Added mas_remakeConstraints

[nickynick]

When working with Masonry, sometimes there is a need to update existing constraints for a view to do some kind of animation or whatever. mas_updateConstraints is only capable of updating constraint constants, so for more advanced changes we are forced to store created constraints in a separate array and then call uninstall on each prior to creating new ones.

This is quite tedious, so I had an idea for a new method, mas_remakeConstraints. It is very simple: it removes all constraints previously defined and installed for the view, allowing you to provide replacements without hassle.

Under the hood, it stores a MASViewConstraint in a private associated set of its firstViewAttribute.view, so it's quite simple implementation-wise too.

TODO list:

• tests :)

[cloudkite]

This would be very handy no doubt! however there are a few pitfalls with this.

[superview.subview remakeConstraints:^(MASConstraintMaker *make){
   make.top.equalTo(100); //this constraint actually gets added to superview
}];

// superview constraints will get removed, meaning the above constraint gets deleted
[superview remakeConstraints:^(MASConstraintMaker *make){
   make.width.equalTo(10);
}];

The other reservation I have is that this adds a strong reference between the view and constraint. Where as before this relationship is managed by UIKit. This means constraints which would normally be deallocated when UIKit automatically removes constraints ie in response to removeFromSuperview will now not be deallocated.

[cloudkite]

Whoops ignore the last comment assumed you where holding onto NSLayoutConstraints not MASViewConstraints. Nice work looks good 👍

[cloudkite]

I guess one thing to think about, is there is now a little bit of extra memory overhead. As MASViewConstraint and MASCompositeConstraint instances will now hang around, previously they would have been deallocated by ARC after a makeConstraints, updateConstraints etc block has been completed.

However they are quite light-weight so don't think it would be a big deal?

[nickynick]

Well yeah, each constraint is merely a hundred bytes, so I don't think it is a problem. There should be literally thousands of constraints for them to take up any considerable amount of memory, and if you have a thousand of constraints, then you are likely screwed in other ways already :)

[cloudkite]

Yeah sounds good :)

[nickynick]

Added some tests, guess it should be fine :)

[cloudkite]

Nice 👍 thanks for the contribution :)