Introducing the pkg.go.dev API
25 points by jmillikin
25 points by jmillikin
Developers building tools, IDE integrations, and automated workflows have historically relied on fragile workarounds like web scraping to access this data.
Why don't these tools go get the package and run toolchain tools like go doc on them?
You would have to download every public package and do a bunch of compute on them to be able to identify what packages depend on a particular package. Also, having to download the source code in the first place to answer metadata questions is a waste.
While this sounds like a straightforward option, one argument I can imagine for scraping pkg.go.dev is that this avoids downloading the package and cluttering the GOPATH with it.
Because they are not written in go and don't want to be tied to a specific implementation of a protocol.
this is really awesome, helps a lot with tooling which will has become more critical lately
Really cool that they expose symbols and imported-by. Hopefully replacing what a lot of tools and data aggregators do in downloading all the binaries and analysing them.
Is it still not possible to download a go package in the standard UNIX tarball form: package-name-version.tar.gz ?
I'm excited about this because I have never been able to get the "/" shortcut (for searching) to work on pkg.go.dev in Firefox. Am I the only one? It works in Chrome, which makes me want to get out my tin foil hat.
Maybe this means someone will create a TUI or desktop app for browsing packages that I can use instead.
The behavior you're experiencing is happening "locally," in your copy of Firefox, and is not pkg.go.dev's fault. You will experience it on the front page of https://lobste.rs/ which I doubt very seriously is in the pocket of Big Chromium
about:keyboard does not expose the / feature to customization so I'd presume there are 155 different bugzillas that are 20 years old about it