!lymvtcwDJ7ZA9Npq:lix.systems

Lix Development

416 Members
(Technical) development of Lix, the package manager, a Nix implementation. Please be mindful of ongoing technical conversations in this channel.139 Servers

Load older messages


SenderMessageTime
5 Jun 2024
@jade_:matrix.orgjade_i would be tempted to use GHA for this since it is relatively likely to probably be trustworthy01:04:53
@jade_:matrix.orgjade_and is lower effort01:05:00
@lunaphied:lunaphied.meLunaphied
In reply to @pennae:matrix.eno.space
all MD should go through the same preprocessor, and nix binary ideally shouldn't contain an MD processor at all
It'll need one for doc comments but otherwise yeah
02:12:08
@lunaphied:lunaphied.meLunaphied
In reply to @raitobezarius:matrix.org
maybe in that case, what we can do is to have GHA infrastructure perform regular builds and push it to our cache?
This means delegating trust to GHA
02:13:22
@delroth:delroth.netdelrothOh god, I've been looking at the static initializers again, and there are some major crimes being done there. For example, GlobalConfig::Register is being used across shared libraries, so now it's really not trivial to invert the dependency to have a single initialization function.03:31:40
@jade_:matrix.orgjade_sigh05:31:18
@jade_:matrix.orgjade_hm05:31:34
@jade_:matrix.orgjade_you could fix this tho05:31:37
@jade_:matrix.orgjade_you would just have to have init_libWhatever05:31:45
@jade_:matrix.orgjade_and then that gets called on startup05:31:51
@jade_:matrix.orgjade_and you could even have init_libwhatever call its dependent ones, and make them idempotent05:32:10
@delroth:delroth.netdelrothyeah, it's far from impossible, it's just more effort :)05:52:42
@vigress9:matrix.orgV. 🏳️‍⚧️ changed their profile picture.10:49:22
@polygon:catgirl.cloudpolygon system (it/she/they) changed their display name from polygon system (she/it/they) [DECT 7659] to polygon system (it/she/they).11:57:43
@terru:raccoon.collegeterru joined the room.12:32:48
@aprl:uwu.isaprl changed their display name from aprl (GPN; DECT 2774; SIP GAYS) to aprl.12:52:47
@qyriad:katesiria.orgQyriad
In reply to@pennae:matrix.eno.space
all MD should go through the same preprocessor, and nix binary ideally shouldn't contain an MD processor at all
absolutely, but unfortunately this means those docs changes might be blocked on fixing our "docs -> nix binary" build dependency
15:57:20
@pennae:matrix.eno.space⚠️ eldritch horrors operating in this area ⚠️and lunaphied is right about docs rendering for the repl <,<16:12:16
@ippetkov:matrix.orgivan joined the room.16:58:52
@jade_:matrix.orgjade_

so i think i found a clang bug :( this clangd file does not remove the arg -include precompiled-headers.hh

CompileFlags:
  Remove: [-fpch-preprocess, "-include precompiled-headers.hh"]
23:25:32
@qyriad:katesiria.orgQyriadCan you remove two arguments with a single string like that?23:27:41
@jade_:matrix.orgjade_well its ostensibly supposed to remove the canonicalized form23:27:58
@pennae:matrix.eno.space⚠️ eldritch horrors operating in this area ⚠️ (wasn't the canonical form for these =-separated? idr) 23:28:58
@jade_:matrix.orgjade_https://github.com/llvm/llvm-project/blob/8c5a7a1fc4890fcae50f8e8a61d5a2e2b1ebd7e5/clang-tools-extra/clangd/ConfigCompile.cpp#L268-L279 idk but heres your clang code23:29:08
@jade_:matrix.orgjade_i did try that also, i think23:29:14
@jade_:matrix.orgjade_well i have def tried it now, no difference23:29:29
@pennae:matrix.eno.space⚠️ eldritch horrors operating in this area ⚠️oh :(23:29:38
@pennae:matrix.eno.space⚠️ eldritch horrors operating in this area ⚠️great :/23:30:13
@jade_:matrix.orgjade_https://github.com/llvm/llvm-project/blob/aff197ff2163a4b7732d08b833906cc21f4a5c89/clang-tools-extra/clangd/CompileCommands.cpp#L593 part two of the code23:31:22
@jade_:matrix.orgjade_i might just . build a clang and shove a debugger on it. i should not do this bc i have releng scripts to write23:32:29

There are no newer messages yet.


Back to Room ListRoom Version: 10