<pixelherodev>
What's the libc target for windows/
<pixelherodev>
?
<pixelherodev>
For cross-compilation
<pixelherodev>
Is there no supported one?
dingenskirchen has quit [Read error: Connection reset by peer]
dingenskirchen has joined #zig
aerona has joined #zig
foobles has joined #zig
<foobles>
how do a turn an LLVMValueRef representing an instance of some type T into an LLVMValueRef representing a pointer to T?
<foobles>
I also have it as an IrInstGen
<foobles>
do I just use result_loc?
<foobles>
do I have to store it first?
_Vi has quit [Ping timeout: 272 seconds]
StateOff has joined #zig
dddddd has quit [Remote host closed the connection]
aerona has quit [Remote host closed the connection]
aerona has joined #zig
ur5us has joined #zig
frenata has quit [Remote host closed the connection]
nephele has quit [Ping timeout: 244 seconds]
nephele has joined #zig
metaleap has joined #zig
aerona has quit [Ping timeout: 244 seconds]
benjif has quit [Ping timeout: 264 seconds]
decentpenguin has joined #zig
klltkr_ has quit [Ping timeout: 264 seconds]
decentpenguin has quit [Quit: decentpenguin]
xackus_ has joined #zig
xackus has quit [Ping timeout: 244 seconds]
<Cadey>
i wonder how far the TCP chat server example is from being an IRC server
<foobles>
anyone here available to talk about generating LLVM IR? I am trying to unwrap an optional, which requires a pointer to the optional object, to do LLVMBuildGEP
<foobles>
im not sure how to accomplish that
<foobles>
also there are some strange things i'm seeing, like handle_is_ptr
<foobles>
what does that actually check, when called on an optional type?
waleee-cl has quit [Quit: Connection closed for inactivity]
layneson has joined #zig
foobles has quit [Remote host closed the connection]
layneson has quit [Ping timeout: 260 seconds]
foobles has joined #zig
slowtyper has joined #zig
<andrewrk>
pixelherodev, x86_64-windows-gnu or i386-windows-gnu
cow-orker has quit [Remote host closed the connection]
metaleap has quit [Quit: Leaving]
metaleap has joined #zig
metaleap has quit [Client Quit]
metaleap has joined #zig
ifreund[m] has joined #zig
pmwhite has joined #zig
fengb has joined #zig
afontain_ has joined #zig
BitPuffin has joined #zig
Dominic[m] has joined #zig
AlexMax has joined #zig
dtz has joined #zig
metaleap has quit [Client Quit]
metaleap has joined #zig
ur5us has joined #zig
cow-orker has joined #zig
ifreund has quit [Read error: Connection reset by peer]
ifreund has joined #zig
kubkon has quit [Ping timeout: 240 seconds]
dingenskirchen1 has joined #zig
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen1 is now known as dingenskirchen
dingenskirchen has quit [Client Quit]
dingenskirchen has joined #zig
ifreund has quit [Read error: Connection reset by peer]
ifreund has joined #zig
Kingsquee has quit [Quit: Konversation terminated!]
dddddd has joined #zig
<alexnask>
andrewrk, Are there any plans to change the way generic functions are type erased in stage2?
<alexnask>
More specifically, currently every arg after the first generic arg (+ the return type) are erased, I believe it would be more convenient for metaprogramming code to only erase generic args (and the return type if it depends on them)
<alexnask>
Sorry if I've missed some issue or discussion on this, I looked at the issue tracker but can't find anything relevant
<mikdusan>
fn(type,var,var) var
<mikdusan>
and maybe require `comptime` keyword for each comptime arg
_Vi has joined #zig
<ikskuh>
mikdusan: i don't think that was alexnask' point
<mikdusan>
which part? the comptime?
<alexnask>
By 'erased' I am talking about the type turning to 'var'
<ikskuh>
two functions:fn(size: usize, comptime T: type)
<ikskuh>
and
<ikskuh>
fn(comptime T: type, size: usize)
<ikskuh>
second one has signature fn(type,var), first one has fn(usize,type)
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen has joined #zig
<mikdusan>
ikskuh: all I'm saying is (probably because of) type erasure, `fn(comptime T0: type, T1: type)` is allowed today. there's a smell to that.
<alexnask>
Wait, this actually compiles? I thought ccomptime only types needed to be comptime to appear as arguments
<mikdusan>
2nd, 3rd,.... comptime keyword not required afaik
<alexnask>
Right, it could be that the check happens after they are erased?
<alexnask>
I should look into the funcdecl code in detail
<mikdusan>
right, so this is particularily odd to me... `fn(one: bool, comptime T0: type, T1: type)`
FireFox317 has quit [Ping timeout: 256 seconds]
marijnfs has joined #zig
<marijnfs>
Is it feasible now to add external cpp libraries by source, like you might a c library in Zig?
<marijnfs>
since zig can do c++ now right?
<mikdusan>
marijnfs: there were some recent successes with that but couple of caveats; it's fairly new and since this is c++ we are talking about, any other native dep (recursively) using c++ must also be built with zig
<ikskuh>
marijnfs: my experience is that you have to add all deps then
<ikskuh>
but it's worth it for crosscompilation :)
<tdeo>
if i have a ZigTypeIdPointer pointing to a single element, i can use ir_read_const_ptr to read the value from it. but what if i have one to multiple items (from a slice)?
<tdeo>
oh, maybe i can change ConstPtrValue.data.base_array.elem_index?
<mikdusan>
slices are repr as structs with 2 fields. [slice_ptr_index] [slice_len_index]
<tdeo>
yeah i have that part already, but i'm struggling with using the ptr value from it
ur5us has quit [Ping timeout: 252 seconds]
<mikdusan>
maybe this will help: inside ir_analyze_ptr_cast() look for `// For slices, follow the `ptr` field` and how it sets up ir_analyze_struct_field_ptr()
st4ll1 has joined #zig
FireFox317 has joined #zig
wootehfoot has joined #zig
riba has joined #zig
<marijnfs>
mikdusan: ikskuh: I see, interesting
<marijnfs>
I basically had this issue that I was linking from build.zig to libc++, but the library I linked to was probably built with another version and I got crazy crashes
<marijnfs>
took ages to figure out, but I figure these issues would be relieved if I build the whole library in my project directly.
<marijnfs>
Or stop blindly linking to libc++
riba has quit [Ping timeout: 246 seconds]
<tdeo>
can someone who knows the compiler review #5203 thanks
dingenskirchen has quit [Remote host closed the connection]
dingenskirchen has joined #zig
jaredmm_ is now known as jaredmm
Snektron has joined #zig
dimenus has joined #zig
dimenus has quit [Remote host closed the connection]
<FireFox317>
yo, anyone in here knows a good single board computer (like raspberrypi) but that doesnt have a close source gpu/bootloader and has good documentation?
ifreund has quit [Quit: WeeChat 2.8]
BaroqueLarouche has joined #zig
<FireFox317>
i'm interested in low level stuff, but the raspberry pi is just so annoying to work with since broadcom -_-
<ikskuh>
FireFox317: relatable
<ikskuh>
but i would not use a SBC on PI level
<ikskuh>
those SOCs are so large-scale that it's a pain to work with
<FireFox317>
i mean there should be SOCs that have good documentation and stuff right?
<ikskuh>
hah you wish
<ikskuh>
there are SOCs
<ikskuh>
but smaller ones
<ikskuh>
as soon as you have a GPU on board, it's mostly "boots linux with blob"
<FireFox317>
hmm yeah i see, so what with smaller socs you are more talking about stm32 stuff i think?
dimenus has joined #zig
<ikskuh>
yeah
<ikskuh>
there was one board though
<ikskuh>
ultra-high powered soc
<ikskuh>
cost-level of stm32
<FireFox317>
maybe rpi is the easiest then because there has been quite some research to the videocore and stuff
<ikskuh>
yeah probably
<fengb>
I've been thinking about getting a toy FPGA
<ikskuh>
fengb: prepare for trouble
<ikskuh>
"learn programming anew"
<fengb>
Then you can deal with your own bootloader! :P
<ikskuh>
also, reserve ~20GB of your disk then
<fengb>
ikskuh: aren't you supposed to be pimping out your new CPU?
<dimenus>
RPI also just recently released the actual datasheet for the rpi4
<BaroqueLarouche>
Get a DE10 Nano so you can run MiSTer too
<ikskuh>
and prepare for *really bad tooling*
<dimenus>
the downside being QEMU doesn't really support the rpi4
<FireFox317>
Ah lattice, they have some fpga chips that are completely reversed enigneered by people <3
<ikskuh>
yep
<ikskuh>
lattice is nice anyways
<ikskuh>
cheap chips
<ikskuh>
not that powerful, but 20$ per chip is okay
<fengb>
Oh I forgot about DP
<ikskuh>
i still do gool ol' vga
<ikskuh>
because it's WAY easier than anything newer
Cadey has joined #zig
chapl has joined #zig
<nephele>
I wonder, does zig allow for comments aimed at code editors only?
<nephele>
I was thinking about hungarian notation and so, in many cases knowing immidiently what a variable belongs to (and where it doesn't belong) could be quite powerfull
<nephele>
But i would like to do this with colors instead of a text notation, so basically telling my editor somehow "this var be red!"
<chapl>
Somebody mind helping me out real quick? Just a tini tiny issue. `var cursor: i32 = 1;` and `cursor += 1;` doesn't actually increment the cursor by one, but by three instead. At least that's what `io.getStdOut().outStream().print()` is telling me.
<ikskuh>
huh
<ikskuh>
it should not…
<chapl>
Wanna see the code? Maybe I made some stupid newbie mistake. It's about 21 lines.
<ikskuh>
note that it will output 1024 zeroes every time!
<chapl>
I'm aware of that. I'll fix that later. :)
<ikskuh>
okay
<ikskuh>
also, entering '1\r' will yield two bytes ;)
foobles has joined #zig
<chapl>
OHHH
<chapl>
fuck me
<ikskuh>
readByte will read the char entered as well as the carriage return
<ikskuh>
hey foobles!
<foobles>
yo!
<foobles>
:)
<chapl>
Thank you ikskuh!
<foobles>
:)
<chapl>
But that doesn't actually fix the integer inc issue I have
<ikskuh>
why?
<chapl>
Did you have a look at the code? Line 19: `answer_cursor += 1;` - I don't read the integer in. I just do plain integer addition and instead of +1 I get +2 and sometimes +3
r4pr0n has joined #zig
<mikdusan>
1\r\n is 3 bytes
<ikskuh>
yes
<ikskuh>
the error is '\r'
<ikskuh>
you will just override your first increment with the second
<ikskuh>
and it's so fast you can't see that
<chapl>
Oh now I get it. \r and \n account as separate read bytes.
<ikskuh>
yeah, sure ;)
<chapl>
That makes sense. Alright. Thank you. Much obliged!
puzzleddev has joined #zig
<scientes>
are zig programs PIE?
<scientes>
is there such thing as static-linked PIE?
<scientes>
and you could just run it by jumping to it
<nycex>
scientes: you can use zig build-exe -fPIE, then it is
<scientes>
and how do you know the entry point?
<scientes>
i.e. could you load it from another program without exec()?
<nycex>
uhm nvm i confused it with PIC
<scientes>
I guess it still is an ELF with a number of sections
<scientes>
nycex, they are very similar
<scientes>
wait, but the loader does most of the ELF handling
<scientes>
I guess the kernel does some too
<nycex>
scientes: I know, but i'm not sure if -fPIC does what you want
<scientes>
./fs/binfmt_elf.c
<mikdusan>
scientes: I believe what happens is the elf loader decides base address to load at. being pic/pie, the baseaddr doesn't matter. then elf is used to figure offsets and determine entry
<scientes>
linux/fs/binfmt_elf.c
layneson has joined #zig
waleee-cl has joined #zig
Akuli has joined #zig
tdeo has quit [Quit: Quit]
tdeo has joined #zig
chapl has quit [Quit: Leaving]
cole-h has joined #zig
wootehfoot has quit [Read error: Connection reset by peer]
alva has joined #zig
ifreund has joined #zig
ifreund1 has joined #zig
puzzleddev has quit [Remote host closed the connection]
puzzleddev has joined #zig
ifreund has quit [Ping timeout: 260 seconds]
r4pr0n has quit [Remote host closed the connection]
decentpenguin has joined #zig
Nypsie has joined #zig
dimenus has quit [Remote host closed the connection]
pystub has joined #zig
TheLemonMan has joined #zig
dingenskirchen has quit [Remote host closed the connection]
<FireFox317>
uhm, the docs say that `readNoEof` returns the number of bytes read, but it doesnt. should it?
r4pr0n has joined #zig
<ifreund>
alright so my hack seems to be ok with a minimal example https://paste.rs/LoI
<fengb>
How did you steal Lol? :(
dddddd__ is now known as dddddd
<ifreund>
fengb: L o I as in iguana actually :D
<fengb>
lol
<ifreund>
oh well this is interesting, maybe it's not my hack that's causing the crash https://paste.rs/M9k
marijnfs_ has joined #zig
Nypsie has quit [Quit: WeeChat 2.8]
r4pr0n has quit [Quit: r4pr0n]
r4pr0n has joined #zig
dimenus has joined #zig
pystub has quit [Ping timeout: 246 seconds]
dimenus has quit [Remote host closed the connection]
dimenus has joined #zig
r4pr0n has quit [Client Quit]
hazeycode has joined #zig
urluck has quit [Remote host closed the connection]
urluck has joined #zig
jwmerrill has quit [*.net *.split]
gazler has quit [*.net *.split]
terinjokes has quit [*.net *.split]
WilhelmVonWeiner has quit [*.net *.split]
tralamazza has quit [*.net *.split]
dottedmag has quit [*.net *.split]
pixelherodev has quit [*.net *.split]
butlerian has quit [*.net *.split]
dom96 has quit [*.net *.split]
dottedmag has joined #zig
terinjokes has joined #zig
urluck has quit [Read error: Connection reset by peer]
gazler has joined #zig
dom96 has joined #zig
urluck has joined #zig
<hazeycode>
andrewrk: The crash I was having is something to specific to my code, with evented io on macOS it's easy to invoke the Cocoa event pump off the main thread, which is no good
<andrewrk>
hazeycode, with the way that zig's async/await/resume/suspend works, it's possible to make a function called `switchToMainThread();` which works as advertised
FireFox317 has quit [Ping timeout: 256 seconds]
Akuli has quit [Quit: Leaving]
ur5us has joined #zig
layneson has joined #zig
Chris660 has joined #zig
pystub has joined #zig
pixelherodev has joined #zig
utzig has joined #zig
<hazeycode>
I'm having some trouble with this andrewrk, I think I'm missing something, I'll try again tomorrow. I've opened a PR to enable async io on macOS, it was easier than I thought, again, I think I'm missing something
<hazeycode>
It's also not really been tested at all, and I'm quite sure some of it is incorrect. But perhaps it will attract comments such that I can go and fix :)
TheLemonMan has quit [Remote host closed the connection]
puzzleddev has quit [Quit: Leaving]
pystub has quit [Ping timeout: 265 seconds]
Chris660 has quit [Ping timeout: 265 seconds]
<andrewrk>
nice!
Invader_Bork has quit [Ping timeout: 260 seconds]
metaleap has quit [Quit: Leaving]
lanodan has quit [Quit: WeeChat 2.7.1]
lanodan has joined #zig
hazeycode has quit [Ping timeout: 240 seconds]
<ikskuh>
andrewrk: quick question: can i somehow emit files in subdirectories in zig-build?
<ikskuh>
i want to build for three platforms in a single build file
<andrewrk>
yes
<andrewrk>
setOutputDir works. but probably zig build should gain some recognition of multiple install directories for this use case
<ikskuh>
yeah, true
<ikskuh>
i'll work it out with setOutputDir for now :)
<pixelherodev>
Would it be worth improving the glibc for Zig update process?
<pixelherodev>
Rather, would it be possible?
ifreund has quit [Quit: Lost terminal]
dddddd has quit [Quit: dddddd]
<ikskuh>
okay, seems like my emulator doesn't run well on windows yet…