ubuntu loader doesn't support the "--preload" parameter
<g-w1>
same thing on debian that happened to me
<andrewrk>
same thing on WSL
<marler8997>
I think LD_PRELOAD is more widely supported
<fengb>
Hasn’t 1.0 been “2 years from now” for about 5 years now? :P
<andrewrk>
for at least 2 years yeah haha
<marler8997>
probably better to use LD_PRELOAD, then if we want, we can remove the values we added on the second invocation
sord937 has quit [Quit: sord937]
<TheLemonMan>
andrewrk, yo, have you seen my comment on the TLS changes?
<andrewrk>
not yet! looking for it now
<g-w1>
I am a little confused. Wasn't the project to compile it into a static binary? Why is there a dynamic linker being used? Is it being used to make the static bianry and then run it?
<andrewrk>
TheLemonMan, ah. yeah I didn't really think this through very well
skuzzymiglet has joined #zig
<TheLemonMan>
ABI compatibility is a huuuge problem
<andrewrk>
want to brainstorm how to get further on the graphics window problem? we made more progress yesterday, and have a binary that runs on a few different distros: https://andrewkelley.me/temp/static-window4
<andrewrk>
it's almost as if we would need builtin.link_libc to be a runtime value
<andrewrk>
g-w1, if I recall correctly, not because of any sensible reason, just because of accidental complexity in the way it is set up
<ifreund>
yeah, that's my understanding as well
<g-w1>
ok
Jeanne-Kamikaze has joined #zig
<marler8997>
I should have a PR for LD_PRELOAD soon
<andrewrk>
TheLemonMan, we *could* have official support for this use case, and have something like std.builtin.runtime_link_libc (comptime value - if true a different, runtime value would tell whether this was the static or dynamic execution). in this case thread.zig would use this mechanism to take the pthread path
<andrewrk>
it would add more complexity to the std lib
<andrewrk>
but this is a pretty obviously valuable use case
<TheLemonMan>
more is an understatement, we're already drowning in complexity
<andrewrk>
valid concern
<andrewrk>
what do you think, should we give up on this?
msingle has quit [Ping timeout: 240 seconds]
<TheLemonMan>
the idea is cool, trying to use a single executable as dynamic and static at the same time is the biggest problem
<TheLemonMan>
maybe if you stick to statically linked glibc/musl and try to run it with a matching dynamic loader... dunno
<TheLemonMan>
on a related note, do you remember why `resizeFn` has a `buf_align` parameter?
<TheLemonMan>
bfredl, ping
<andrewrk>
TheLemonMan, the idea is that the client is responsible for tracking all the allocation metadata, including pointer, length, and requested alignment
<TheLemonMan>
but `buf_align` can be simply obtained by the slice ptr, and it'd be even correct if the user down-aligned the slice
<TheLemonMan>
eg. a []align(16) u32 can be safely thrown around as a []align(4) u32, trying to free the latter will call `resizeFn` with `buf_align` of 4 instead of 16
<marler8997>
I think the point of buf_align isn't to know the alignment of the pointer, but to be used in something like an alignment lookup table
<marler8997>
there are cases where an allocator could use it, however, I might argue that the use cases would be rare, and in the cases where it is useful, the allocator could manage that information itself
<andrewrk>
I'm not strongly in favor of defending the idea. As far as I'm concerned, client tracking of length and alignment, and the thing where you can get more bytes than requested are all experimental
<andrewrk>
IMO if we can't show some real data on perf improvements or debugging experience being better, we should simplify the interface and remove those things
<marler8997>
this decision is what my generic allocator design addresses, it doesn't specify who owns allocation metadata, it can be managed by the caller or the allocator
<TheLemonMan>
the C allocator may make use of that info and choose between plain malloc and posix_memalign
<TheLemonMan>
I hesitated because of this possible problem, picking the wrong function here means a runtime crash so it's pretty critical
<ifreund>
hrm, can't figure out how to work around or isolate this stage1 compiler crash I've found :/
<andrewrk>
ifreund, are you able to get a gdb stack trace?
<andrewrk>
marler8997, yeah we should test the perf of your generic allocator with one that does its own tracking and see if we can use that to justify the interface
<ifreund>
andrewrk: yeah poked around with gdb a bit already, one sec
<marler8997>
andrewrk, yeah that's a good idea
<TheLemonMan>
andrewrk, annnd, what do you think of the proposal in #6870 ?
<TheLemonMan>
marler8997, what's this generic allocator thing?
<TheLemonMan>
two nested `inline for`, you're really pushing your luck here
<andrewrk>
TheLemonMan, I'm in favor of breaking {} and forcing people to be in the habit of using {s} to force string interpretation. I already think in status quo, {} is a footgun
<ifreund>
yeah, though it's been working alright till now :/
<TheLemonMan>
andrewrk, awesome!
<andrewrk>
was that the only open question?
<TheLemonMan>
yeah, unless you don't like the brackets I've used :P
<andrewrk>
for representing slices?
<TheLemonMan>
yeah, both slices and arrays now print as `{ x, xx, xxx, ... }`
<andrewrk>
I'm fine with that
<TheLemonMan>
ifreund, try `up 5` and `p source_node->src()`
<TheLemonMan>
finding and replacing all the stray `{}` is gonna take a while
<ikskuh>
andrewrk: i agree on forcing {s} for strings :)
<ikskuh>
annoying, but we still have writeAll :)
<ikskuh>
well, "annoying"
<ikskuh>
:D
<ikskuh>
typing one character more
<ifreund>
TheLemonMan: that points me to line 71 of the code I linked
<ikskuh>
let vs const again! :D
<fengb>
5076
<ifreund>
(I had to go up a little more than 5)
<TheLemonMan>
awesome, what's the result of `p val->data` in the `ir_resolve_const_type` frame?
<TheLemonMan>
I'm afraid it's tripping up that assertion because of the `destroy: void` field
<TheLemonMan>
hmm, the data payload is completely zeroed... what's `p val->special` ?
<ifreund>
ConstValSpecialStatic
<TheLemonMan>
and `p val->type->special` ?
<ifreund>
"there is no memeber named special"
<TheLemonMan>
arr, `p val->type->id`
<ifreund>
ZigTypeIdMetaType
osa1 has quit [Quit: osa1]
osa1 has joined #zig
<TheLemonMan>
can I reproduce the crash by pulling the zig-wayland repo?
<ifreund>
no, you'll need my WIP branch of river
<andrewrk>
TheLemonMan, we *could* make {} a compile error and you'd have to put some character in there to mean "figure it out based on the type"
<ifreund>
Give me a second and I'll push my current local changes
<andrewrk>
similar to how you have to put `: anytype` in a parameter instead of leaving it blank
dumenci has quit [Ping timeout: 260 seconds]
<ikskuh>
"{?}"
<andrewrk>
yeah something like that
<TheLemonMan>
that's an interesting idea, people may not want to type that extra character tho
<andrewrk>
I agree. however right now we do have a footgun that C printf doesn't have. in printf you get a compile error if you do %s and put a non-string. in zig you get incorrect output at runtime
<ikskuh>
↑
<TheLemonMan>
#1358 reserved "{?}" for the debug output format, makes sense
<ikskuh>
which is kinda a "just give me *something*"
<andrewrk>
if you figure out something that works we can add runtime logic to detect it
segmentationfaul has joined #zig
<segmentationfaul>
wait
segmentationfaul has quit [Client Quit]
segfaults has joined #zig
<segfaults>
there we go
<TheLemonMan>
ifreund, the problem is in deps/zig-wayland/generated/zriver_status_v1_server.zig, the Request type is generated multiple times with different bodies
<TheLemonMan>
removing the extra definitions fixes the crash
<ifreund>
TheLemonMan: they are generated in different namespaces though, shouldn't that be fine?
<TheLemonMan>
yeah, but that's the quickest workaround I thought of :)
<ifreund>
interesting
<TheLemonMan>
ok, the problem are the zero-sized enums
<ifreund>
the ones with just destroy: void?
<TheLemonMan>
yeah
<ifreund>
well it makes sense why nobodies hit this yet, it's not something that would come up in non-generated code
<g-w1>
I now got to VulkanInitalizationFailed by just installing more vulkan packages from apt. (I think a mesa package did the trick)
Xavi92 has joined #zig
ur5us_ has joined #zig
LanceThePants has quit [Read error: Connection reset by peer]
LanceThePants has joined #zig
frett27 has joined #zig
<TheLemonMan>
I really hate when writing a small test case takes more time than fixing the problem :(