You are not logged in.
There's currently this rss feed which seems like the optimal solution ( https://www.archlinux.org/feeds/packages/added/x86_64 ) but it also includes packages that weren't added recently like python or syslong-ng for example. Shouldn't this ideally have only packages like toastify with the time of the entry being the initial commit ?
Offline
It should, but the current implementation is not (yet?) clever enough to tell the difference between a package that has been newly added to the general set of Arch packages, and a package which has been newly added to the testing repo but has existed in the repos in general for quite some time.
Hence you see lots of false positive hits for packages like python that have been rebuilt in testing.
Managing AUR repos The Right Way -- aurpublish (now a standalone tool)
Offline
I guess Arch Archive could be used to differentiate. At least to some degree: the archive doesn’t go infinitely long in the past too.
Offline
Couldn't you just filter out all the testing entries (assuming you only care about new packages in the main repos)?
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
Well the core and community repos have the same thing. The only one that seems to work properly is extra.
Offline
Couldn't you just filter out all the testing entries (assuming you only care about new packages in the main repos)?
This feed is the union of the feeds for all repositories. There are a number of feeds listed at https://www.archlinux.org/feeds/
What you would need to do is build a union of the core, extra, community and multilib feeds (rather than filtering out the testing, community-testing, multilib-testing, staging, community-staging, multilib-staging, gnome-unstable, and kde-unstable feeds).
Unfortunately there is still no convenient way to do this short of assembling the feed yourself. The defaults are not ordinarily very useful. But it should be doable with a bit of work, yes.
Last edited by eschwartz (2018-12-31 18:52:52)
Managing AUR repos The Right Way -- aurpublish (now a standalone tool)
Offline
What you would need to do is build a union of the core, extra, community and multilib feeds (rather than filtering out the testing, community-testing, multilib-testing, staging, community-staging, multilib-staging, gnome-unstable, and kde-unstable feeds).
Why? Filtering out anything with testing/unstable in the repo name would be much easier. Depending on how you view it, this could be as simple as a short 'grep -v ...` pipe.
Although, merging the results of the individual feeds would be pretty easy too.
Unless of course one must view this in a feed reader without any capacity to pipe through a filter. But if that's the case, the problem seems to be at the client end.
These few lines of python will do it:
#!/bin/env python3
import feedparser
feed = feedparser.parse('https://www.archlinux.org/feeds/packages/added/x86_64/')
hide = [ '/testing/', '/community-testing/', '/multilib-testing/', '/staging/', '/community-staging/', '/multilib-staging/', '/gnome-unstable/', '/kde-unstable/' ]
[ print(x.title) for x in feed.entries if not any(y in x.link for y in hide) ]Last edited by Trilby (2019-01-12 17:38:58)
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
I don't think the problem resides in the staging and testing repositories.
For example running Trilby's script right now outputs among results bpf which is not a new package.
Offline
It is a new package, though. it was just added as part of the linux-tools split PKGBUILD, 61 minutes ago.
Managing AUR repos The Right Way -- aurpublish (now a standalone tool)
Offline
Hello
Long enough I had found this link on this forum:
https://tokyo.cuoan.net/apw/output.json
Unfortunately, I couldn't find the post that was discussing it in the forum.
A+
Offline
@eschwartz I believe buttcake may after packages that are being packaged by arch for the first time.
Offline
@eschwartz
Ohh yeah I'm retarded. I saw all the commits and assumed it can't be new. Thanks.
@jean_no
This is nice, I will probably end up switching to this and jq since it's easier to manipulate than a rss feed ( for me ). Thanks.
@loqs
I honestly don't really care that much about that differentiation. The current two solutions provided in this thread are great.
Last edited by buttcake (2019-01-04 02:16:16)
Offline