Open Xerkus opened 2 months ago
I don't think it's entirely apparent that OpenSwoole isn't getting industry support; I actually see an equal number of mentions of both projects, and it's never clear if a mention of "Swoole" is specifically regarding the originating project, or just a generic term used for any Swoole installation, unless there's context pointing either way.
That said, I support this, as it would also be a first step towards creating a separate package specifically for OpenSwoole, which has a diverging API and feature-set.
For the signs of the industry support I was specifically looking at the issue activity, PR activity and commit activity in the openswoole repositories. For the extension last release was a minor release one year ago with the biggest highlight being the php 8.3 support. Next prior release was a major release exactly two years ago.
openswoole pecl page sees around 10k average monthly installs, openswoole/core composer package for the last 30 days had around 20k installs. Usage wise openswoole is not exactly abandoned. Contributions and maintenance of the extension itself however... are not promising.
Previously, we discussed splitting this package into two each declaring explicit dependency on either ext-swoole or ext-openswoole. A need for such split was dictated by the diverging API in those two extensions. See https://github.com/laminas/technical-steering-committee/blob/9e73fae14dd1bc8bdcdf875485353269af018de0/meetings/minutes/2023-05-08-TSC-Minutes.md
However, a year and a half passed since that TSC meeting. It is now apparent that openswoole does not have the kind of industry support and project activity required for such a fundamental extension.
At this point, I propose to drop the openswoole support entirely with no replacement. We will immediately remove it from documentation and from composer suggests.
At the same time we will begin a work on the new major release. Starting with a new major release this component will depend on a specific major versions of ext-swoole. This would allow us to utilize latest features in ext-swoole and to handle any BC breaking changes coming from the adoption of the current PHP features or from dependencies like #134