Package Details: circleci-cli-bin 0.1.30549-1

Git Clone URL: https://aur.archlinux.org/circleci-cli-bin.git (read-only, click to copy)
Package Base: circleci-cli-bin
Description: CircleCI's new command-line application.
Upstream URL: https://github.com/CircleCI-Public/circleci-cli
Licenses: MIT
Conflicts: circleci-cli
Provides: circleci-cli
Submitter: stefanc_diff
Maintainer: igemnace
Last Packager: igemnace
Votes: 12
Popularity: 0.000000
First Submitted: 2018-12-13 08:05 (UTC)
Last Updated: 2024-03-15 09:01 (UTC)

Dependencies (0)

Required by (0)

Sources (2)

Pinned Comments

igemnace commented on 2021-07-06 01:00 (UTC) (edited on 2021-07-06 01:02 (UTC) by igemnace)

Indeed, here is an issue link for anyone else viewing this thread:

https://github.com/CircleCI-Public/circleci-cli/issues/589

I haven't run the CLI in a long time outside of routinary version checks, since I have no current use for the CLI and am maintaining the PKGBUILD out of habit.

If anyone discovers that this is due to some defect in the PKGBUILD, I'd be happy to merge a patch (or perhaps @stefanc_diff might allow for yet another co-maintainer).

From the looks of that issue, it seems like it might be requiring a kernel compiled with certain parameters (see https://github.com/CircleCI-Public/circleci-cli/issues/589#issuecomment-862544538). If so, hopefully we can express this simply as a dependency on some linux package. Otherwise, if this proves too brittle (e.g. no such linux package maintained both in the repos and in the AUR), we might have to make do with post-installation instructions, both in the PKGBUILD and in a pinned comment here in the AUR.

This comment suggests it must be resolved in circleci-cli itself, however (see https://github.com/CircleCI-Public/circleci-cli/issues/589#issuecomment-872387555). If so, hopefully this will be resolved upstream with a simple version bump.

Latest Comments

« First ‹ Previous 1 2 3

rosatolen commented on 2018-05-13 01:21 (UTC)

According to their github (https://github.com/circleci/local-cli/blob/master/circleci.sh), this local CLI is a wrapper around an internal tool called picard. When run, the wrapper will try to update the local version of picard by comparing a hash of the local vs what it pulls using docker.

Re: hash validation for pacman They offer no means of validating the executable, and creating our own system for updating a hash we maintain is very brittle. :( The best thing we have is that their documentation reflects the same HTTPS link that we use as a source in the PKGBUILD.

Re: License The license is MIT according to the github link above. It's included in the repository.

Looks like 7 hours ago they released for the first time something that we can effectively integrate with: https://github.com/circleci/local-cli/releases

I've created a corresponding PKGBUILD and pushed it.

alexpe87 commented on 2018-05-12 19:42 (UTC)

Hi! i added you as co-contributor - feel free to update.

A couple of things to mention (which you might know, but also might not):

a.) The URL is not stable at all. They frequently update, thus breaking any md5sum we can provide. Maybe there is a better one, but searching on GitHub/Bitbucket and similar platforms didn't help here either. b.) There is no licence specified on their webpage https://circleci.com/docs/2.0/local-cli/ - either we state "no licence", or cite the licence for the documentation (CC-BY-SA 4.0), which woudl also be incorrect.

The description was simply filled out to get this done as fast as possible.

Cheers

rosatolen commented on 2018-05-12 19:28 (UTC)

Unfortunately, there are several things wrong with the PKGBUILD here: the md5sum, the License, the description, and the dependency list. CircleCI's cli is not only for conda packages. Are you planning to update the package to fix these errors?

If not, I can update the AUR package.

« First ‹ Previous 1 2 3