[RFC|Rejected] Revert all of ajax in 3.1

[tbackoff]

What I believe your referring to (quick mod, if not disregard this) is not still broken

I was referring to "Mark forums/topics read". Has this been fixed? Also, my previous question/statement still stands:

if we are going to ajax, it needs to be throughout the entire core, not bits and parts. For example, why are we using trigger_error in some parts, but ajax in others?
[snip]
Choose one, the other, or none at all.

[MichaelC]

What I believe your referring to (quick mod, if not disregard this) is not still broken

I was referring to "Mark forums/topics read". Has this been fixed? Also, my previous question/statement still stands:

if we are going to ajax, it needs to be throughout the entire core, not bits and parts. For example, why are we using trigger_error in some parts, but ajax in others?
[snip]
Choose one, the other, or none at all.

It works, but it doesn't update the icons, but yes, that still needs fixing (I think)

The whole point is AJAX'ing it bit by bit, not all at the same time. That's what caused all the problems in the first place.

[imkingdavid]

if we are going to ajax, it needs to be throughout the entire core, not bits and parts. For example, why are we using trigger_error in some parts, but ajax in others?
[snip]
Choose one, the other, or none at all.

...because sometimes you need a trigger_error() and other places make more sense to display a modal box or a flash message or something. We aren't trying to turn this into ajaxBB. We are just trying to implement AJAX technology into places in which it makes sense.

It is better to get the basic AJAX framework merged and then work on features bit by bit, testing and reviewing each one separately, than to try and fit everything into one huge, unreadable PR that will inevitably be merged without proper review.

Think of it like making a new software. If you try to implement everything at once, you get very sketchy, disorganized and orphaned (i.e. unused but retained because you forgot to remove it) code. If you focus your efforts on one feature at a time and give each one the attention it deserves, you get a much cleaner and more maintainable code base.

[Danielx64]

If you want to revert this because it lacks the proper documentation of what it does, then you might as well just scrap the majority of the phpBB codebase, which also lacks proper documentation, and rewrite it from scratch (that's where 4.0 comes in).

^ This

[naderman]

That's really not a reason to add any new code lacking documentation.

[Erik Frèrejean]

That's really not a reason to add any new code lacking documentation.

Though neither a reason to remove already added code, especially as it is hard to expect that anyone will jump in to redo this based upon the way this iteration was handled.

[imkingdavid]

That's really not a reason to add any new code lacking documentation.

No one ever suggested, as far as I can tell, adding any further code without documentation. All that has been suggested so far (by me, at least) is to keep what we have and work on documenting it, rather than scrapping it and starting over. A bunch have time has already been spent on it, it would be a waste to toss that out just because the code is currently undocumented. Instead, let's just document what we have.

[igorw]

I was mostly in charge of finishing up the pull request and the docs in the wiki.

There have been quite a lot of different issues discussed here. But I feel like some things are missing from the discussion. Before I comment on what I think we should or should not do, I'd like to hear some answers to these questions:

• Where exactly (apart from bugs) do you see shortcomings in the current code, and how would you improve it?
• Where exactly do you see shortcomings in the current documentation, and how would you improve it?

[Erik Frèrejean]

A better solution rather then just nuking this would be to fix up the issues for 3.1, and create a new RFC which is aimed to improve/rewrite parts of the current codebase for 3.2. Software can evolve over time and I believe that this is a case that can evolve quite a bit.

[Erik Frèrejean]

not to mention it's still broken.

Are there tickets for those "still" broken? If there are then those issues will be resolved, if not it is hard to expect them to be solved at any point