Louis2d ago
daniel:// stenberg://
If you want automated machine-readable info about the most recent #curl release, fetch this URL: https://curl.se/info - always accurate and available (mastodon hides the "https://" scheme, but I know you get it)
Jun 11 at 8:48pmWeb
Show threadMathias Hasselmann2d ago
@bagder How about adding a checksum of the tarball to the output?
Show threaddaniel:// stenberg://2d ago
@taschenorakel coming up: https://github.com/curl/curl-www/pull/595
info: add a `sha256` field by bagder · Pull Request #595 · curl/curl-www

Since the 'info' file is an excellent way for the world to get automated information about the most recent curl release and tarball, providing a sha256sum for the file might help users to m...

GitHub
Show threadAnton Ohorodnyk2d ago
@bagder Does it allow to choose a rendering format?
At least JSON would be nice
Show threaddaniel:// stenberg://1d ago
@safigo https://github.com/curl/curl-www/pull/596
info.json: provide the /info in JSON as well as /info.json by bagder · Pull Request #596 · curl/curl-www

The curl.se website. Contribute to curl/curl-www development by creating an account on GitHub.

GitHub
Show threadMathias Hasselmann1d ago

@safigo The file is static content generated using the C++ preprocessor during upload.

Some kind of simplicity that has become rare in these days of web bloat and AI slop. I like this simplicity. Even more as the info file can be parsed with one-liners:

info = dict(line.strip().split(":", maxsplit=1) for line in open("info"))

Why would one like to throw JSON on that?

@bagder

Show threadMathias Hasselmann1d ago

@safigo A more readable version of the one-liner would be:

with open("info", mode="rt", encoding="utf-8") as file:
    lines = file.read().splitlines()
    info = dict(l.split(":", maxsplit=1) for l in lines)

@bagder

Show threadMathias Hasselmann1d ago

@safigo Even more such trivial file formats are not only accessible to high-level programming languages. They also work for shell scripts using only common stock commands:

echo $(sed -ne 's/^Download: //p' info)

@bagder

Show threadJP Mens1d ago

@bagder brilliant, and I wish more projects did that, but might I POST a suggestion? :-)

Edit: forgive me: I should have read the thread first: https://github.com/curl/curl-www/pull/596

Show threaddaniel:// stenberg://1d ago
@jpmens negotiating content I don't know, but I did https://github.com/curl/curl-www/pull/596
info.json: provide the /info in JSON as well as /info.json by bagder · Pull Request #596 · curl/curl-www

The curl.se website. Contribute to curl/curl-www development by creating an account on GitHub.

GitHub
Show threadStéphane Bortzmeyer1d ago
@jpmens @bagder Another suggestion because users love to create tasks for maintainers: work with other free software projects to have a way to get this information for several projects, without having to know their specific URL.
Show threaddaniel:// stenberg://1d ago
@bortzmeyer @jpmens that's left for someone else to do!