Closed sreichel closed 1 year ago
ps. would not remove downloader as of yet
I think the Venn diagram of power users and connect users would be quite small.
FWIW, I'm totally on board with this idea. I would love to see some real love applied to the M1 codebase.
A project I've started multiple times and would love to complete is fixing the broken modularization of M1. One aspect of that is controller locations. Since in the early days you could only have one module per shortname, they moved all the admin code into the Mage_Adminhtml module. After they added the ability to register multiple modules on a shortname (I like to think my module pushed them to do that, hah!), they started adding admin related controllers in their respective modules, but they never moved the original controllers to their correct modules. Moving them is fairly easy. And we can replace the moved controllers with stubs that call to the correct controller and log a deprecation warning. Another aspect of the modularization issue is underclared and circular deps in the modules. That's a tad harder to resolve, but is still possible. I know that M1 is an EOL platform at this point, but I think with some community effort it could live on in parallel for quite a while.
In any case, I think we should make a todo list of all the suggestions and get some community feedback
A project I've started multiple times and would love to complete is fixing the broken modularization of M1. One aspect of that is controller locations.
W'd like to see this ... but how to start? Guess waiting for a todo lit will take ages, so why not start a feature
or dev
branch and start commiting?
Intrigued by #435 I ran PhpStorm's code inspection on the repo. Oh my. :) I guess a lot of them are to be expected due to the custom aspects of Magento's core but def. plenty of code smells like broken function signature calls.
Was only interested in undefined vars like in the linked issue today, not too bad, but still a few that should be looked at. (untick "Variable might have not been defined" in inspection). Created a PR for an obvious one, the rest would need a closer look for fixing.
Yep, there are a lot of them ...
Im not sure, about fixing or just remove the whole method ... there are some that could not work and would produce a warning or fatal error (#419). This code is not used anywhere ... remove it.
@sreichel BTW, I've done the layout refactor, I just need to get a PR issued for it at some point.
i would be in with removing core modules!
@sreichel @Flyingmana Is this list still WIP? Could you please update tasks? Maybe with references to other issues?
@kkrieger85 Had no time during pasts weeks. I'll update ASAP.
Excellent job. It is a pity for so much effort if the work is not completed.
@ADDISON74 biggest part is done. DOCblocks for Adminhtml would be a large PR. So i dont worry,. The only PR i'd like to see get merged is #249 ... (still needs testing)
@sreichel - PR #249 is closed. What shall we do in this case? Re-open it?
I doubt there will be a M1.10 with bigger code changes, so why not to tidy up this project?
Cleanup:
Zend
as @colinmollenhour already suggested #374) *lib/Mage/?
lib/PEAR
(#952)lib/Zend/?
Mage_Compiler
(#534)Mage_Connect
(#952)Mage_GoogleBase
(deprecated after 1.5.1.0) (#1002)Mage_Usa
(now the process of creating the label is on DHL side)Mage_XmlConnect
(disabled by default, extendsEnterprise_
) (#1040)Phoenix_Moneybookers
(#1044)Mysql4
)"Improvements beyond bug fixes" (features):
"full" IDE support / DOC block updates
Connect(removed, see #952)GoogleBase(removed, see #1002)XmlConnect(removed, see #1040)development
I'd keep/update the master (
1.9.3.x
) branch as usual to not break anything, but start a new1.9.3.x-dev
that includes also BC breaking changes (like PHP7 required, or removing deprecated methods)