|
Shaggar posted:maven's package management stuff works. nuget would be ok if it didnt have support for scripting lol maven suffers from java’s problem of only one version of a package being available on the classpath at a given time unless your dependencies generated shaded jars
|
#
?
Jul 28, 2022 16:38
|
|
|
|
| # ? May 27, 2024 08:39 |
|
|
crazypenguin posted:They’re publicly testing the new http index now at least 
|
|
#
?
Jul 28, 2022 16:53
|
|
|
Blinkz0rz posted:maven suffers from java’s problem of only one version of a package being available on the classpath at a given time unless your dependencies generated shaded jars right, it works
|
|
#
?
Jul 28, 2022 16:56
|
|
|
Blinkz0rz posted:maven suffers from java’s problem of only one version of a package being available on the classpath at a given time unless your dependencies generated shaded jars i dont see the problem
|
|
#
?
Jul 28, 2022 17:05
|
|
|
it works until you need to use guava, which is "maintained" by googleoids unfortunately most Java things directly or indirectly use guava. the smarter projects vendor whatever version they're using.
|
|
#
?
Jul 28, 2022 17:09
|
|
|
breaking interface change? sure just run an automated refactor across our gigantic monorepo nbd anyway can't stay and chat i have a promotion package i need to assemble
|
|
#
?
Jul 28, 2022 17:10
|
|
|
i would just not use guava or w/e. i mean i havent written java in years at this point cause .net has completely usurped it, but its not surprising that goog is doing poo poo to gently caress up what was the one decent build system
|
|
#
?
Jul 28, 2022 17:11
|
|
|
Gaukler posted:Are you programming on top of a civil war battle site? Do you naturally produce unsafe levels of RF radiation? good ideas; I'll book appointments with a radiologist and an exorcist right away crazypenguin posted:They’re publicly testing the new http index now at least cocoapod or cargo?
|
|
#
?
Jul 28, 2022 17:37
|
|
|
cool av posted:right, it works i’m the runtime exception that occurs when one of your dependency’s dependency uses a different version of another dependency than another one of your dependencies with a slightly different api that you’ll never see until your app blows up in production
|
|
#
?
Jul 28, 2022 18:01
|
|
|
barkbell posted:new guy on the team opened his first pr to change all the linting rules because he didnt like them. what a power move nudgenudgetilt posted:plot twist: no linter was currently being used and the new guy was proposing they try one out early on at my current job I tried to add some linting rules after discussing it because there were none and people got mad at me.
|
|
#
?
Jul 28, 2022 18:03
|
|
|
Blinkz0rz posted:i’m the runtime exception that occurs when one of your dependency’s dependency uses a different version of another dependency than another one of your dependencies with a slightly different api that you’ll never see until your app blows up in production iirc maven will warn you about this so you can fix it, but since not all version changes are breaking it doesnt error.
|
|
#
?
Jul 28, 2022 18:27
|
|
|
Blinkz0rz posted:i’m the runtime exception that occurs when one of your dependency’s dependency uses a different version of another dependency than another one of your dependencies with a slightly different api that you’ll never see until your app blows up in production im the conceit that dependencies having their own differently versioned copy of their dependencies solves this problem
|
|
#
?
Jul 28, 2022 18:37
|
|
|
this is why i use as few dependencies as possible and why i'll never ever ever EVER touch node
|
|
#
?
Jul 28, 2022 19:34
|
|
|
pokeyman posted:wait, cargo looked at cocoapods's decision to make the package index a git repo that caused github to say "pls no we can't handle the traffic" and said "let's copy that"? And it's not the only newish tool doing this, vcpkg (MS's C++ pkg manager) also does this and I can already feel the performance getting worse linearly over time.
|
|
#
?
Jul 28, 2022 19:36
|
|
|
maven rules. i've never run into an issue with weird mismatched dependencies because i don't touch libraries that bring in that much bullshit.
|
|
#
?
Jul 28, 2022 19:39
|
|
|
tazjin posted:cocoapod or cargo? cocoapods switched over to a http thing a few years ago so i assume cargo
|
|
#
?
Jul 28, 2022 20:17
|
|
|
CRIP EATIN BREAD posted:maven rules. yeah it’s a theoretical problem but it really doesn’t come up much in the real world even with guava, the apis people actually use are pretty stable so I’ve yet to run into a case where naively picking the newest version doesn’t work
|
|
#
?
Jul 28, 2022 21:01
|
|
|
Shaggar posted:i mean i havent written java in years at this point cause .net has completely usurped it
|
|
#
?
Jul 28, 2022 21:20
|
|
|
Dependency conflicts in Java has been a problem at my current and previous job, it's definitely not theoretical. For all it's faults, Gradle warns you about version conflicts while maven picks one of indirect dependencies using some bad heuristic. You need to print the dependency tree and ctrl+f for the same artifact with different versions to discover the cause. Since some artifacts provide the same java classes (jakarta vs javax for a while e.g.) it can sometimes be a pain to debug as the JVM only tells you that some class was incorrect/missing. These problems usually occur with libraries like Guava that became "too commonly used to introduce breaking changes" because they provided utility that the standard library should have provided to begin with. Examples include collection classes, stream utilities, http clients, logging facilities, maybe more. Some libraries like Apache Commons, OkHttp and JUnit wised up and released major versions under new package names to avoid this problem.
|
|
#
?
Jul 28, 2022 21:50
|
|
|
simply use some crazy plugin with XML config to rewrite symbol names in your binaries, that's totally sane, no need to worry about name clashes!
|
|
#
?
Jul 28, 2022 22:17
|
|
|
lol if you rely on anyone else's code aka "dependencies"
|
|
#
?
Jul 28, 2022 22:25
|
|
|
Only trust your fists Code will never help you
|
|
#
?
Jul 28, 2022 22:30
|
|
|
CPColin posted:Only trust your fists
|
|
#
?
Jul 28, 2022 22:39
|
|
|
CPColin posted:Only trust your fists
|
|
#
?
Jul 28, 2022 23:26
|
|
|
nuget out
|
|
#
?
Jul 28, 2022 23:32
|
|
|
nuget what nugive
|
|
#
?
Jul 28, 2022 23:59
|
|
|
and my javax
|
|
#
?
Jul 29, 2022 01:51
|
|
|
Sapozhnik posted:it works until you need to use guava you don't. ever.
|
|
#
?
Jul 29, 2022 06:08
|
|
|
CRIP EATIN BREAD posted:maven rules. this is the truth. nuget might suck, idk, but I've never found out because I hardly need any deps and they rarely have transitive 3rd party ones (other than newtonsoft)
|
|
#
?
Jul 29, 2022 07:08
|
|
|
cool av posted:im the conceit that dependencies having their own differently versioned copy of their dependencies solves this problem it doesn’t completely but having a sealed ownership pattern for dependencies means that at least you don’t have to worry that your direct dependencies are quietly being replaced by indirects
|
|
#
?
Jul 29, 2022 12:59
|
|
|
Deffon posted:
Cowards
|
|
#
?
Jul 29, 2022 13:59
|
|
|
distortion park posted:this is the truth. nuget might suck, idk, but I've never found out because I hardly need any deps and they rarely have transitive 3rd party ones (other than newtonsoft) yeah nugets failings are largely stuff thats been removed like scripts that run on package install/uninstall (lmao what a terrible idea) or its poor integration into msbuild which is more of a problem with msbuild which sucks
|
|
#
?
Jul 29, 2022 14:46
|
|
|
my favourite thing to do with our internal nuget repo is name packages with stupid version suffixes that the package browser then won't display by default and confuse the poo poo out of everyone
|
|
#
?
Jul 29, 2022 16:03
|
|
|
meanwhile people continue not to give any shits about jdk9 modules (despite them looking fairly reasonable i guess?)
|
|
#
?
Jul 29, 2022 16:25
|
|
|
Sapozhnik posted:meanwhile people continue not to give any shits about jdk9 modules (despite them looking fairly reasonable i guess?) they’re pretty good, op. most popular libraries have a fixed module name now, so it’s feasible to use them, and an increasing number are fully modularised. we switched our code over to modules a couple of years ago and haven’t regretted it at all. it’s very nice having more control over api surface. the tooling is not great tho
|
|
#
?
Jul 29, 2022 18:00
|
|
|
are the git tools in jetbrains ides a pile of poo poo or am i just an idiot? i'm used to sublime merge and i've been trying to gently caress around with git in PyCharm but everything except for a simple commit ends up being a loving nightmare. every time i click "merge main into poo poo-branch". the commit that gets created shows poo poo-branch reverting all of the changes that would have been applied from main instead of applying them like i would expect
|
|
#
?
Jul 29, 2022 22:34
|
|
|
did car forum poster deploy something today
|
|
#
?
Jul 29, 2022 23:01
|
|
|
i vomit kittens posted:are the git tools in jetbrains ides a pile of poo poo or am i just an idiot? i'm used to sublime merge and i've been trying to gently caress around with git in PyCharm but everything except for a simple commit ends up being a loving nightmare. every time i click "merge main into poo poo-branch". the commit that gets created shows poo poo-branch reverting all of the changes that would have been applied from main instead of applying them like i would expect I use those tools only for making commits, diffs, rollbacks, merge conflicts, and looking at the history for which it does really well. Anything more complicated including merging other branches I pop a terminal
|
|
#
?
Jul 29, 2022 23:05
|
|
|
i vomit kittens posted:Everything except for a simple commit ends up being a loving nightmare.
|
|
#
?
Jul 29, 2022 23:07
|
|
|
|
| # ? May 27, 2024 08:39 |
|
|
the problem with git is it was made by moon man linux tovalds who is both insane and an rear end in a top hat
|
|
#
?
Jul 29, 2022 23:15
|
|