#Go 1.24 interactive tour
1 messages ยท Page 1 of 1 (latest)
omg - we can retire identity.NewID()???? https://pkg.go.dev/crypto/rand@master#Text
cc @narrow pine @unreal pulsar ๐
"Main module's version" also relevant to us - now you get v0.0.0-20250111143208-a7857c757b85+dirty instead of (devel)
t.Context() is also almost relevant to us but only does a very narrow part of testctx (cc @karmic hare) - it's just a context that gets canceled on test completion. I guess we can switch to it as the base context
gotta hope rand.Text() followed the spirit of that RFC4648 and not the letter, bc the letter has you use = for padding and it can break urls
actually, don't have to hope, i can read and it does indeed follow the spirit and not the letter
os.OpenRoot is also a very nice feature... makes it significantly less spooking touching the filesystem on a web-exposed server... although like should you really be doing that in the first place lol
omfg they added fake clocks so you don't have to dep inject fake time any longer
lol yea was just gonna say
the way it's integrated with goroutines blocking is really slick
so you don't even need to babysit time advancing
really a lot of good stuff, VCS info in BuildInfo.Main.Version is also mega convenient... i think i've seen like 5 different ways ppl do that same thing themselves to get shas into their logs
but in terms of time saved i think the fake clock addition wins the release
soooo when can we upgrade? I'd like to see how the t.Context() fits into testctx while testctx is still loaded in my brain
should be out next month, we usually upgrade pretty much immediately
https://antonz.org/go-1-24/#tool-dependencies
^ tool dependencies seem kinda interesting ๐ค
kinda interesting to imagine integrating that into common dagger commands
since i could definitely see language servers picking up these tools as code actions
And @small nest, look at what the author built for the running examples: https://codapi.org