gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
<ashkitten> andi-: thanks for trying
drakonis has quit [Quit: Leaving]
pie_ has quit [Ping timeout: 268 seconds]
pie_ has joined #nixos-chat
pie_ has quit [Ping timeout: 240 seconds]
endformationage has quit [Ping timeout: 258 seconds]
Guanin has quit [Ping timeout: 260 seconds]
rardiol has quit [Ping timeout: 240 seconds]
__monty__ has joined #nixos-chat
pie_ has joined #nixos-chat
pie_ has quit [Ping timeout: 258 seconds]
rardiol has joined #nixos-chat
pie_ has joined #nixos-chat
pie_ has quit [Ping timeout: 248 seconds]
pie_ has joined #nixos-chat
Guanin has joined #nixos-chat
jD91mZM2 has quit [Quit: ZNC 1.7.4 - https://znc.in]
jD91mZM2 has joined #nixos-chat
drakonis has joined #nixos-chat
Profpatsch has joined #nixos-chat
Profpatsch has left #nixos-chat ["WeeChat 0.4.3"]
endformationage has joined #nixos-chat
<pie_> is it possible to look up some sort of build expression for a binary?
<pie_> I want to write a tool that rebuilds things with debug info, then extracts the debug info and allows passing it to gdb
<pie_> drv files provide some amount of information but Im assuming its not sufficient for this
<pie_> and im guessing theres no way to do this manner of reverse lookup that I want
<pie_> and its probably not enough to add some debug flags to the drv or something
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos-chat
<ashkitten> question: would it be possible to do some sort of delta-rpm-like thing for nix?
<MichaelRaskin> I think it was supported at some point, but dropped when moving to binary cache model
<ashkitten> ahh
<ashkitten> so it's probably not very technically easy now?
<gchristensen> it is hard to know what to diff it against
<ashkitten> yeah for sure
<ashkitten> i wonder if we could just have nix send the store path of the same attribute from the current generation? i'm not sure if that's easy to figure out though
<ashkitten> and also what about things that aren't part of a generation
<ashkitten> so yeah this is a challenge :/
<MichaelRaskin> ashkitten: also, without bit-perfect reproducibility, if you have the path from another source, a binary patch can go wrong
<ashkitten> oof, yeah
<samueldr> maybe with the CAS it would be more possible, but still, do you diff all 10 last channel updates against the current version?
<ashkitten> so we'd have to send a checksum too
<samueldr> (and that'd only work for what was implemented using content-addressable store paths!)
<ashkitten> this all becomes extremely complicated by the fact that we can have any number of versions of the same package in the store
<samueldr> there is no coordination for software updates, it "just happens" :)
<samueldr> and sometimes it "just happens" to rebuild all your system!
<ashkitten> i suppose we could have the binary cache preemptively diff against the last x number of updates to the attribute, but then we run into the same issue as deltarpm where you don't get the benefits if you don't update all the time
__monty__ has quit [Quit: leaving]
pie_ has quit [Quit: No Ping reply in 180 seconds.]
pie_ has joined #nixos-chat
pie_ has quit [Excess Flood]
pie_ has joined #nixos-chat
pie_ has quit [Ping timeout: 265 seconds]
pie_ has joined #nixos-chat
pie_ has quit [Ping timeout: 248 seconds]