|
Captain Foo posted:I use this for wireshark they're moving to QT so there's no more xquartz and gtk piss to deal with I've got one of the nightly builds on my phone work machine and it feels way nicer
|
# ? Oct 8, 2014 11:59 |
|
|
# ? May 10, 2024 11:35 |
|
ahmeni posted:they're moving to QT so there's no more xquartz and gtk piss to deal with rad, xquartz seems kind of janky a lot of the time
|
# ? Oct 8, 2014 12:00 |
|
oh hey look systemd is going to replace virtual terminals with systemd-consoled. this is in no way a complete waste of time because
|
# ? Oct 8, 2014 17:35 |
|
.
Sassafras fucked around with this message at 23:21 on Oct 12, 2014 |
# ? Oct 8, 2014 17:42 |
|
Sassafras posted:try to pretend that computers haven't changed since 1960? this is the entire point of Linux
|
# ? Oct 8, 2014 17:43 |
|
Lennart Poettering posted:
I've actually seen people unironically suggest that systemd is a military-industrial complex plot to destroy linux this is what happens when you try to be an adult in a community full of unhinged mentalists, apparently
|
# ? Oct 8, 2014 17:44 |
|
Sassafras posted:How could it possibly make sense to start eliminating ten layers of emulation in ttys that try to pretend that computers haven't changed since 1960? except they aren't really doing this and all the DEC VT emulation, ECMA-48, etc is just being reimplemented because all the bug tested stable as poo poo kernel code is worthless and this needs to be in userspace because of reasons. all this effort to support the 0.5% of people who actually use console ttys (as opposed to terminal emulators, or ssh, or what have you) is obviously worthwhile
|
# ? Oct 8, 2014 17:46 |
|
Mr Dog posted:I've actually seen people unironically suggest that systemd is a military-industrial complex plot to destroy linux Nah, it's just a Redhat plot to claim ownership of Linux.
|
# ? Oct 8, 2014 17:57 |
|
that would probably be the best thing for Linux
|
# ? Oct 8, 2014 18:02 |
|
Zombywuf posted:Nah, it's just a Redhat plot to claim ownership of Linux. If you believe Red Hat management is smart enough to even start to plan a coup like systemd, you're an idiot. When I left they were too busy chasing over the cloud with
|
# ? Oct 8, 2014 19:16 |
|
orphean posted:except they aren't really doing this and all the DEC VT emulation, ECMA-48, etc is just being reimplemented because all the bug tested stable as poo poo kernel code is worthless and this needs to be in userspace because of reasons. all this effort to support the 0.5% of people who actually use console ttys (as opposed to terminal emulators, or ssh, or what have you) is obviously worthwhile lol if you believe fbcon is worth keeping around. fbcon is the reason we have to write a simple hardware blitter / rendering engine in the kernel for every graphics driver we bring up. If you have a UMA system, sure, that's nice and simple, just use the generic bitblt, but if you're on discrete video memory, you have to power up the GPU core and write some assembly code to pump the 2D core to do a bitblt to simply show some text and paint it on the screen. Turns out graphics drivers people are tired of writing that poo poo, and so in userspace it goes so we can maintain the driver *once*.
|
# ? Oct 8, 2014 19:21 |
|
Suspicious Dish posted:If you believe Red Hat management is smart enough to even start to plan a coup like systemd, you're an idiot. When I left they were too busy chasing over the cloud with It doesn't take a lot of smarts to
|
# ? Oct 8, 2014 19:36 |
|
Suspicious Dish posted:lol if you believe fbcon is worth keeping around. fbcon is the reason we have to write a simple hardware blitter / rendering engine in the kernel for every graphics driver we bring up. can't the GPU vendors just subclass the generic video driver and get that for free by implementing a few methods? oh, wait, OO in drivers? that's crazy talk! eschaton fucked around with this message at 19:39 on Oct 8, 2014 |
# ? Oct 8, 2014 19:36 |
|
They do. As I said, if you have a UMA system, you can use the standard blitting methods for the most part. But they still have to write the driver twice: once for fbcon, once for literally everything else. Reminder, fbcon/fbdev are the framebuffer devices that don't support such modern graphics features as "double-buffering" or "vertical blank synchronization". We want to kick it to the curb.
|
# ? Oct 8, 2014 19:40 |
|
Suspicious Dish posted:They do. As I said, if you have a UMA system, you can use the standard blitting methods for the most part. But they still have to write the driver twice: once for fbcon, once for literally everything else. So consoled doesn't use fbdev? How does it work?
|
# ? Oct 8, 2014 19:43 |
|
*consolidates related process supervision functionality into pid 1* You're putting too much stuff into one program!!! Do One Thing Well! *factorises out cruddy kernel-mode functionality and moves it to user-space to better support contemporary GPU hardware* OMG why are you over-engineering core system functionality???
|
# ? Oct 8, 2014 20:09 |
|
Zombywuf posted:So consoled doesn't use fbdev? How does it work? It uses DRM and KMS, the new modesetting API that's actually sane that the rest of the system has used for years.
|
# ? Oct 8, 2014 20:16 |
|
Mr Dog posted:*consolidates related process supervision functionality into pid 1* *bundles newly refactored code into a spaghetti mass of an init system* *tries to look innocent*
|
# ? Oct 8, 2014 20:16 |
|
Zombywuf posted:*bundles newly refactored code into a spaghetti mass of an init system* You've never looked at the systemd codebase if you think it's a spaghetti mess. It's one of the best codebases I've ever worked on.
|
# ? Oct 8, 2014 20:16 |
|
Suspicious Dish posted:You've never looked at the systemd codebase if you think it's a spaghetti mess. It's one of the best codebases I've ever worked on. It's easy to make clean looking code that's still a spaghetti mess. The bit where it gets messy is where you lsof and find that everything just has a socket into dbus and you have 0 introspection without writing a bunch of C to even get the remotest idea what's going on. (yes I still hate dbus)
|
# ? Oct 8, 2014 20:19 |
|
orphean posted:except they aren't really doing this and all the DEC VT emulation, ECMA-48, etc is just being reimplemented because all the bug tested stable as poo poo kernel code is worthless and this needs to be in userspace because of reasons. all this effort to support the 0.5% of people who actually use console ttys (as opposed to terminal emulators, or ssh, or what have you) is obviously worthwhile The most likely outcome is that the user concludes their computer broke and power cycles it. If you're really lucky the distro might have included instructions into getty to press ctrl+alt+f7 to get back to the actual system. But those instructions aren't possible to translate, or even update if there was a successful boot, because all these dumb VTs run in kernel space. It's like when we disabled X11's "press these keys to crash all programs on the computer button" by default and idiots complained that we were removing functionality.
|
# ? Oct 8, 2014 20:29 |
|
What users are just mashing Ctrl+Alt+F1? How does that even happen?
|
# ? Oct 8, 2014 21:02 |
|
Zombywuf posted:What users are just mashing Ctrl+Alt+F1? How does that even happen? Put some ctrl+alt bindings in for various keyboard shortcuts or gaming keys
|
# ? Oct 8, 2014 21:17 |
|
ShadowHawk posted:On my keyboard my F keys are about as far from the number keys as the number keys are from the letters. So a user who's rebound a bunch of keys for special functions? They're pretty rare.
|
# ? Oct 8, 2014 21:19 |
|
also Kinesis supremacy
|
# ? Oct 8, 2014 21:19 |
|
Zombywuf posted:also Kinesis supremacy sorry about your crippling rsi
|
# ? Oct 8, 2014 21:23 |
|
raruler posted:sorry about your crippling rsi I think you're projecting
|
# ? Oct 8, 2014 21:38 |
|
ShadowHawk posted:On my keyboard my F keys are about as far from the number keys as the number keys are from the letters. lol, gaming on a linux imagine.
|
# ? Oct 8, 2014 21:46 |
|
like most desktop related changes, the removal of the useful x server termination shortcut was imposed for the alleged benefit of people who will never use linux, and caused nothing but inconvenience to the people who actually do use it every day. apparently this is progress.
|
# ? Oct 8, 2014 22:15 |
|
Soricidus posted:like most desktop related changes, the removal of the useful x server termination shortcut was imposed for the alleged benefit of people who will never use linux, and caused nothing but inconvenience to the people who actually do use it every day. apparently this is progress. :"these lovely programs are hanging up my x server" :"then kill the x server" :"ugh... i have to do this all the time. can u fix it?" :"here i made u a shortcut to kill the x server" [the lovely programs continue to be lovely]
|
# ? Oct 8, 2014 22:23 |
|
Soricidus posted:like most desktop related changes, the removal of the useful x server termination shortcut was imposed for the alleged benefit of people who will never use linux, and caused nothing but inconvenience to the people who actually do use it every day. apparently this is progress.
|
# ? Oct 8, 2014 22:29 |
|
Soricidus posted:like most desktop related changes, the removal of the useful x server termination shortcut was imposed for the alleged benefit of people who will never use linux, and caused nothing but inconvenience to the people who actually do use it every day. apparently this is progress. And if I did, power cycling my computer wouldn't be that big of a difference anyway.
|
# ? Oct 8, 2014 22:40 |
|
ShadowHawk posted:I use linux every day and haven't felt a need to forcefully kill X in maybe 6 years. I used a Mac for a few years, I felt a need to forcibly kill my display server once every week or so. Usually when I wanted long running processes to not have to be restarted by a reboot. Linux on the other hand, I have not felt the need to restart the X server for many years as you say.
|
# ? Oct 8, 2014 22:45 |
|
i miss ctrl alt backspace but i havent needed it. bascially ShadowHawk posted:I use linux every day and haven't felt a need to forcefully kill X in maybe 6 years.
|
# ? Oct 8, 2014 22:48 |
|
bobbilljim posted:i miss ctrl alt backspace
|
# ? Oct 8, 2014 22:55 |
|
bobbilljim posted:i miss ctrl alt backspace same it was a cool thing that I used to show friends and be smug about
|
# ? Oct 8, 2014 23:00 |
|
Zombywuf posted:I used a Mac for a few years, I felt a need to forcibly kill my display server once every week or so. Usually when I wanted long running processes to not have to be restarted by a reboot. why in the gently caress would you forcibly kill the osx window server just because you're scared of processes running for a week??? you could just log out and then log in. that would do precisely the same thing without danger of data loss. same applies to killing x11. killing the parent process of your entire gui session is unnecessary unless it's actually unresponsive to user input. (also if you're doing either logout or kill of windowserver on osx just loving go ahead and reboot it takes like 10 extra seconds these days jfc, can't say the same about red hate linux tho it still takes forever to reboot for me)
|
# ? Oct 8, 2014 23:06 |
|
BobHoward posted:why in the gently caress would you forcibly kill the osx window server just because you're scared of processes running for a week??? you could just log out and then log in. that would do precisely the same thing without danger of data loss. I think the implication here is that OS X is locked up somehow, and a restart would kill those processes, so if there was yet another escape hatch you could relaunch windowserver
|
# ? Oct 8, 2014 23:10 |
|
Suspicious Dish posted:If you believe Red Hat management is smart enough to even start to plan a coup like systemd, you're an idiot. When I left they were too busy chasing over the cloud with Oh god , you said the magic summoning words and now we'll need to have evol white knighting red hat virt for the next two pages
|
# ? Oct 8, 2014 23:14 |
|
|
# ? May 10, 2024 11:35 |
|
By the way, "RDO" *totally* doesn't stand for "Red Hat's Distribution of OpenStack". No how, no way. It's just "RDO" and it's a distribution of OpenStack that works well with Red Hat software.
|
# ? Oct 8, 2014 23:20 |