| 4 Aug 2023 |
 Lily Foster | In reply to @raitobezarius:matrix.org
I remember roughly we had discussions about it with Lily Foster a long time ago We did? š | 13:37:21 |
| Lily Foster | I have no memory of that (but that's not that surprising...) | 13:37:31 |
 raitobezarius | I remember we had the discussions on the closure size of nodejs-slim vs nodejs | 13:37:33 |
| raitobezarius | maybe in NixOS and not Nix NodeJS | 13:37:38 |
 gdamjan | In reply to @lily:lily.flowers
Does -slim include headers? it does. I've tried removing them in some post* hook, didn't change the deps | 13:37:46 |
| Lily Foster | Nix why depends can tell you the exact files with the reference iirc | 13:38:15 |
| Lily Foster | So knowing which files may help | 13:38:20 |
| gdamjan | let me see,first time I hear about nix why-depends :D | 13:39:20 |
| Lily Foster | Maybe it's pkg-config files though? | 13:39:22 |
| Lily Foster | Nodejs doesn't have a dev output right now, correct? | 13:39:41 |
| Lily Foster | If it did, headers and pkg config files and such would go to it | 13:39:51 |
| Lily Foster | So maybe it should if that's why | 13:40:01 |
| gdamjan | maybe I'm using this wrong:
# nix why-depends nixpkgs#nodejs-slim nixpkgs#openssl.dev
/nix/store/5vm8fh4q21xlk7i1y6rqykqar296dglr-nodejs-slim-18.16.1
<E2><94><94><E2><94><80><E2><94><80><E2><94><80>/nix/store/l0vnx3ik63qrlj086gq06fjmalsi5dpy-openssl-3.0.9-dev
?
| 13:43:08 |
| gdamjan | let me try a real VM :/ | 13:45:12 |
| Lily Foster | It's a flag i think. By default it just shows the chain of derivation/outpath references | 13:45:12 |
| Lily Foster | If I wasn't on mobile abd at $dayjob I'd check myself | 13:45:38 |
| Lily Foster | * If I wasn't on mobile and at $dayjob I'd check myself | 13:45:44 |
| gdamjan | sure - take care :D | 13:45:59 |
| Lily Foster | (I mean I work for a state government in the US and it's a friday -- I can be on mobile, it's not like I'm getting much done anyway š
) | 13:46:39 |
| gdamjan | --precise | 13:46:41 |
| gdamjan | is there a way to tell it not to shorten strings like this bin/node: ā¦.46.0-dev/include", "/nix/store/l0vnx3ik63qrlj086gq06fjmalsi5dpy-openssl-3.0.9-dev/include", "/nā¦ ? | 13:48:17 |
| Lily Foster | In reply to @gdamjan:spodeli.org
is there a way to tell it not to shorten strings like this bin/node: ā¦.46.0-dev/include", "/nix/store/l0vnx3ik63qrlj086gq06fjmalsi5dpy-openssl-3.0.9-dev/include", "/nā¦ ? Hmmmmm not sure. Also why would the include paths be embedded in the binary???? | 13:49:09 |
| Lily Foster | Unless it's wrapped and that's a wrapper. I forget if it's wrapped | 13:49:25 |
| gdamjan | /nix/store/5vm8fh4q21xlk7i1y6rqykqar296dglr-nodejs-slim-18.16.1/bin/node is 45M ā¦ seems like the real deal :D | 13:50:28 |
| gdamjan | ok thanks for why-depends, let me make some experiments | 13:51:16 |
| Lily Foster | In reply to @gdamjan:spodeli.org
/nix/store/5vm8fh4q21xlk7i1y6rqykqar296dglr-nodejs-slim-18.16.1/bin/node is 45M ā¦ seems like the real deal :D Damn, I wonder if it just doesn't have LTO or anything on it. Highly templated C++ stuff can explode in size for sure, especially without that | 13:52:24 |
| Lily Foster | But 45 MB is a lot... | 13:52:32 |
| gdamjan | /usr/bin/node is 39M on Arch ā¦ I guess a static library or two difference?! | 13:53:35 |
| Lily Foster | Maybe | 13:55:39 |
| gdamjan | this is gonna compile for a while ā¦ | 13:56:33 |
