justanotheruser has quit [Ping timeout: 264 seconds]
justanotheruser has joined #nixos-dev
<aanderse>
jtojnar: ping
ris has quit [Ping timeout: 258 seconds]
drakonis has joined #nixos-dev
drakonis1 has quit [Quit: WeeChat 2.8]
drakonis_ has quit [Ping timeout: 265 seconds]
<jtojnar>
aaron?
<aanderse>
hey
<aanderse>
about your httpd-php pr
<aanderse>
1 sec
<aanderse>
in the httpd module you add this line: cat ${php.customConfig} > $out
<aanderse>
i was originally going to modify the module to use PHP_INI_SCAN_DIR to be consistent with the rest of the php implementations
drakonis_ has joined #nixos-dev
<aanderse>
but what you have is more flexible
<aanderse>
any thoughts on that?
<jtojnar>
aaron: do you mean running httpd with the environment variable set?
<aanderse>
Jan Tojnar: yes
<jtojnar>
I would worry about apache filtering environment variables it does not know
<aanderse>
PHPRC is already set, so i figured just add PHP_INI_SCAN_DIR on top of that and it would be consistent with the use of PHP_INI_SCAN_DIR for both php and php-fpm
<aanderse>
Jan Tojnar: using PHP_INI_SCAN_DIR definitely works... but it suffers the same flaw that both php and php-fpm suffer - the user can't set PHP_INI_SCAN_DIR then
<aanderse>
consistent, but far less flexible
<jtojnar>
not sure how often someone wants to use PHP_INI_SCAN_DIR instead of just including the config into the main one
<jtojnar>
does not sound like something reasonable to do
<jtojnar>
but environment variables are still icky
<aanderse>
using mod_php is icky... and effectively requires PHPRC as an environment variable anyways
<aanderse>
i said your method is more flexible, so probably better...
<aanderse>
but i wanted to hear your opinion on what i was thinking to do if you didn't mind sharing
<jtojnar>
yeah, I like the concatenation more than another environment variable
<aanderse>
ok sounds good
<aanderse>
thanks!
<jtojnar>
on that note, would it be possible to use `php -c` in the wrapper instead of PHP_INI_SCAN_DIR?
<jtojnar>
docs are not clear if it adds a config file or sets it
<jtojnar>
hmm, looks like it resets it 😔️
phreedom has quit [Remote host closed the connection]
<gchristensen>
worldofpeace: LnL mentioned he might start on a pipeline this weekend to automatically mark packages as broken in master, so hopefully your 20.09 release will be much more buttoned up :)
<{^_^}>
"An implementation of a subset of the pwd.h, group.h and shadow.h family of functions."
<cole-h>
Oh, maybe tehre isn't a difference lol
<cole-h>
Nope it's an alias... 🤦
<cole-h>
"alias"
<cole-h>
gchristensen: Is it important? Why wouldn't we be using {pwd,shadow}.h from glibc? (group.h isn't in glibc, but appears to be in my linux-headers on Arch)
* cole-h
is admittedly not that knowledgable about the lower-level workings, as he is now
<gchristensen>
I take it back :)
<cole-h>
And for extra "shouldn't really matter": nothing appears to directly use nsss in nixpkgs
<gchristensen>
beautiful
<cole-h>
`rg -uu 'nsss'` are either sha256s, nsss's definition and derivation, and TODOs regarding support for it
<Profpatsch>
worldofpeace: what about skawarePackages?
<Profpatsch>
Should all be green, I updated them recently
<Profpatsch>
I try to keep them well-maintained
<Profpatsch>
gchristensen: ofborg finds the maintainer by unsafeGetAttrPos right?
<arianvp>
niksnut: I see in Nix 2.0 nix-env has a replacement called `nix profile <install|remove>`
<arianvp>
can we also add an nix-env --set equivalent? where your profile is bound to a specifc nix flake (that calls buildEnv or whatever) and every time you call `nix profile upgrade` it updates the entire thing?
<arianvp>
I guess you could just make sure you only install that specific flake; but forcing the profile to contain exactly one flake might be useful for people who want to declaratively describe their environment
<arianvp>
wait let me read the source doe first. maybe this is already what is happening :P
<niksnut>
there already is, nix build --profile
teto has joined #nixos-dev
Jackneill has quit [Ping timeout: 250 seconds]
Jackneill has joined #nixos-dev
chagra has quit [Ping timeout: 260 seconds]
chagra has joined #nixos-dev
drakonis has joined #nixos-dev
drakonis_ has quit [Ping timeout: 264 seconds]
drakonis has quit [Read error: Connection reset by peer]
<worldofpeace>
Profpatsch: I think only two were marked broken on stable. and I think it was because of that commit I mentioned, which you reverted on master so they now build
<worldofpeace>
arianvp: whatt? I need to look at the nix repo sometime 🤣
<worldofpeace>
garbas: awesome.
drakonis_ has joined #nixos-dev
drakonis has quit [Ping timeout: 256 seconds]
drakonis1 has joined #nixos-dev
ixxie has joined #nixos-dev
drakonis_ has quit [Ping timeout: 264 seconds]
phreedom has quit [Ping timeout: 240 seconds]
phreedom has joined #nixos-dev
chagra has quit [Ping timeout: 240 seconds]
<arianvp>
niksnut: awesome!
chagra has joined #nixos-dev
<Profpatsch>
worldofpeace: kk
<Profpatsch>
I’ll backport if I notice something.
drakonis has joined #nixos-dev
justanotheruser has quit [Ping timeout: 246 seconds]
<Profpatsch>
worldofpeace: does it need to be backported? I think you removed it from your PR that marks stuff as broken
drakonis has quit [Ping timeout: 272 seconds]
teto has quit [Ping timeout: 246 seconds]
<worldofpeace>
Profpatsch: it was removed from my PR on master because that commit is on master. In this commit you say "The patch removed `outputs` for some reason
<worldofpeace>
(possible oversight), so building nsss failed." and on stable I marked nsss as broken, so this commit backported would unbreak that (I think)
<worldofpeace>
anyways, I'm hoping marking things as broken is done by a robot so I don't become the correspondent for every package in nixpkgs :D
<sphalerite>
andi-: would you happen to know things about this? :)
<andi->
sphalerite: I haven't invested the time yet
<sphalerite>
alright, no problem. I might have a look this weekend if I succeed at making the time :)
<sphalerite>
(though I'm not really sure where to start yet)
<andi->
I have the feeling we just have to add that libOpenGL fixup hook or whatever that was we have ther that reidrects searches to /run/opengl-drivers
<sphalerite>
hm, if that were the case setting LD_LIBRARY_PATH to that would make it work, wouldn't it?
<sphalerite>
also, strace shows that it does successfully open a libEGL.so.1 (from libglvnd)
<sphalerite>
lol, ltrace segfaulted when I tried to trace firefox's use of that lib
<cole-h>
`addOpenGLRunpath` is the hook I think
<sphalerite>
nope that didn't do the trick
tdeo has quit [Remote host closed the connection]
drakonis1 has joined #nixos-dev
tdeo has joined #nixos-dev
tdeo has joined #nixos-dev
tdeo has quit [Changing host]
chagra has joined #nixos-dev
drakonis_ has quit [Ping timeout: 258 seconds]
drakonis_ has joined #nixos-dev
drakonis1 has quit [Ping timeout: 252 seconds]
<pie_[bnc]>
for your consideration: recently published google book (thats libgen) 93.174.95.29/main/239A340B0D274937627136B6DBC004F0 Building Secure and Reliable Systems: Best Practices for Designing, Implementing, and Maintaining Systems
<srk>
is it good? I've only skimmed it so far
<pie_[bnc]>
dunno yet
<pie_[bnc]>
srk: hows the skim
<srk>
not much formal methods
<pie_[bnc]>
eh. i can probably forgive them for not being a haskell type system book
<pie_[bnc]>
srk: the table of contents has a section that uses the word "resilience" thats already a +1 from me xD
<srk>
:))
<pie_[bnc]>
srk: right now i happen to be looking for stuff like "use CI, dont merge to master, write tests", if you have any suggestions
<srk>
not really, only browsed it on a phone for a bit
<pie_[bnc]>
i mean not related specifically to the book
<srk>
trying to work my way thru TAPL currently, very slowly
<srk>
my experimental uniplate CAS thingie turns out quite complicated when you try to test the language with quickCheck so I've started exploring Agda and TAPL to learn some new things
<srk>
it was good I've learned uniplate a bit as hnix is based around it and it is now a bit less painful to use :D
<pie_[bnc]>
nice, i forgot about that book
<srk>
it even has a section on dependent types! albeit a short one
<pie_[bnc]>
re: agda, i *started* but didnt have time to get far with Verified Functional PRogramming in Agda , but this might be more suited for -chat :D
<srk>
hah, ok :)
<pie_[bnc]>
at the moment im looking for resources on how to not be shitty at software engineering in general
<pie_[bnc]>
something something "software is hard, use all the help you can get"
<srk>
same same, it's a constant learning experience :)
<pie_[bnc]>
yeah but having to cobble thigs together from random resources maybe could be more efficient
<pie_[bnc]>
hence my neverending quest for references
<pie_[bnc]>
though it would be nice if i stopped to actually read sometimes
<srk>
thanks for the polysemy link btw, I'm gonna try that as one of my projects is stuck on lift proliferation
WilliButz has quit [Remote host closed the connection]
WilliButz has joined #nixos-dev
<matthewbauer>
worldofpeace: it might be useful to have in there for 20.09 release - i'm not sure if we're marking broken the same things over and over again
__monty__ has quit [Quit: leaving]
<worldofpeace>
matthewbauer: well, I considering that marking things as broken as we do for a branched off 20.03 is much much different than even touching master
<worldofpeace>
I think gchristensen mentioned LnL might work on having that happen continuously
<LnL>
yeah, or at least some of the building blocks needed to build the automation
<worldofpeace>
It seems a lot of people are now (seems sudden) interested having that be a thing, so "building block wanted"
<worldofpeace>
gchristensen: so we're going to share lego's
<cole-h>
Heh
<gchristensen>
sometimes it is hard to share legos
<gchristensen>
but the cool thing about projects as big as ours is there are always more legos to play with
ixxie_ has quit [Quit: leaving]
ixxie_ has joined #nixos-dev
ixxie_ has quit [Client Quit]
ixxie_ has joined #nixos-dev
ixxie_ has quit [Client Quit]
ixxie_ has joined #nixos-dev
<LnL>
right but this is pretty opaque, just exposing a stream of bad packages with an indication of how bad
<LnL>
consumers could do many things: notify maitainers, mark as broken, bisect regressions, ...
drakonis_ has quit [Ping timeout: 256 seconds]
<infinisil>
srk: I have a bunch of practice with polysemy, it's pretty neat
drakonis_ has joined #nixos-dev
drakonis has quit [Quit: WeeChat 2.8]
justanotheruser has quit [Ping timeout: 265 seconds]
<ryantm>
Is there something like "wrapProgram" that wraps every binary in a directory, or should I just do that with a script?
justanotheruser has joined #nixos-dev
teto has quit [Ping timeout: 272 seconds]
teto has joined #nixos-dev
<pie_[bnc]>
ryantm: use globbing on the p-wait no thats not how that works