gemini - kennedy.gemi.dev

💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2021-09-26.txt captured on 2024-05-12 at 16:04:15.
-=-=-=-=-=-=-
[2021-09-26T00:02:49Z] <claudia02> illiliti: https://gitlab.freedesktop.org/wayland/wayland/-/issues/72
[2021-09-26T00:03:04Z] <claudia02> there is this ubikbsd link which is I think obsd related.
[2021-09-26T00:03:22Z] <claudia02> It was accessible without account. maybe its gone now. dunno
[2021-09-26T00:12:31Z] <illiliti> i don't understand why they can't just use poll/select
[2021-09-26T00:13:33Z] <ryoshu> in wayland?
[2021-09-26T00:13:38Z] <ryoshu> upstream
[2021-09-26T00:13:57Z] <illiliti> yep
[2021-09-26T00:14:30Z] <illiliti> https://gitlab.freedesktop.org/wayland/wayland/-/blob/main/src/event-loop.c
[2021-09-26T02:54:50Z] <tleydxdy[m]> hmm, when I'm trying to make the kernel the build script seems to want diff -I
[2021-09-26T02:58:14Z] <tleydxdy[m]> a few other errors, like for find
[2021-09-26T05:29:15Z] <testuser[m]> Hi
[2021-09-26T05:29:20Z] <testuser[m]> tleydxdy: it doesn't matter
[2021-09-26T05:57:07Z] <tleydxdy[m]> hmm, it only built a tiny part of the kernel
[2021-09-26T05:57:18Z] <tleydxdy[m]> guess my config is wrong somewhere
[2021-09-26T05:57:45Z] <tleydxdy[m]> perhaps I should use my current one from arch as the base
[2021-09-26T06:02:04Z] <tleydxdy[m]> noticed the use of underscore
[2021-09-26T06:02:15Z] <tleydxdy[m]> didn't know that was a thing in the shell
[2021-09-26T10:13:58Z] <claudia02> illiliti: heres some wip work from 3-4 ago https://github.com/jasperla/openbsd-wip/tree/master/x11/wayland
[2021-09-26T10:24:29Z] <ryoshu> 3-4 years ago
[2021-09-26T16:14:11Z] <tleydxdy[m]> ree, compiling the kernel still gives __always_inline error, but I've already patched swab.h
[2021-09-26T16:16:33Z] <testuser[m]> just find the file with the error and add the same linux/stddef.h include
[2021-09-26T16:17:00Z] <tleydxdy[m]> yeah, it's still complainingabout swab
[2021-09-26T18:06:00Z] <tleydxdy[m]> ah, wth
[2021-09-26T18:06:07Z] <tleydxdy[m]> it's using the headers from the system
[2021-09-26T18:06:14Z] <tleydxdy[m]> rather than the unpacked directory
[2021-09-26T18:06:48Z] <GalaxyNova> wierd
[2021-09-26T18:15:38Z] <tleydxdy[m]> lol samba.org is down
[2021-09-26T18:17:37Z] <tleydxdy[m]> is there someway to do split package?
[2021-09-26T18:17:58Z] <tleydxdy[m]> or are the source file downloaded automatically deduped
[2021-09-26T20:05:49Z] <acheam> tleydxdy[m]: sources are not deduped between pacakges
[2021-09-26T20:12:10Z] <tleydxdy[m]> hmm
[2021-09-26T20:13:11Z] <tleydxdy[m]> would make sense to dedup using the checksum for example
[2021-09-26T20:14:50Z] <tleydxdy[m]> ah, I guess that would be a problem if say the checksum was outdated
[2021-09-26T20:15:35Z] <tleydxdy[m]> still fighting the initramfs, fun
[2021-09-26T20:40:21Z] <acheam> i guess it could, but that might be slow
[2021-09-26T20:40:38Z] <acheam> would require storing the checksum of every downloaded file and grepping for it each time you download a package
[2021-09-26T20:41:06Z] <acheam> for a not too common use case
[2021-09-26T20:53:02Z] <rio6> checksums are already stored in package db
[2021-09-26T20:54:14Z] <rio6> oh hm nvm
[2021-09-26T20:54:37Z] <rio6> the files need them too
[2021-09-26T20:55:57Z] <tleydxdy[m]> well, it would just be a symlink
[2021-09-26T20:57:18Z] <tleydxdy[m]> so the files are stored in sources with their name as the checksum
[2021-09-26T20:57:35Z] <tleydxdy[m]> and the current way source work are converted to symlinks
[2021-09-26T20:57:50Z] <tleydxdy[m]> but yeah, now I think about it it would has some problem
[2021-09-26T20:58:02Z] <tleydxdy[m]> better to have split package or provides
[2021-09-26T20:59:12Z] <tleydxdy[m]> s/name/checksum/, s/checksum/file name/
[2021-09-26T20:59:12Z] <cotangent> <tleydxdy[m]> better to have split package or provides
[2021-09-26T21:11:45Z] <rio6> I do want a "provides" for kiss, I think the current way of doing it is to symblink alternative package to the package it provides, but it's not the greatest
[2021-09-26T21:41:03Z] <Guest4475> Are there any RSS feed readers that have been packaged for kiss already?
[2021-09-26T22:09:00Z] <tleydxdy[m]> rio6: I tried the symlink method, but it doesn't quite work right (or I'm misunderstanding something)
[2021-09-26T22:09:16Z] <tleydxdy[m]> basically it tries to install the package twice
[2021-09-26T22:09:50Z] <tleydxdy[m]> since some package depends on the alternative directly
[2021-09-26T22:14:52Z] <rio6> yeah I don't like that method either
[2021-09-26T22:15:36Z] <illiliti> does openssl still depend on perl?
[2021-09-26T22:17:47Z] <tleydxdy[m]> I think the fix would be relatively simple
[2021-09-26T22:18:27Z] <tleydxdy[m]> just dereference the symlinks when reading the dependency
[2021-09-26T22:18:42Z] <tleydxdy[m]> rather than using them as is
[2021-09-26T22:33:44Z] <tleydxdy[m]> so I'm reading through the sources of kiss
[2021-09-26T22:34:15Z] <tleydxdy[m]> seems like circular dependencies are not handled
[2021-09-26T22:34:43Z] <tleydxdy[m]> what's the intended way to do it?
[2021-09-26T22:36:02Z] <illiliti> they are handled
[2021-09-26T22:36:07Z] <illiliti> https://github.com/kisslinux/kiss/blob/master/kiss#L552
[2021-09-26T22:36:36Z] <illiliti> it's a very hacky and ugly way but it works
[2021-09-26T22:39:10Z] <tleydxdy[m]> well, die is not really handling it
[2021-09-26T22:40:43Z] <tleydxdy[m]> I guess the right solution would be to ditch the software since it's badly designed
[2021-09-26T22:41:06Z] <tleydxdy[m]> * the software (that has circular dependency) since it's
[2021-09-26T22:41:53Z] <illiliti> die prevents segfault due to stack overflow caused by circular dep
[2021-09-26T22:42:43Z] <tleydxdy[m]> okay, by handling it I mean some way to make the package install correctly 
[2021-09-26T22:45:16Z] <illiliti> circular dep is bug not a feature
[2021-09-26T22:45:45Z] <illiliti> that's why kiss just dies if it sees circular dep
[2021-09-26T22:46:43Z] <illiliti> this is the only *proper* way to handle circular deps imo
[2021-09-26T22:48:18Z] <tleydxdy[m]> > I guess the right solution would be to ditch the software (that has circular dependency) since it's badly designed
[2021-09-26T22:48:18Z] <tleydxdy[m]> so this
[2021-09-26T22:49:49Z] <illiliti> yeah
[2021-09-26T22:50:03Z] <illiliti> see also: freetype-harfbuzz
[2021-09-26T22:50:48Z] <illiliti> https://github.com/kisslinux/repo/tree/master/extra/freetype-harfbuzz
[2021-09-26T22:51:54Z] <tleydxdy[m]> ah ic
[2021-09-26T22:52:05Z] <tleydxdy[m]> so zfs probably need the same thing
[2021-09-26T22:53:06Z] <illiliti> what's the problem with zfs?
[2021-09-26T22:54:58Z] <rio6> all these packages have readme now, nice
[2021-09-26T22:55:13Z] <ryoshu> for what events we get device change in udev?
[2021-09-26T22:55:32Z] <ryoshu> network link up/down?
[2021-09-26T22:56:53Z] <illiliti> ryoshu: display brightness change
[2021-09-26T22:57:35Z] <ryoshu> hmm, is there a comletr 
[2021-09-26T22:57:39Z] <ryoshu> complrtr
[2021-09-26T22:57:49Z] <ryoshu> complete list somewhere
[2021-09-26T22:58:00Z] <illiliti> idk
[2021-09-26T22:58:09Z] <ryoshu> (sorry screen is wet and misclicks)
[2021-09-26T22:58:38Z] <tleydxdy[m]> zfs packages has circular depencies, but I guess it would make sense to jusr package as one
[2021-09-26T22:59:00Z] <illiliti> ryoshu: all good
[2021-09-26T22:59:26Z] <illiliti> grep change action in linux source
[2021-09-26T23:02:13Z] <illiliti> https://grep.app/search?q=KOBJ_CHANGE&filter[repo][0]=torvalds/linux
[2021-09-26T23:05:12Z] <ryoshu> thanks
[2021-09-26T23:06:16Z] <ryoshu> i am working on enumeration, is there a need to store match patterns in udev_list_entry? i used a dedicated data structure
[2021-09-26T23:06:42Z] <ryoshu> list, but i plan to switch to rbtree
[2021-09-26T23:07:34Z] <illiliti> API hardcodes udev_list_entry
[2021-09-26T23:07:42Z] <ryoshu>  in future, for now the goal is to get it operational
[2021-09-26T23:07:42Z] <illiliti> you must comply
[2021-09-26T23:08:43Z] <ryoshu> in what call hardcodes?
[2021-09-26T23:09:00Z] <ryoshu> i plan to return to the code in 0.5h
[2021-09-26T23:09:14Z] <illiliti> udev_enumerate_get_list_entry
[2021-09-26T23:10:39Z] <ryoshu> On success, udev_enumerate_get_list_entry() returns a pointer to the first entry in the list of found devices.
[2021-09-26T23:11:02Z] <ryoshu> but i am thinking about match_subsystem etc
[2021-09-26T23:11:20Z] <ryoshu> match_parent
[2021-09-26T23:11:31Z] <ryoshu> match_property
[2021-09-26T23:11:34Z] <illiliti> the whole match thing is the mess
[2021-09-26T23:11:40Z] <dilyn> tleydxdy[m]: wdym zfs has a circular dep? 
[2021-09-26T23:11:41Z] <illiliti> horrible mess
[2021-09-26T23:12:22Z] <illiliti> i think callback is better than these hardcoded match functions
[2021-09-26T23:12:55Z] <illiliti> anyway, you can copy the logic from libudev-zero code
[2021-09-26T23:15:38Z] <illiliti> ryoshu: also, don't bother with sysattrs
[2021-09-26T23:16:00Z] <illiliti> same for match_parent
[2021-09-26T23:16:48Z] <ryoshu> it uses udev_list_entry and stores value as key, keeping right hand value null
[2021-09-26T23:17:12Z] <ryoshu> kind of surprising design if this is mandated by api
[2021-09-26T23:17:40Z] <illiliti> yeah, the api is shit
[2021-09-26T23:20:06Z] <illiliti> libudev-devd stores devnodes instead of syspaths
[2021-09-26T23:20:36Z] <illiliti> i think you should do this too
[2021-09-26T23:21:07Z] <ryoshu> i only keep /dev/nodeX
[2021-09-26T23:21:22Z] <ryoshu> major minor unit (same as unit?)
[2021-09-26T23:22:10Z] <ryoshu> i need to submit from the kernel subsystem and input discrimination data, like touchpad or mice
[2021-09-26T23:22:40Z] <ryoshu>  /sys is trimmeed fully
[2021-09-26T23:23:18Z] <illiliti> good
[2021-09-26T23:23:56Z] <ryoshu> also parent, but the kernel version is i think buggy
[2021-09-26T23:24:35Z] <ryoshu> but as no api tests it, we get what we have
[2021-09-26T23:26:25Z] <ryoshu> is refcnt design necessary for the api? are there some wins with it?
[2021-09-26T23:27:40Z] <ryoshu> except being fancy
[2021-09-26T23:29:40Z] <illiliti> it's useful if you pass context between many libraries. so you don't need to manually track when you need free context
[2021-09-26T23:30:22Z] <illiliti> it's also useful keeping parent device of udev_device
[2021-09-26T23:30:58Z] <ryoshu> but the library is generally not thread safe, am i right?
[2021-09-26T23:31:01Z] <ryoshu> by design
[2021-09-26T23:31:08Z] <illiliti> yes
[2021-09-26T23:32:59Z] <ryoshu> is this a problem ever?
[2021-09-26T23:33:15Z] <ryoshu> in practice
[2021-09-26T23:33:41Z] <illiliti> no
[2021-09-26T23:33:54Z] <ryoshu> okay, thank you
[2021-09-26T23:35:04Z] <illiliti> np
[2021-09-26T23:37:44Z] <illiliti> refcnt is basically a primitive implementation of garbage colletion