drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-dev
drakonis1 has joined #nixos-dev
drakonis_ has quit [Read error: Connection reset by peer]
drakonis has quit [Ping timeout: 245 seconds]
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-dev
ivan has quit [Remote host closed the connection]
ivan has joined #nixos-dev
Jackneill has joined #nixos-dev
Jackneill has quit [Remote host closed the connection]
teehemkay has quit [Excess Flood]
teehemkay has joined #nixos-dev
orivej has joined #nixos-dev
Jackneill has joined #nixos-dev
__monty__ has joined #nixos-dev
ris has joined #nixos-dev
jonringer has quit [Ping timeout: 276 seconds]
pie_ has quit [Read error: Connection reset by peer]
pie_ has joined #nixos-dev
orivej has quit [Ping timeout: 265 seconds]
avn has quit [Remote host closed the connection]
orivej has joined #nixos-dev
drakonis1 has quit [Ping timeout: 265 seconds]
drakonis1 has joined #nixos-dev
drakonis1 has quit [Ping timeout: 276 seconds]
drakonis has joined #nixos-dev
jonringer has joined #nixos-dev
jtojnar has quit [Read error: Connection reset by peer]
<niksnut>
ouch, my system closure size went from 5.3 GB in 19.03 to 6.4 GB in 19.09
<arianvp_>
Ouch
<niksnut>
anybody experience the mouse pointer being unusably slow in 19.09?
<ivan>
niksnut: yeah, we switched to libinput by default
<ivan>
services.xserver.libinput.enable = false; to restore sanity
<ivan>
I saw that on my Logitech G9 which should have been moving much faster
<niksnut>
yeah, I have a G502
<niksnut>
I'll try that
<niksnut>
or should there be some udev rules to configure the mouse?
<niksnut>
I notice 'xinput set-prop 8 300 4' (to change acceleration) fails with 'integer parameter out of range for operation'
<niksnut>
so I can't set the acceleration higher than 1
<ivan>
libinput is one of those software with incredibly detailed rationale about why it is better than {libev, synaptics} but is hampered by author testing it on his laptop and 'works on my machine'
<samueldr>
let me review what I proposed, it's been a while
<samueldr>
I think it was that, at build time, any [propagated]BuildInput having the qt plugin path prefix would be added to the registered qt plugin paths
<samueldr>
(I still had some unanswered questions)
<niksnut>
QT_PLUGIN_PATH is less of an issue because lib/qt-*/plugins is unlikely to exist in a build-time-only dependency
<niksnut>
but /share is not very discriminating for XDG_DATA_DIRS
<samueldr>
oh, I hadn't realised that hook also managed XDG_DATA_DIRS
<samueldr>
then I figure the answer is "it wouldn't have bloated as much, but also wouldn't have the same results as it lacked XDG_*_DIRS support" or something like that
<niksnut>
it's also a bit of an issue that it leaks into user shells (e.g. in konsole)
<samueldr>
I am really not fond of all wrappers due to environmental contamination
<niksnut>
yeah
drakonis_ has quit [Ping timeout: 246 seconds]
<niksnut>
we need something like RPATH
<samueldr>
what would be nice is environment variable isolation per store path, so that launching a konsole could add what is needed, but anything launching that is not in that path wouldn't have them... but that's a pipe dream; and I wouldn't dare imagine how to implement that sanely
<samueldr>
(and also breaks so many rules)
drakonis_ has joined #nixos-dev
drakonis has quit [Ping timeout: 246 seconds]
<niksnut>
RPATH-like would mean: add an XDG_DATA_DIRS field to the ELF dynamic section (just like RPATH) and patch qt/gtk to read it
<niksnut>
then a program like konsole can have its own XDG_DATA_DIRS without it being inherited by child processes
<samueldr>
that sounds like an interesting idea
<samueldr>
though how would it fare in situations where it's ran under something like a scripting language
<samueldr>
e.g. packaging an app written in python, that uses Qt, and relies on QT_PLUGIN_PATH being set right
<niksnut>
there you'd need to automatically (but in a language-specific way) patch the script to call some function to set the search path
<samueldr>
if the library (e.g. Qt) relies on a well-known store path holding the data, the runtime (e.g. python, lua, ruby) doesn't have to know about it
<samueldr>
though it's maybe less generic, as you can't override any environment variable, only pre-determined ones
<niksnut>
it could look for a path relative to the store path containing the executable, like /nix-support/qt-plugin-path