Mumfrey
Mumfrey
> So, it wouldn't make sense to for the line numbers of the mixin themselves to match in the injected source file, because they need to point to the injecting...
It's the unique-ification of the method names that is done in order to prevent them accidentally overlapping or being targetted by transformers during post-application.
I'll see what I can add to support this.
Yeah I've observed that `remap` is one of the least understood arguments, though it is currently documented in [the javadoc](http://jenkins.liteloader.com/job/Mixin/javadoc/index.html?org/spongepowered/asm/mixin/injection/At.html) (as are most things) though I will make a note...
Following on from our discussion the other night, I think the way to approach this looks like the following: 1. From the point of view of injection points, target selectors,...
> I like the first 2 syntax ideas (`->` and `::`). Just a question on that, is it going to work with nested lambdas? Is a "recurse into all lambdas"...
Any further thoughts on the above @gabizou, @Barteks2x?
Read the above. Short answer, "no". Because otherwise this represents a semantic change to *existing* selectors, which I don't want to do.
The main reason injection points currently only receive the instruction list is for defensive purposes, the instruction list the injection points receive as part of their operation is read-only (and...
Yep. It's clear from some of the questions we get in IRC that it's not as obvious as I thought it was. I'll do a basic explanation now but also...