|
Suspicious Dish posted:btw mr. bsd i'm still waiting your reply about systemd and how "it's not unix" and why that matters and for details about your supposed pulseaudio crashes. if they happened today you should still have the core dumps around and be able to paste a backtrace. systems is weird and no portable and lots of people, have made those arguments before No I do not have core dumps from pulse.
|
# ? Dec 3, 2013 08:06 |
|
|
# ? Jun 5, 2024 07:09 |
|
yes, it's not portable to aix and solaris because it uses a lot of linux-specific functionality like cgroups, fanotify, inotify, tmpfs, devtmpfs, statfs, udev, netlink, /sys, waitid, etc. who the gently caress cares. nobody is putting any r&d into anything but linux and why would we spend time supporting platforms with no features in the name of portability. systemd finally allows us to bring service management into the late 90s by having a central place for system logging for services, a way to stop services so all of their forked children don't become zombies, smart ways to manage services (so you can say that nginx and postgres can both have 50% cpu and one can't stall the other with i/o or something like that). these things aren't possible without hooking into the kernel and poking at kernel-level apis, so what's the point of porting to hp/ux. also, yes, it is "weird". so is your dumb face.
|
# ? Dec 3, 2013 08:14 |
|
Notorious b.s.d. posted:systems is weird and no portable and lots of people, have made those arguments before systemd owns, get hosed if you want to stick with 1980s shell scripts for everything sysvinit is terrible
|
# ? Dec 3, 2013 10:17 |
|
Notorious b.s.d. posted:systems is weird and no portable and lots of people, have made those arguments before i love systemd and the loving antis never seem to come up with any logic against it other than "it offends my aesthetic sensibilities" and "it is a red hat project being aggressively promoted via political and not technical channels" (which is clearly a load of complete poo poo). i think the closest thing to a reasonable argument they have is that "pid 1 does too much stuff now", and i don't buy it anyway because it's all logically related (being the hub of a systemwide and kernel-supported D-Bus might be a bit much, though. Initially I'd have thought that they should have just used libdbus as an implementation detail and just privately used the D-Bus wire protocol for IPC, however systemd's job is launching and supervising system processes, and D-Bus activation is one of the major ways that gets initiated, so it makes sense. linux needs something like a system-wide and performant D-Bus, otherwise there'll be five different half-baked D-Bus equivalents that don't interoperate, which is useless). anyway why does a piece of software that is a major part of linux's unique identity need to be "portable" to other operating systems? this is a system service manager, not an office suite. take this argument to its logical conclusion: what if every kernel and piece of core system software could be mixed freely? what would be the point of having a slate of five operating systems that are independently-developed but otherwise functionally identical to each other? that would just be a ridiculous amount of duplicated effort. you seem like you're probably one of those "linux is about choice" guys so why would you deny people the choice of a system that actually uses a post-1983 kernel feature or two because said features aren't supported by your particular People's Front of Judea fork of BSD? i mean i see this come up time and time again in Debian and it really pisses me off, because these people really think that having multiple selectable kernels is something that users will (a) even give a poo poo about and (b) have an opinion about other than "the one that works". a kernel's job is to drive hardware and securely multiplex the resources provided by that hardware. nobody gives a flying gently caress about HURD, it's been 20 years, rms' project has utterly failed. nobody gives a gently caress about a FreeBSD kernel on a linux userland, they can use the FreeBSD kernel with the FreeBSD userland it was loving designed for and compile their ~*portable*~ user apps on top of that. people give a gently caress about having one thing that more or less works that everybody agrees on and can rely upon as a stable foundation for something that's actually useful to an end user, not a free choice between three things that DON'T work because a bunch of people like to turbonerd about uncompromisingly different answers to important questions like "what shall we call the file that stores the system hostname" Sapozhnik fucked around with this message at 10:30 on Dec 3, 2013 |
# ? Dec 3, 2013 10:22 |
|
lol at the backend programmer guy getting mad about UI. gnome 3 owns.
|
# ? Dec 3, 2013 11:26 |
|
Any system config that's not fully scriptable is gonna get in my way eventually so why try it in the first place
|
# ? Dec 3, 2013 12:25 |
|
today i drew a fractal http://scratch.mit.edu/projects/14957030/#editor
|
# ? Dec 3, 2013 13:46 |
|
no turtle sprite
|
# ? Dec 3, 2013 14:04 |
|
Mr Dog posted:i think the closest thing to a reasonable argument they have is that "pid 1 does too much stuff now", and i don't buy it anyway because it's all logically related (being the hub of a systemwide and kernel-supported D-Bus might be a bit much, though. Initially I'd have thought that they should have just used libdbus as an implementation detail and just privately used the D-Bus wire protocol for IPC, however systemd's job is launching and supervising system processes, and D-Bus activation is one of the major ways that gets initiated, so it makes sense. linux needs something like a system-wide and performant D-Bus, otherwise there'll be five different half-baked D-Bus equivalents that don't interoperate, which is useless). none of this stuff is inside pid1. it's all in a separate systemd-dbus daemon. and most of the dirty work will be inside the kernel in kdbus, which provides a good datagram message passing solution. i actually think this is the biggest problem with systemd, right now: branding. systemd is the name of the binary pid1, but also the name of the project as a whole. 90% of the work going on in the project is not in the pid1. the pid1 is actually fairly small and doesn't do a lot, and isn't where any active development is going, but because people hear "this is going in systemd", they think "this is going in /usr/bin/systemd", when it's not.
|
# ? Dec 3, 2013 16:39 |
|
tef posted:today i drew a fractal http://scratch.mit.edu/projects/14957030/#editor that little fox just does to dance
|
# ? Dec 3, 2013 16:40 |
|
Brain Candy posted:no turtle sprite http://scratch.mit.edu/projects/14957030/#fullscreen fixed and it now waggles
|
# ? Dec 3, 2013 17:11 |
|
how much is that turtle in the window
|
# ? Dec 3, 2013 17:13 |
|
Suspicious Dish posted:that little fox just does to dance it's a catte: http://wiki.scratch.mit.edu/wiki/Scratch_Cat
|
# ? Dec 3, 2013 17:15 |
|
kids do really neat things http://scratch.mit.edu/projects/14292779/ the code is :O
|
# ? Dec 3, 2013 17:19 |
|
i wrote some actual elisp for the first time today dynamic scoping is terrible and so is the lack of real namespacing
|
# ? Dec 3, 2013 17:38 |
|
The major problem with Gnome is that it's nearly impossible to find out what the gently caress is going on. Things fail and you have no idea how to even report a bug as it could be anything that failed. For a while logging out stopped working. hosed if I know which component broke because it sure as hell ain't just sending a signal to the session manager. Which brings me to another complaint, I used to be able to manage the session manager through a GUI, now there are files scattered everywhere on the system, some are in system level root only writable files, if I want to log in and not have a bunch of poo poo running I have to switch to root? I mean, I like the autoruns managed through .desktop files, but not like this. gnome-keyring-manager seems to be the worst offender in this case. Also the online account stuff is terrible and simply doesn't work.
|
# ? Dec 3, 2013 17:59 |
|
Opinion Haver posted:i wrote some actual elisp for the first time today doesn't elisp support lexical scoping as of emacs 24?
|
# ? Dec 3, 2013 18:05 |
|
Deus Rex posted:doesn't elisp support lexical scoping as of emacs 24? yes but it's still terrible i was doing some stuff with closures and i couldn't figure out why the hell it wasn't working until i remembered
|
# ? Dec 3, 2013 18:07 |
|
Zombywuf posted:The major problem with Gnome is that it's nearly impossible to find out what the gently caress is going on. Things fail and you have no idea how to even report a bug as it could be anything that failed. For a while logging out stopped working. hosed if I know which component broke because it sure as hell ain't just sending a signal to the session manager. Which brings me to another complaint, I used to be able to manage the session manager through a GUI, now there are files scattered everywhere on the system, some are in system level root only writable files, if I want to log in and not have a bunch of poo poo running I have to switch to root? I mean, I like the autoruns managed through .desktop files, but not like this. gnome-keyring-manager seems to be the worst offender in this case. the major problem with gnome is that it exists in a linux-on-the-desktop-in-tyool-21XX-vacuum where none of its goals or purposes is clear. if gnome ceased to exist tomorrow I truly suspect it would be a net positive for redhat and at best a wash for everyone except possibly suspicious dish. no one us asking for it. the industry tasks Linux is performing were perfectly well handled by fvwm2
|
# ? Dec 3, 2013 18:25 |
|
Deus Rex posted:doesn't elisp support lexical scoping as of emacs 24?
|
# ? Dec 3, 2013 18:38 |
|
Notorious b.s.d. posted:no one should take it personally i'm glad that you guys seem to be less openly hostile now; i was getting uneasy we can all be friends here
|
# ? Dec 3, 2013 18:41 |
|
Suspicious Dish posted:i work for red hat. feel free to complain about rhel all you want at me. the company i'm at uses centos for their linux, and i feel a little guilty about it
|
# ? Dec 3, 2013 18:43 |
|
tef posted:today i drew a fractal http://scratch.mit.edu/projects/14957030/#editor the little wiggling turtle legs are
|
# ? Dec 3, 2013 18:44 |
|
Zombywuf posted:The major problem with Gnome is that it's nearly impossible to find out what the gently caress is going on. Things fail and you have no idea how to even report a bug as it could be anything that failed. For a while logging out stopped working. hosed if I know which component broke because it sure as hell ain't just sending a signal to the session manager. File a bug. I don't care where you put it: gnome-shell, gdm, gnome-session, systemd. We'll investigate problems like this, and figure out what component actually broke. Zombywuf posted:Which brings me to another complaint, I used to be able to manage the session manager through a GUI, now there are files scattered everywhere on the system, some are in system level root only writable files, if I want to log in and not have a bunch of poo poo running I have to switch to root? I mean, I like the autoruns managed through .desktop files, but not like this. gnome-keyring-manager seems to be the worst offender in this case. The reason we have multiple modes of operations is because login needs to be carefully ordered as to which components start when. As more and more services are moving into the kernel or alternate daemons this will become less of an issue. So a large part of session startup is carefully managed, because there's 10 or 12 phases that the session manager will go through for login. If you simply want to start an application at runtime, gnome-tweak-tool now contains what gnome-session-properties used to. That will launch apps in the "Applications" phase.
|
# ? Dec 3, 2013 19:04 |
|
lol gnome guys still trying to figure out autoruns. just put shortcuts in the startup folder you goofballs
|
# ? Dec 3, 2013 19:18 |
|
prefect posted:the little wiggling turtle legs are variable scoping, OTOH...
|
# ? Dec 3, 2013 20:01 |
|
As long as we're having gnome chat with an insider, any chance of that whole "grouping windows by application" experiment being considered a failure, so we can have alt-tab back, and not alt-` and alt-esc (which doesn't show the window list) and all that garbage? It really seems like they redesigned alt-tab around people who don't use alt-tab, ruining it for those who do. It's impossible to build a mental model of what's going on with the grouped windows, so you just plain can't use it quickly. Most of my interactions went from a 200ms unthinking thing to a 10 second ordeal where I give up and use the mouse, and angrily try to figure out what I was even doing. btw, I went looking for that ctrl+alt+t shortcut you said firefox had that caused trouble for a default open terminal shortcut, and it doesn't seem to have one, at least anymore, as far as I can tell.
|
# ? Dec 3, 2013 20:10 |
|
crazypenguin posted:As long as we're having gnome chat with an insider, any chance of that whole "grouping windows by application" experiment being considered a failure, so we can have alt-tab back, and not alt-` and alt-esc (which doesn't show the window list) and all that garbage? yep. you can switch the bindings in the keyboard panel to be back to the standard alt-tab.
|
# ? Dec 3, 2013 20:13 |
|
firefox is ctrl+shift+t. it reopens the last closed tab you had. it's super useful.
|
# ? Dec 3, 2013 20:13 |
|
prefect posted:the little wiggling turtle legs are the cutest turtle
|
# ? Dec 3, 2013 20:25 |
|
i don't recall off the top of my head, is there an option or supported extension to limit alt-tab to just the active workspace? also it'd be nice if you could drag entire workspaces up and down the stack in the overview, because they are too tedious to manage in their current form alt-tab is useless on windows too fwiw because it cycles through the 3 or 4 windows i have unminimised and then gets to work on the 39 i have minimised. workspaces are a good way to partition all this poo poo though yeah alt-tab switching applications instead of windows is a bit of a weird thing that nobody seems to have asked for, it just kinda got lumped in with "hey well as long as we're redesigning everything ANYWAY..."
|
# ? Dec 3, 2013 20:58 |
|
just do like osx and make the search work good. cmd-space is my task switcher
|
# ? Dec 3, 2013 20:59 |
|
is it gnome or kde that does that thing where minimized windows will suddenly demand attention and the little minimized icon will start blinking and put itself in every workspace panel and when you click it it turns out it was just lonely or something because that poo poo sucks
|
# ? Dec 3, 2013 22:09 |
|
that would be kde, gnome doesn't have minimized icons
|
# ? Dec 3, 2013 22:31 |
|
crazypenguin posted:As long as we're having gnome chat with an insider, any chance of that whole "grouping windows by application" experiment being considered a failure, so we can have alt-tab back, and not alt-` and alt-esc (which doesn't show the window list) and all that garbage? Shorter form: stop copying MacOS features, especially the terrible ones.
|
# ? Dec 4, 2013 01:47 |
|
Suspicious Dish posted:If you simply want to start an application at runtime, gnome-tweak-tool now contains what gnome-session-properties used to. That will launch apps in the "Applications" phase. Not seeing it in gnome-tweak-tool.
|
# ? Dec 4, 2013 01:49 |
|
it's here for me.
|
# ? Dec 4, 2013 02:06 |
|
Suspicious Dish posted:
that dashed line is the keyboard focus indicator, not some sort of stitching effect on the button, right?
|
# ? Dec 4, 2013 03:00 |
|
the core problem is that change is intrinsically bad. removing any feature, ever, even if it doesn't fit your ux ~*~ vision ~*~ or is a pain in the rear end to re-implement on your spanking new rewrite, is bad. every change has to justify itself in use cases for existing users. i put up with pulseaudio's crashiness and poor audio quality because it delivers on cool new use cases that make my life, individually, better. systemd is a big set of changes that don't help with any of my use cases. it's a whole bunch of changes i have to cope with, training documents i have to write, scripts i gotta modify, and it doesn't make my life better in any way whatsoever it doesn't matter if it's cool and it helps some laptop user in south america. that's not me, or anyone i know. i'm going to complain bitterly that it has changed 30 year old designs, and my complaints are legitimate.
|
# ? Dec 4, 2013 03:16 |
|
|
# ? Jun 5, 2024 07:09 |
|
on the one hand, i'm glad that legacy unix is dying on the other hand, we threw out a lot of baby with that bathwater. commercial unix had to have volumes of documentation, written by professional tech writers, before they made sweeping changes. even utter turds like CDE had well-documented behavior and reams of training material. change was undertaken infrequently and after long consideration. nothing got rewritten for fun
|
# ? Dec 4, 2013 03:17 |