ChanServ changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | "Does XXX work yet?": https://alx.sh/fs | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-alt #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
Brainium has quit [Quit: Konversation terminated!]
ddxtanx has quit [Remote host closed the connection]
nesnas has quit [Ping timeout: 480 seconds]
chrisl has joined #asahi
httpszach has quit [Remote host closed the connection]
httpszach has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
httpszach has quit [Ping timeout: 480 seconds]
httpszach has joined #asahi
httpszach has quit [Ping timeout: 480 seconds]
shawn_xu has quit [Remote host closed the connection]
shawn_xu has joined #asahi
tobhe has joined #asahi
httpszach has joined #asahi
nesnas has joined #asahi
pb17 has quit [Ping timeout: 480 seconds]
tobhe_ has quit [Ping timeout: 480 seconds]
nesnas has quit [Ping timeout: 480 seconds]
httpszach has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi
shiggitay has joined #asahi
ddxtanx has joined #asahi
chrisl has joined #asahi
nesnas has joined #asahi
nesnas has quit [Ping timeout: 480 seconds]
marvin24 has joined #asahi
shawn_xu has quit [Read error: Connection reset by peer]
marvin24_ has quit [Ping timeout: 480 seconds]
chrisl has quit [Ping timeout: 480 seconds]
verlio has joined #asahi
chrisl has joined #asahi
nightstrike has joined #asahi
verlio has quit [Quit: verlio]
pb17 has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi
Xe_ has joined #asahi
Xe has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
nesnas has quit [Ping timeout: 480 seconds]
chrisl has quit [Ping timeout: 480 seconds]
chrisl has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
nesnas has quit [Ping timeout: 480 seconds]
balrog_ has joined #asahi
chrisl has joined #asahi
balrog has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
monthebiff has joined #asahi
<monthebiff>
last night i spun up halo masterchief collection via steam, and it launched which was cool. but the machine sounded like it was trying to take off. tried setting everything to low, but couldnt get the machine to calm down and there was some feeling off choppiness when it came to input. (not sure on actual frame rate). any suggestions?
nesnas has quit [Ping timeout: 480 seconds]
monthebi1f has joined #asahi
monthebiff has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
monthebi1f has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
nesnas has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
pb17 has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
nesnas has quit [Read error: Connection timed out]
shiggitay has quit [Ping timeout: 480 seconds]
chrisl has joined #asahi
nesnas has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
<chadmed>
monthebiff: which machine?
<chadmed>
muvm/fex are really quite heavy on the application cores
nesnas has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
Calandracas_ has joined #asahi
ddxtanx has quit [Remote host closed the connection]
Calandracas__ has joined #asahi
Calandracas has quit [Ping timeout: 480 seconds]
Calandracas_ has quit [Ping timeout: 480 seconds]
riker77 has quit [Quit: Quitting IRC - gone for good...]
riker77 has joined #asahi
doti[mds] has joined #asahi
roxfan2 has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
chrisl has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
Method_ has joined #asahi
pb17 has joined #asahi
Method has quit [Ping timeout: 480 seconds]
dcavalca8569 has quit [Ping timeout: 480 seconds]
malte has quit [Ping timeout: 480 seconds]
<fl0_id>
I don't know about how heavy this collection is, but even on macos, with AI stuff or gaming, fans spin up a lot. / take off
<fl0_id>
though a while after closing it should settle down again. (but not immediately)
<fl0_id>
I assume on m1 and m2 it should be similar
ddxtanx has joined #asahi
dcavalca8569 has joined #asahi
chrisl has joined #asahi
<runxiyu_>
my firefox always freezes for about half a minute each time and looks like 20250309_19h02m17s_grim.png
<runxiyu_>
... it primarily happens on my own webapps, and when using mastodon
<runxiyu_>
also happened when i was using MS Outlook from School, but very arerly
<runxiyu_>
... okay it also freezes on forgejo and github now
chrisl has quit [Ping timeout: 480 seconds]
nimprod3l has joined #asahi
roxfan has joined #asahi
<bgtlover[mds]>
<monthebiff> "last night i spun up halo..." <- well, it's more or less normal. Remember, you're running wine, on virtualisation due to page size differences, on emulation. I don't think you can expect anything more than an actual takeoff!
Sebastian2 has joined #asahi
<bgtlover[mds]>
does asahi support running gpu-heavy applications from flatpak? for example, there's this app called alpaca, it can use either the cpu or gpu. It has support for the amdgpu by pulling in a platform extension apparently, but I'm not sure it works with normal mesa and vulkan
<chadmed>
bgtlover[mds]: yes, we have a flatpak runtime extension that bundles up our mesa
<chadmed>
afaik it should be installed automatically
<bgtlover[mds]>
hang on, your mesa? doesn't mainstream mesa have support for the apple gpu?
mesa has quit [Quit: WeeChat 4.0.4]
Sebastian2 has quit [Quit: Sebastian2]
<j`ey>
bgtlover[mds]: kind of
<j`ey>
bgtlover[mds]: it has some support, but not the actual gpu/userspace API
<j`ey>
that is still in a fork
chrisl has joined #asahi
<fl0_id>
for ai stuff, you can also try ramalama. that's should automatically pull right support for linux on mac, or linux container on macos
monthebiff has joined #asahi
<monthebiff>
chadme, m1pro macbook, 16GB ram
<monthebiff>
chadmed *
pb17 has quit [Ping timeout: 480 seconds]
<chadmed>
yeah nothing out of the ordinary then
chrisl has quit [Ping timeout: 480 seconds]
<monthebiff>
bgtlover[mds] i was just comparing it to say... doom 2016 which runs much smoother, even running at full native resolution. im not sure what performance i should be able to get for any given title on this machine,
<chaos_princess>
it will be all over the place, and even "how new is it" is quite often a bad predictor
<monthebiff>
but given that it launched, i thought the first halo game should be relatively lightweight
<chaos_princess>
can you launch fexconfig and tell me if you have "reducted x87 precision" enabled?
<monthebiff>
... feel like i've missed a trick somewhere. command not found, dnf search returns no matches for fexconfig
<chaos_princess>
it is `FEXConfig`
<monthebiff>
one moment, had to figure out it was in fex-emu-utils package
<monthebiff>
that checkbox is currently not checked
<chaos_princess>
check it, lets see if it improves things
Calandracas_ has joined #asahi
monthebi1f has joined #asahi
monthebiff has quit [Ping timeout: 480 seconds]
Calandracas__ has quit [Ping timeout: 480 seconds]
<chadmed>
did microslop put any effort into MCC? i wouldnt expect any x87 code in there if they did
<monthebi1f>
wow.... magic button
<monthebi1f>
the machines silent
<chadmed>
okay i guess they didnt put any effort into MCC :p
<chaos_princess>
implying they have done anything beyond getting the original x86 packaged with a "pretty" launcher
<chaos_princess>
*original x86 build
<monthebi1f>
just had a play around with with the graphics, definitely some better performance rendering at 50% instead of 100%, but whichever im using the machines quiet
<chadmed>
for context: x87 is the _original_ floating point ISA for x86. it's called x87 because it used to be implemented in an off-chip coprocessor of the x87 family (e.g. 8087, 387, 487, etc). it uses an 80-bit format and is generally just extremely fucked up and weird which is why it is slow to emulate
<monthebi1f>
where as yesterday before i even got into the game it was going nuts processing vulkan shaders before the game even launched
<chadmed>
unless of course you just slice off 16 bits so values fit in a 64-bit register
<monthebi1f>
thanks for context i was going to ask
<monthebi1f>
is cool though, halo is working on mac
<monthebi1f>
its come home :P
monthebi1f is now known as monthebiff
<monthebiff>
out of interest im gonna re launch guild wars and see if thats also working
<chaos_princess>
if it did not work without this thing enabled, it will not start working with it
<chaos_princess>
that feature is more of a "emulator speedhack"
<chadmed>
x87 really had no use whatsoever after SSE2 was released... unless youre nvidia in 2008 and you want to gimp physx so your "physx cards" are actually useful
<chadmed>
("physx cards" being a flash chip on a serial bus on a pcie card that had a licence key read by the physx runtime that would switch physx from x87 code to SSE and cuda once nvidia bought AEGIA)
Tenkawa has joined #asahi
<monthebiff>
and it crashed :P
<chaos_princess>
i sorta wonder when nvidia or sth will go full ibm and start charging for *PU utilization
<monthebiff>
guild wars launched, it just was really laggy.
<maz>
chaos_princess: that's @cloud computing@ for you! :)
<chaos_princess>
i usually get charged by vcpu, not by utilization for that
<maz>
depends on the model. in oversubscribed scenarios, you can get charged by the actual runtime.
<chaos_princess>
huh. interesting, but i'd expect this more to be in "faas" scenarios, no?
<chadmed>
chaos_princess: nah microslop offer something similar on azure
<chadmed>
but you can actually save money if your workload is not mission critical by eating up the scrap cycles wasted by folks paying for guaranteed capacity
<chadmed>
instead of buying vm sizes with guaranteed availability your VM just runs in the wasted cycles of people who _are_ paying for it
<maz>
the amusing thing is that at this stage, the biggest waste of cycle is the context-switch...
<chadmed>
i was about to say... all this does is make all VMs slower as they switch out the VMs
<chadmed>
fun hack: write a thing that will hog wakeup time on every single core so that utilisation on ms's side is always 100%, but will yield to your actual workloads when they need to run
<chaos_princess>
hahaha
<chaos_princess>
i remember oracle having free vms on offer, but then taking them away if the vm sits idle, so people were launching a shitcoin miner every week for a hour or so
<chadmed>
lmfao
<chadmed>
burning cycles in an endless spin is genuinely preferable imo :)
<chadmed>
at least fire up folding@home or something...
<monthebiff>
lol interesting. so switch that option back off, guild wars launches, switch it on, guild wars crashes on launch
<chaos_princess>
yep, it is a speedhack :P
<chaos_princess>
in hl2 you can sometimes fall through the map with it enabled
<monthebiff>
o_O
<monthebiff>
lol
<chadmed>
i never said those last 16 bits werent useful
<chadmed>
99% of the time though theyre not that useful
<monthebiff>
99% of the time it works 100% of the time
<LarstiQ>
what's their signifigance? <pun/>
Calandracas__ has joined #asahi
pb17 has joined #asahi
Moprius has joined #asahi
<monthebiff>
ill have a play around and see how other games are affected
<chadmed>
most things made after like 2010 should be fine
<monthebiff>
while im here though, when guild wars 1 is running, its very slow. but its directX 9 on the graphics front. anything i should be aware of when trying to run that?
<chadmed>
most things made before should also be totally fine, but some may be less fine if they actually use all 80 bits of precision
<chadmed>
dx feature level doesnt really correllate with performance. tf2 is dx9 to vulkan/opengl and it runs like ass in certain situations
<chadmed>
guild wars could also have bottlenecks elsewhere, e.g. unoptimised paths in fex
<chadmed>
e.g. deus ex human revolution runs flawlessly in game, but its menu system is a slide show if you dont have reduced x87 precision enabled
<monthebiff>
is it all vulkan/ all opengl, or is there a way to say... try the other one, see if it runs smoother?
<chadmed>
the linux version of tf2 lets you switch between them
<chadmed>
either way it's dx9->vk or dx9->opengl
<chaos_princess>
for x87 precision - if it is a 64 bit binary - in general it will have zero effect
Calandracas_ has quit [Ping timeout: 480 seconds]
<monthebiff>
guild wars one was 2005
<monthebiff>
i suspect its 32bit
chrisl has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
<bgtlover[mds]>
<monthebiff> "bgt lover i was just comparing..." <- well, doom is doom, it runs on pretty much everything
<bgtlover[mds]>
<fl0_id> "for ai stuff, you can also try..." <- yeah, I'm just asking at this point, to see how good it is. Is that gpu any good, can it run games and AI and such well enough, or is amdgpu generally better?
<monthebiff>
weirdly enough original doom, ie floppy disk doom i actually can't play atm :P or couldn't some graphical glitching particulary in the menus
<chaos_princess>
is that even protected mode 32 bit?
<monthebiff>
i've never checked, i can do in a minute
<monthebiff>
was one of the first things i thought to install because... doom runs on everything :P but i never dug into it
<monthebiff>
doom 2016 was very shiny and i got distracted
<chaos_princess>
(i sorta expect them to ship a dos emulator and run it in that, from what i remember doom directly pokes the gpu registers)
<chadmed>
winevdm wen :p
<chaos_princess>
just run x86box or sth at that point
Moprius has quit [Quit: bye]
<chadmed>
i should try dosbox in fex
<chaos_princess>
why
<chaos_princess>
run x86box, it has a x86 interpreter
<chadmed>
oh wow i didnt know that was a thing
<chadmed>
this looks neat
<monthebiff>
sorry for being a pain, but how am I checking that 32 bit mode? is that an fex-emu thing, or a steam thing?
<bgtlover[mds]>
<chaos_princess> "is that even protected mode 32..." <- nope, probably still real mode
<chaos_princess>
monthebiff: list the game contents, what files are there
<bgtlover[mds]>
<chaos_princess> "just run x86box or sth at that..." <- isn't that box86, or was it box64? btw, is fex better than box64?
<chaos_princess>
it is "86box", i forgot the name, compared to box86/box64/fex, that is a full system emulator
<chaos_princess>
we recommend fex here in general
<monthebiff>
side note, steam provides a 'launch dos' option
<monthebiff>
and that works, though at limited screen size
<chaos_princess>
yep, they shipped dosbox
<bgtlover[mds]>
chaos_princess: ahh, so like...qemu on steroids?
<chaos_princess>
sorta?
<chaos_princess>
they kinda do similar things
<bgtlover[mds]>
monthebiff: ok, now I'm speechless
<chaos_princess>
not like they had any other choice for running original doom on modern computers
<monthebiff>
i can do screenshots/ captures later
<monthebiff>
but i gotta go be an adult and make dinner
<monthebiff>
cheers for the help all o/
<bgtlover[mds]>
I'm surprised they'd go to that length to preserve stuff like that, most companies don't care. By that logic, portal should be playable too?
<chaos_princess>
portal is very playable
<monthebiff>
portal works great can confirm
<monthebiff>
so does hollow knight
<bgtlover[mds]>
oho, awesome! y'all really did something never done before, I never heard of something like this. Emulation is one thing, but a vm on top, and having the thing still playable at the end of the day? now that's awesome
<chaos_princess>
it is not quite the same as a normal vm. yes, it is using virtualization, but it does not emulate almost any hardware, and lets the vm directly access host hardware
<chaos_princess>
yes, but those are not that expensive for cpu
<bgtlover[mds]>
that should normally add some overhead anyway, the fact that it's playable enough is still awesome
<chaos_princess>
it is the rest of it that creates overhead
<bgtlover[mds]>
chaos_princess: ahh, then it's because of my ancient computer over here, back when this thing was new, hypervisors were quite costing on the cpu
<chaos_princess>
The more you have to emulate the worse it is and we emulate almost nothing.
<bgtlover[mds]>
did you make the whole vm from scratch? I'm kinda curious to see the code of that if so, but I'll probably have my brain fold upon reading it :p
<bgtlover[mds]>
O, it's rust? even better, I know that
<chaos_princess>
are you sure :P
<bgtlover[mds]>
chaos_princess: I know normal rust, I dk, if this involves kernel bindings, then I don't because I never worked with that. I maintain some rust programs, namely the odilia screenreader is what me and a couple of people are working on, so yeah, I know some basics :p
<chaos_princess>
it is more that it contains probably the most cursed rust i've ever written :P
<chadmed>
cursed or genius
<bgtlover[mds]>
chaos_princess: ahh, oof! is it worse than using tower and basically recreating axum for accessibility events sent over dbus? only reviewing that stuff made my head spin
<bgtlover[mds]>
chadmed: that's always the question, isn't it?
<monthebiff>
it is wickedly cool
<chaos_princess>
bgtlover[mds]: it is "swap out file descriptors and memory mappings in other processes"
<j`ey>
wayland native wen
<bgtlover[mds]>
chaos_princess: yes, that's so much worse! oho, this is starting to sound like fun
tertu2 has quit [Quit: so long...]
<chaos_princess>
j`ey: looked at it, and it is annoying, if i got this right, wayland's protocol is not self-delimiting, so i will have to parse the entire protocol, instead of only messages i care about like i can do with x11 and pw
<j`ey>
what does self-delimiting mean, it doesn't have a size or?
<j`ey>
size field
<chaos_princess>
although would be funny to run hsr via waydroid
<chaos_princess>
yes, you need to know what the message is about to know it's size
<j`ey>
ah
<bgtlover[mds]>
chaos_princess: self-delimiting? and wait what, you're parsing the pipewire protocol by hand? now that's wickid!
<chaos_princess>
there is like 3 messages i care about, so yes? same for x11, except there it is like 10
tertu has joined #asahi
<bgtlover[mds]>
chaos_princess: huh? you want to run waydroid apps from the vm?
<chaos_princess>
yes
<chaos_princess>
android also has the 4k page issue
<bgtlover[mds]>
chaos_princess: ahh shit, those are hard to parse indeed
<bgtlover[mds]>
that's why I hate xml and such too
<bgtlover[mds]>
chaos_princess: ahh, gotcha. And rn asahi can't run waydroid because of that?
<chaos_princess>
yes
<chaos_princess>
not only that, but that is one of the issues
<j`ey>
chaos_princess: but on the other hand, there's rust crates for wl-p already
chrisl has joined #asahi
<chaos_princess>
hmm, there were people who were interested in contributing :P
<bgtlover[mds]>
j`ey: depends, some have too many dependencies, like the cosmic stuff
<bgtlover[mds]>
I didn't know this stuff has wireplumber code in it...I suspect it's for passthrough?
<chaos_princess>
it is kind of a dirty hack because i did not want this thing to also handle dbus and portals
<chaos_princess>
this just assigns correct permissions for webcam passthrough
Gues__________________________ has joined #asahi
Gues__________________________ has quit []
chrisl has quit [Ping timeout: 480 seconds]
JayBeeFOSS has quit [Ping timeout: 480 seconds]
nesnas has quit [Ping timeout: 480 seconds]
JayBeeFOSS has joined #asahi
<monthebiff>
dirty hack... working code... *shrug*
<monthebiff>
ship it :P
<chadmed>
it does actually take out the host's pipewire occasionally but i cant reliably repro :p
<bgtlover[mds]>
<monthebiff> "dirty hack... working code... *..." <- correct
<chaos_princess>
It could be trying to negotiate explicit sync, or maybe allocate memory in vm and send to host.
<chaos_princess>
But if you have repro, or crash log, im interested
<chaos_princess>
(My only test cases were "play music via pw-play, and get camera via firefox)
rubin55 has joined #asahi
<bgtlover[mds]>
chaos_princess: oho, now that's definitely not pleasent
<bgtlover[mds]>
and...krun, I remember I heard that before. It's in...the container context. Now I'm curious how can this be a vm? I'm reading the rust bindings to said library
<chaos_princess>
krun is a thing to run containers as a full vm
<chaos_princess>
kinda?
okb has joined #asahi
<bgtlover[mds]>
chaos_princess: so...like firecracker? I learned afew new things today
<chaos_princess>
libkrun is somewhat like firecracker as a library, muvm is a thing that sets up the vm and does some hardware/desktop environment passthrough
<monthebiff>
looking at the aus apple store, i think they've stopped selling the m2/m3 airs completely. can only see pre orders for the m4 now.
<monthebiff>
m3/4 supports still a way off right? how goes the testing stuff mentioned in the recent blog post? CI/ testing etc?
<j`ey>
chaos_princess: youre running firefox in muvm?
<chaos_princess>
was running it as a test for webcams, yes
<chaos_princess>
for daily usage - ofc running it directly on host os
<j`ey>
ah testing
<bgtlover[mds]>
speaking of that, widevine and such stuffage still works?
<chaos_princess>
run widevine-installer and it should do the magic
<bgtlover[mds]>
when I read that blog post a while ago about widevine, I couldn't believe someone can patch executables like that
<j`ey>
i still havent installed rpmfusion ffmpeg, I really should do that
<chaos_princess>
why not. As long as it does not checksum itself, it should be fine
<j`ey>
just modify the expected checksum :D
<bgtlover[mds]>
chaos_princess: because...it sounded so outlandish that I thought it couldn't possibly work! turns out, it totally does! now I know more about how that's done, because I followed amos's writing an executable packer series, you probably already know who that is, but yeah, I didn't before
<bgtlover[mds]>
j`ey: unless it's a code signature they're getting from a server, or which they put in the secure onclave, think secure boot
pb17 has quit [Ping timeout: 480 seconds]
<chaos_princess>
j`ey: yea, and then you realize it checksums in multiple places, and some of them are obfuscated, and so on
<chaos_princess>
at that point, just use alternative methods to acquire your media
<bgtlover[mds]>
yeah, there should be a point where those monopoly practices have to stop, even corrupt politicians probably realise that
<fl0_id>
bgtlover[mds] even with good intent, there is no certainty that you get good faith compliance, and that you get a well written law
pb17 has joined #asahi
<monthebiff>
dont suppose anyone has stumbled into a working WoW configuration since i was last asking about it?
<monthebiff>
a boy can dream
nesnas has joined #asahi
nesnas has quit [Ping timeout: 480 seconds]
catred has quit [Read error: Connection reset by peer]
catred has joined #asahi
catred2 has joined #asahi
catred has quit [Read error: Connection reset by peer]
creechy has quit [Quit: ZNC 1.8.2+deb2ubuntu0.1 - https://znc.in]
creechy has joined #asahi
nesnas has joined #asahi
creechy has quit [Quit: ZNC 1.8.2+deb2ubuntu0.1 - https://znc.in]
<monthebiff>
night all o/
monthebiff has quit [Quit: leaving]
creechy has joined #asahi
Tenkawa has quit [Quit: Was I really ever here?]
<bgtlover[mds]>
<fl0_id> "bgt lover even with good intent,..." <- true, but not trying isn't a good option, for example in the US :p
Calandracas__ has quit [Ping timeout: 480 seconds]
chrisl has joined #asahi
<fl0_id>
bgtlover[mds] yeah for sure. nowhere. and like here f.e. we have a similiar thing where right now the supposedly more moderate party rolled over to work together with conservatives, either because power or because they tell themselves they want to prevent the far right party, but then what use is that if they adopt their points...
<fl0_id>
(sorry for the OT)
glem81005488 has quit [Quit: bye.]
pb17 has quit [Ping timeout: 480 seconds]
chrisl has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi
nimprod3l has quit [Ping timeout: 480 seconds]
Calandracas has joined #asahi
nimprod3l has joined #asahi
glem81005488 has joined #asahi
chrisl has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
nimprod3l has quit [Ping timeout: 480 seconds]
snuck has joined #asahi
Mary2 has joined #asahi
timefractal has quit [Quit: Ping timeout (120 seconds)]
Mary has quit [Read error: Connection reset by peer]
sneak has quit [Write error: connection closed]
js6pak has quit [Quit: Ping timeout (120 seconds)]
js6pak has joined #asahi
timefractal has joined #asahi
Mary2 has quit []
chrisl has joined #asahi
Mary2 has joined #asahi
Mary2 is now known as Mary
sand4rt has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
bjoto` has joined #asahi
bjoto has quit [Ping timeout: 480 seconds]
Calandracas_ has joined #asahi
<bgtlover[mds]>
<fl0_id> "bgt lover yeah for sure. nowhere..." <- let me guess, germany?
Calandracas has quit [Ping timeout: 480 seconds]
sand4rt has quit [Quit: WeeChat 4.5.2]
sand4rt has joined #asahi
sand4rt has quit []
Calandracas__ has joined #asahi
sand4rt has joined #asahi
Calandracas_ has quit [Ping timeout: 480 seconds]
<bgtlover[mds]>
chaos_princess (IRC): I'm looking at the muvm sources, particularly the raw bindings to krun. In the crate root, you have a file called wrapper.h, which just includes the system header. Wouldn't it be better to use pkgconfig to get the path of the header you want? is it not done because it's good enough as it is, or because that's the only way bindgen supports of doing that?
<bgtlover[mds]>
<chaos_princess> "https://github.com/AsahiLinux/..."; <- not system wide installs? ahh, that explains it lol. But wrapper.h still includes the system header, unless...you use chroot to make an entirely different prefix? that sounds like a buildroot for cross-compilation, with C libs
<PaulFertser>
pkg-config can give -I flags that tell pre-processor to prepend a path to the system-wide list, so no chroot needed.
<bgtlover[mds]>
I looked at the commit, still doesn't explain much :p, wrapper.h is still there
<bgtlover[mds]>
* still there and it wasn't added by that commit
Calandracas_ has joined #asahi
<chaos_princess>
I am not 100% sure as to what is the proper way, and pkg-config support is very recent
<bgtlover[mds]>
PaulFertser: aha, awesome! luckily we're no longer in the autotools without pkgconfig world, that's brutal
nesnas has quit [Ping timeout: 480 seconds]
<PaulFertser>
bgtlover[mds]: brutal was having to maintain a separate set of makefiles for each supported OS I'd say ;)
nesnas has joined #asahi
<bgtlover[mds]>
PaulFertser: possible, I probably wasn't born then, so I didn't experience it. I did recently have to write autotools when contributing to speech dispatcher though, and even with pkgconfig it was horrible
Calandracas__ has quit [Ping timeout: 480 seconds]
<PaulFertser>
bgtlover[mds]: what was horrible, m4 or shell? I mean clearly it's somewhat fragile to use a text macro language plus POSIX shell language (which was never that great even for imperative programming) to express notions that should be normally declarative and purely functional when more flexibility is needed. But e.g. CMake is essentially the same but you need to learn yet another imperative
<PaulFertser>
language to write it.
<bgtlover[mds]>
more like m4, also the way dependencies are declared, coming from meson it all felt weird. Shell, I don't like it a lot, but I'm not so ignorant about it
chrisl has joined #asahi
nesnas has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi
chrisl has quit [Ping timeout: 480 seconds]
nesnas has joined #asahi
ten157237743246305066182150355 has joined #asahi
bjoto`` has joined #asahi
bjoto` has quit [Ping timeout: 480 seconds]
halloy4565 has joined #asahi
Calandracas has joined #asahi
Calandracas_ has quit [Ping timeout: 480 seconds]
Calandracas has quit [Remote host closed the connection]
Calandracas_ has joined #asahi
<halloy4565>
hi, i plan on reinstalling alx-fedora with sway via a "minimal" install, and i was wondering, what are the set of packages that a normal installation of alx-fedora GNOME/KDE would install? if no one knows exactly, could someone point to me where i could figure this out?