Should the Config and GlobalConfig classes and the associated RequestContext methods be reverted from 1.23 as an incomplete feature? As far as I can tell, it is not yet used anywhere, so reverting it should be easy.
getConfig() was added to IContextSource in 101a2a160b05[1]. Then the method was changed to return a new class of object (Config) instead of a SiteConfiguration object in fbfe789b987b[2]; however, the Config class faces significant changes in I5a5857fc[3].
[1]: https://gerrit.wikimedia.org/r/#/c/92004/ [2]: https://gerrit.wikimedia.org/r/#/c/109266/ [3]: https://gerrit.wikimedia.org/r/#/c/109850/
I'd suggest a revert from the branch, yes.
-- View this message in context: http://wikimedia.7.x6.nabble.com/Config-class-and-1-23-tp5026223p5026236.htm... Sent from the Wikipedia Developers mailing list archive at Nabble.com.
I agree. I was going to attempt to fix the newest patch, but until the semester ends I won't have a lot of time (and it seems neither does the current patch owner).
*-- * *Tyler Romeo* Stevens Institute of Technology, Class of 2016 Major in Computer Science
On Fri, Apr 18, 2014 at 12:16 PM, Aaron Schulz aschulz4587@gmail.comwrote:
I'd suggest a revert from the branch, yes.
-- View this message in context: http://wikimedia.7.x6.nabble.com/Config-class-and-1-23-tp5026223p5026236.htm... Sent from the Wikipedia Developers mailing list archive at Nabble.com.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Hey,
Is there some kind of description of the responsibility of the context source stuff anywhere? And the design vision behind it? I find the whole thing extremely dubious, as it appears to try make you bind to a whole group of rather scary classes. Perhaps I am missing something?
Cheers
-- Jeroen De Dauw - http://www.bn2vs.com Software craftsmanship advocate Evil software architect at Wikimedia Germany ~=[,,_,,]:3
On 04/18/2014 05:35 PM, Jeroen De Dauw wrote:
Hey,
Is there some kind of description of the responsibility of the context source stuff anywhere? And the design vision behind it? I find the whole thing extremely dubious, as it appears to try make you bind to a whole group of rather scary classes. Perhaps I am missing something?
Cheers
-- Jeroen De Dauw - http://www.bn2vs.com Software craftsmanship advocate Evil software architect at Wikimedia Germany ~=[,,_,,]:3 _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Jeroen, thanks for your message.
Looks like we are still working on merging the "Make abstract Config class truly implementation-agnostic" changeset.[0]
We came out of the Architecture Summit with a lot of momentum[1] for replacing our direct use of global variables throughout the system. We agreed on "Basic hygiene of taking what's already being used (JSON configuration like EventLogging, Zero, and UploadWizard campaigns) and separating into separate extension or integrating into Core." But we don't really have a clear design document/discussion of the current path forward, as far as I know.
We've also talked about the "Graphical configuration interface" RfC (formerly "Configuration database 2") a little bit[2][3] but could use a lot more speccing out of requirements.
[0] https://gerrit.wikimedia.org/r/#/c/109850/ [1] https://www.mediawiki.org/wiki/Talk:Architecture_Summit_2014/Configuration [2] https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2013-11-20 [3] https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-03-12
On 4/29/14, 1:56 PM, Sumana Harihareswara wrote:
Looks like we are still working on merging the "Make abstract Config class truly implementation-agnostic" changeset.[0]
The patch currently has three +1's (including myself) and one -1; it's just waiting for someone to +2 it :)
-- Legoktm
On 5/23/14, 9:26 PM, Legoktm wrote:
On 4/29/14, 1:56 PM, Sumana Harihareswara wrote:
Looks like we are still working on merging the "Make abstract Config class truly implementation-agnostic" changeset.[0]
The patch currently has three +1's (including myself) and one -1; it's just waiting for someone to +2 it :)
Super-delayed update: Tyler merged the patch, and it was backported into 1.23. Thanks to everyone who contributed during the 4 month journey of the patchset!
Now, we begin the fun part of migrating our code to use the new classes. Reedy has [1] open which switches all of core's API to use it, and I've submitted [2] for MassMessage.
I've also written up some basic documentation[3] about how to migrate to the new classes.
[1] https://gerrit.wikimedia.org/r/#/c/109271/ [2] https://gerrit.wikimedia.org/r/#/c/137216/ [3] https://www.mediawiki.org/wiki/Manual:Configuration_for_developers
-- Legoktm
On 04/18/2014 04:40 AM, Kevin Israel wrote:
Should the Config and GlobalConfig classes and the associated RequestContext methods be reverted from 1.23 as an incomplete feature? As far as I can tell, it is not yet used anywhere, so reverting it should be easy.
The implementation in core right now is incomplete. Once [3] is merged though, it should be suitable for people to start using it, and I know Reedy had a patch that converted all of the API modules in core to use it.
I would like to have Config make it into 1.23, mainly since it's an LTS, which would allow more extensions to take advantage of it without breaking backwards-compatability.
I was actually working on [3] last night and just updated the patchset, which should address the concerns the reviewers had. But if we're not able to get it merged in time for the 1.23 release, then I would recommend reverting the current implementation out of the release branch.
-- Legoktm
On Fri, Apr 18, 2014 at 7:17 PM, Legoktm legoktm.wikipedia@gmail.comwrote:
I would like to have Config make it into 1.23, mainly since it's an LTS, which would allow more extensions to take advantage of it without breaking backwards-compatability.
I don't mind getting the Config class into 1.23. However, at this moment, Legoktm's patch is not merged, and while it'd be nice to get it merged before the release, as we all know the MediaWiki review process is not always as fast as we'd like it to be.
I'd recommend reverting the merged Config patch, and then backporting Legoktm's patch when it's finished and merged.
*-- * *Tyler Romeo* Stevens Institute of Technology, Class of 2016 Major in Computer Science
On Fri, Apr 18, 2014 at 5:35 PM, Tyler Romeo tylerromeo@gmail.com wrote:
I'd recommend reverting the merged Config patch, and then backporting Legoktm's patch when it's finished and merged.
That'd be messy in the history. Let's just wait a few days or so and see what comes of the patch. There's no rush...the revert will literally take seconds if we do need to back it out of REL1_23.
-Chad
wikitech-l@lists.wikimedia.org