2025/05/16

Newest at the top

2025-05-16 20:57:15 +0200 <EvanR> that can't be right
2025-05-16 20:57:12 +0200 <wbrawner> transitive dependencies meaning dependencies of my dependencies, right?
2025-05-16 20:57:02 +0200 <EvanR> in go you download transitive dependencies manually?
2025-05-16 20:56:54 +0200leah2(~leah@vuxu.org) leah2
2025-05-16 20:56:42 +0200 <__monty__> Are we actually using the term "transitive" in the same way? Because I never deal with transitive dependencies directly. That's why they're transitive.
2025-05-16 20:56:35 +0200leah2(~leah@vuxu.org) (Ping timeout: 252 seconds)
2025-05-16 20:56:20 +0200mistivia(~mistivia@user/mistivia) mistivia
2025-05-16 20:55:57 +0200yin(~z@user/zero) zero
2025-05-16 20:55:36 +0200 <wbrawner> But having to do that for a ton of transitive dependencies becomes a chore
2025-05-16 20:55:30 +0200 <EvanR> but I get the impression haskell ecosystem changes faster than normal
2025-05-16 20:55:28 +0200 <wbrawner> well like with go for example, I can just clone the git repos to ~/go/src and then I can pull whatever version I need from there since I have the whole repo locally
2025-05-16 20:55:15 +0200tromp(~textual@2001:1c00:3487:1b00:a44a:50e6:3df5:3b66)
2025-05-16 20:55:12 +0200Tuplanolla(~Tuplanoll@91-159-69-59.elisa-laajakaista.fi) Tuplanolla
2025-05-16 20:55:05 +0200yin(~z@user/zero) (Ping timeout: 272 seconds)
2025-05-16 20:54:52 +0200 <EvanR> (without any obvious benefit)
2025-05-16 20:54:45 +0200Tuplanolla(~Tuplanoll@91-159-69-59.elisa-laajakaista.fi) (Ping timeout: 252 seconds)
2025-05-16 20:54:43 +0200 <EvanR> to update it all would take work
2025-05-16 20:54:42 +0200philopsos(~caecilius@user/philopsos) philopsos
2025-05-16 20:54:38 +0200 <EvanR> some projects are on old version of libraries though
2025-05-16 20:54:26 +0200philopsos(~caecilius@user/philopsos) (Ping timeout: 272 seconds)
2025-05-16 20:54:21 +0200 <EvanR> that sounds like a would be nice situation
2025-05-16 20:54:05 +0200Lord_of_Life(~Lord@user/lord-of-life/x-2819915) Lord_of_Life
2025-05-16 20:53:56 +0200 <wbrawner> I use renovate to keep all of my projects on the latest versions so really I just need the latest version of said dependency and I'm good to go
2025-05-16 20:53:56 +0200 <__monty__> Does the *total* number actually matter? You don't have to manually track down and install/update every transitive dependency with any of the tooling I'm familiar with.
2025-05-16 20:53:50 +0200gorignak(~gorignak@user/gorignak) gorignak
2025-05-16 20:53:39 +0200Lord_of_Life(~Lord@user/lord-of-life/x-2819915) (Ping timeout: 252 seconds)
2025-05-16 20:53:27 +0200gorignak(~gorignak@user/gorignak) (Read error: Connection reset by peer)
2025-05-16 20:53:27 +0200 <EvanR> now I cabal sandbox everything
2025-05-16 20:53:19 +0200 <EvanR> because versions
2025-05-16 20:53:10 +0200 <EvanR> the idea that you just need "the dependency" and any project that needs it will work is alien to me
2025-05-16 20:53:02 +0200mistivia(~mistivia@user/mistivia) (Ping timeout: 244 seconds)
2025-05-16 20:52:49 +0200rembo10(~rembo10@main.remulis.com) rembo10
2025-05-16 20:52:32 +0200rembo10_(~rembo10@main.remulis.com) (Ping timeout: 272 seconds)
2025-05-16 20:52:31 +0200 <wbrawner> Also I rely on renovate to keep dependencies up to date and the node projects I have are the noisiest in that regard
2025-05-16 20:52:13 +0200 <wbrawner> yeah it mostly comes down to total number of dependencies, because I can install a few locally and be ready to go, but when there are dozens needed to do anything useful it's more of a chore to install and make sure they're up-to-date
2025-05-16 20:52:12 +0200user363627(~user@user/user363627) (Remote host closed the connection)
2025-05-16 20:50:53 +0200monochromonce exceeded his data plan by a "simple" "docker pull currylang/kics2". That was 2GB.
2025-05-16 20:50:37 +0200 <EvanR> dang __monty__ asked this already
2025-05-16 20:49:57 +0200 <EvanR> that would make the size of dependencies the problem not discrete number
2025-05-16 20:49:34 +0200 <monochrom> I think it's the "greater than my data plan" problem. :)
2025-05-16 20:48:47 +0200 <EvanR> yes is it the number greater than zero problem or number greater than 10 problem
2025-05-16 20:48:14 +0200 <EvanR> if they combined all the functionality being pulled in into 1 dependency would that be better for you
2025-05-16 20:48:06 +0200 <c_wraith> Yeah, that's not about transitive dependencies. Just dependencies at all
2025-05-16 20:47:57 +0200 <EvanR> wbrawner, is it the number of transitive dependencies that annoys you. Like how in C people will rewrite a library just to avoid *any* dependencies, because that's a number greater than zero
2025-05-16 20:47:53 +0200 <__monty__> What does that have to do with *transitive* dependencies? That sounds like behemoth stdlib v. minimal stdlib.
2025-05-16 20:47:10 +0200 <wbrawner> __monty__ ^
2025-05-16 20:47:04 +0200 <wbrawner> transitive dependencies. I travel a lot and write code offline. If I have go installed I can pretty much write whatever I want without internet but if I try to write something in Node I have to hope that I remember to cache all of the dependencies I _may_ need before I get internet again
2025-05-16 20:47:04 +0200Lord_of_Life(~Lord@user/lord-of-life/x-2819915) Lord_of_Life
2025-05-16 20:47:02 +0200 <monochrom> https://ro-che.info/ccc/23
2025-05-16 20:46:39 +0200ChanServ+v haskellbridge