|
looks like cpuid ran into a few cpuegos
|
#
?
Dec 29, 2017 03:16
|
|
|
|
| # ? Apr 23, 2024 12:27 |
|
|
nobody can provide a patch because nobody has one of these fukcing museum pieces to hand for testing with but then a modern linux distribution will, in the course of being compiled from source, check for the existence of unistd.h approximately 400,000 times, so actually i suppose this is perfectly in keeping with the rest of the system.
|
|
#
?
Dec 29, 2017 03:34
|
|
|
Sapozhnik posted:nobody can provide a patch because nobody has one of these fukcing museum pieces to hand for testing with the patch is really loving simple: execute the check for cpuid support, and exit(1) if the support check fails.
|
|
#
?
Dec 29, 2017 03:52
|
|
|
Sapozhnik posted:nobody can provide a patch because nobody has one of these fukcing museum pieces to hand for testing with except the filer did have one of those fukcing museum pieces to hand and they did test it and someone is sending a patch lmao
|
|
#
?
Dec 29, 2017 04:44
|
|
|
bear in mind this is a runtime issue. systemd is dumping core if an x86 cpu lacks cpuid. that's not good. if poo poo doesn't compile on your 486, i don't really care
|
|
#
?
Dec 29, 2017 04:49
|
|
|
Sapozhnik posted:nobody can provide a patch because nobody has one of these fukcing museum pieces to hand for testing with LuigiThirty make the patch and also make systemd draw rotating cubes
|
|
#
?
Dec 29, 2017 06:28
|
|
|
Notorious b.s.d. posted:the patch is really loving simple: execute the check for cpuid support, and exit(1) if the support check fails. no, I’m pretty sure the patch should be “if the chip in question doesn’t support CPUID, return some default value” not “exit(1)” (probably something I should have done in my patch to add 68040 cache support to MINIX)
|
|
#
?
Dec 29, 2017 06:45
|
|
|
eschaton posted:LuigiThirty make the patch make the linux kernel qr code bounce around and morph like the windows screensaver
|
|
#
?
Dec 29, 2017 06:58
|
|
|
tbh that's cool that someone is trying it out on a 486 in the first place
|
|
#
?
Dec 29, 2017 10:26
|
|
|
from skimming the patch it looks like theyre just using cpuid to infer whether systemd is running in a VM anyway: https://github.com/systemd/systemd/pull/7758/files
|
|
#
?
Dec 29, 2017 10:30
|
|
|
eschaton posted:no, I’m pretty sure the patch should be “if the chip in question doesn’t support CPUID, return some default value” not “exit(1)” ok this is fine too just, core dumping is definitely the wrong behavior
|
|
#
?
Dec 29, 2017 17:00
|
|
|
eschaton posted:LuigiThirty make the patch dnf install systemd-skulltrumpetd
|
|
#
?
Dec 29, 2017 18:45
|
|
|
Soricidus posted:dnf install systemd-skulltrumpetd slot it into compizconfig next to Paint Fire On The Screen
|
|
#
?
Dec 29, 2017 19:32
|
|
|
xiaomi laptop trip report: everything works on arch, including dynamic gpu with bumblebee coming from arch on a macbook air, it’s simple overall the quality/specs can’t be beat for 850 usd
|
|
#
?
Dec 29, 2017 20:06
|
|
|
VikingofRock posted:looks like cpuid ran into a few cpuegos
|
|
#
?
Dec 29, 2017 21:22
|
|
|
the sound for my lovely-rear end hobby game that im writing is really lagged on linux. works fine on windows. its honestly probably the version of sdl . I still blame linux because, you know, audio
|
|
#
?
Dec 29, 2017 23:50
|
|
|
Best Bi Geek Squid posted:the sound for my lovely-rear end hobby game that im writing is really lagged on linux. works fine on windows. if you mean that it cuts out you probably just want to up the number of samples requested with sdl_openaudio. windows is actually pretty good by default on that stuff, as the thread doing the audio operates under a separate scheduler (the multimedia class scheduler service) tuned for that task
|
|
#
?
Dec 30, 2017 00:34
|
|
|
Cybernetic Vermin posted:if you mean that it cuts out you probably just want to up the number of samples requested with sdl_openaudio. windows is actually pretty good by default on that stuff, as the thread doing the audio operates under a separate scheduler (the multimedia class scheduler service) tuned for that task  ill check it out
|
|
#
?
Dec 30, 2017 01:57
|
|
|
Who do I talk to if I want an ubuntu desktop environment built around i3. I know the whole "Linux means choice" crowd particularly glom onto i3, but why can't my one choice be i3 and the rest of my choices be the standard network manager and keyboard shortcut handler.
|
|
#
?
Dec 31, 2017 15:52
|
|
|
TimWinter posted:Who do I talk to if I want an ubuntu desktop environment built around i3. I know the whole "Linux means choice" crowd particularly glom onto i3, but why can't my one choice be i3 and the rest of my choices be the standard network manager and keyboard shortcut handler. kde will allow you to replace its window manager
|
|
#
?
Dec 31, 2017 17:18
|
|
|
TimWinter posted:Who do I talk to if I want an ubuntu desktop environment built around i3. you need to talk to a shrink
|
|
#
?
Dec 31, 2017 17:34
|
|
|
akadajet posted:you need to talk to a shrink beaten
|
|
#
?
Dec 31, 2017 17:46
|
|
|
TimWinter posted:Who do I talk to if I want an ubuntu desktop environment built around i3. I know the whole "Linux means choice" crowd particularly glom onto i3, but why can't my one choice be i3 and the rest of my choices be the standard network manager and keyboard shortcut handler. turns out "linux is a choice" is a meme perpetuated by people building inherently broken, unstable systems. they don't allow you to choose network manager because they think it's bad because it has a point-and-click gui.
|
|
#
?
Dec 31, 2017 18:08
|
|
|
Suspicious Dish posted:turns out "linux is a choice" is a meme perpetuated by people building inherently broken, unstable systems. they don't allow you to choose network manager because they think it's bad because it has a point-and-click gui. the biggest problem with network manager is that it doesn't have a point and click gui. it presents a very shiny api, and it comes out of the box with some extremely unfriendly CLI tools. (e.g. it is impossible to provide a wpa passphrase at the command line!)
the whole thing is kind of a mess. i mean, the api is fine, and network-manager the daemon is really helpful, but it would have been nice to have a more complete reference set of config tools! tl;dr: nmcli is badly incomplete and this makes me sad
|
|
#
?
Dec 31, 2017 18:53
|
|
|
i thought network manager was initially gui only
|
|
#
?
Dec 31, 2017 18:55
|
|
|
no, the guis have always been external to n-m. and the tools that actually came with it, the command line, suck. it is effectively gui-only because the reference tool sucks. but the guis come from outside developers who tie them to other products end result: everything sucks, burn it down
|
|
#
?
Dec 31, 2017 19:05
|
|
|
http://islinuxaboutchoice.com/ https://www.google.com/search?q=nmcli+command+line+passphrase
|
|
#
?
Dec 31, 2017 19:14
|
|
|
Notorious b.s.d. posted:it comes out of the box with some extremely unfriendly CLI tools. Sounds like Linux in a nutshell.
|
|
#
?
Dec 31, 2017 19:24
|
|
|
In the continuing saga of "systemd is cool and good": systemd decided after version 234 to drop Autotools support entirely and go to pure meson/ninja. This caused a huge headache for me because I help with the systemd package with BuildRoot. Problems: - Meson wasn't even supported in BuildRoot until last week. - When Meson was merged in, it created a broken cross-compile config - systemd no longer allowed building unit tests to be skipped. - Converting all of the options from Autotools to meson - Needing python now to build systemd (meson and ninja both require it) Once that was all done, I did some timing tests: Compile times for systemd-236 with meson: real 0m54,483s user 5m6,496s sys 0m24,760s Compile times for systemd-236 with Autotools: real 3m25,001s user 10m39,520s sys 1m10,672s Conclusion: systemd converted to meson because they can't handle using Autotools. I did some digging and that the Autotools build was spending a huge amount of time running the libtool shell script. In conclusion: systemd is cool and good.
|
|
#
?
Dec 31, 2017 21:53
|
|
|
autotools is trash hth systemd is far from the only thing that uses meson and there's a big shift in that direction why do you need bleeding edge systemd in your embedded system, nothing terribly interesting has happened there in the last year or two
|
|
#
?
Dec 31, 2017 21:57
|
|
|
i'm glad you also think your computer has better things to do than run libtool 8,000 times. happy new year
|
|
#
?
Dec 31, 2017 22:08
|
|
|
TimWinter posted:Who do I talk to if I want an ubuntu desktop environment built around i3. I know the whole "Linux means choice" crowd particularly glom onto i3, but why can't my one choice be i3 and the rest of my choices be the standard network manager and keyboard shortcut handler. Ubuntu has something like minimal install, choose that and go from there. It might be called network or server install, I forgot.
|
|
#
?
Dec 31, 2017 22:32
|
|
|
Sapozhnik posted:autotools is trash hth Dynamic users are cool and good.
|
|
#
?
Jan 1, 2018 07:17
|
|
|
i like cmake
|
|
#
?
Jan 1, 2018 08:19
|
|
|
Progressive JPEG posted:i like cmake nobody likes CMake but lol if you don’t accept it as best of a bad lot
|
|
#
?
Jan 1, 2018 09:35
|
|
|
oh yeah that's true
|
|
#
?
Jan 1, 2018 09:36
|
|
|
Sapozhnik posted:autotools is trash hth did they fix the thing where if you have a hosed up rtc pid0 will crash on boot and leave the system unfixable without a full reflash and possibly a clear of the external rtc cache
|
|
#
?
Jan 1, 2018 19:05
|
|
|
Phobeste posted:did they fix the thing where if you have a hosed up rtc pid0 will crash on boot and leave the system unfixable without a full reflash and possibly a clear of the external rtc cache No, because the main issue was awful kernel vendor drivers. https://github.com/systemd/systemd/issues/1143
|
|
#
?
Jan 1, 2018 19:11
|
|
|
lmao if people think old-style init is in any way "defensive programming"
|
|
#
?
Jan 1, 2018 19:47
|
|
|
|
| # ? Apr 23, 2024 12:27 |
|
|
it's defensive in that the programmer is probably extremely defensive about it
|
|
#
?
Jan 2, 2018 02:52
|
|
