Add support for `extension.json`

SemanticMediaWiki / SemanticScribunto

Provides service functions to support the Scribunto extension

Other

24 stars 14 forks source link

Add support for `extension.json` #61

Closed mwjames closed 6 years ago

mwjames commented 6 years ago

This PR is made in reference to: #

This PR addresses or contains:

Adds support for extension.json and hereby the requirement to use wfLoadExtension( 'SemanticScribunto' ); in the LocalSettings.php
PHP 5.6 as min requirement
Switches to 2.0.0-alpha as version

This PR includes:

[ ] Tests (unit/integration)
[x] CI build passed

mwjames commented 6 years ago

@oetterer @kghbln I leave you both in charge to push/merge for a release (adapt branch-alias in composer.json, modify the CI matrix etc.).

kghbln commented 6 years ago

@oetterer It will be great if you could tackle this. I would like to do the release on my release sprint day on October 11.

oetterer commented 6 years ago

Will look into it on thursday.

kghbln commented 6 years ago

Will look into it on thursday.

Cool. Thanks a lot!

oetterer commented 6 years ago

Hey. Just to confirm: this release just throws the extension.json into the mix. It does not remove the necessity to use composer, right?

Quick question then: why can't we put

wfLoadExtension( "SemanticScribunto" );

into \SemanticScribunto::load()? The method is called through composer before the extension registration process and therefore queues SSC correctly. We would leviate the necessity for admis to load SSC by themselves.

And btw: the repo claims that I was the author of https://github.com/SemanticMediaWiki/SemanticScribunto/pull/61/commits/4e4e8bf51521a4c45477766286fdc5de2f9daa33. Either I was stoned to hell or something is amiss here...

kghbln commented 6 years ago

Hey. Just to confirm: this release just throws the extension.json into the mix. It does not remove the necessity to use composer, right?

Affirmative.

Quick question then: why can't we put ...

This would be fine with me however I think this mechanism was also created to allow administration via some special page etc. at a later stage. So deviating from the standard might not be the best idea. Also other semantic extensions using this mechanism do not do this either.

We would leviate the necessity for admis to load SSC by themselves.

That's true. Since a major version is published we should get the admin's attention to do so.

And btw: the repo claims that I was the author of 4e4e8bf. Either I was stoned to hell or something is amiss here...

Perhaps mwjames used some astray branch or pull created by you still there to do his changes. I do not know either.