mikmoore

I'll add to the sentiment that `concat` or `concatenate` might be a more natural name for this operation. But it seems that this has already been in the works for...

In this particular example you can get away with this, but in general you cannot. For example, `map(cos, Diagonal([1, 2]))` does not preserve diagonal structure. It would be difficult to...

Neat! I didn't notice that it already had that infrastructure in place. So what appears to be the issue here is that it isn't recognizing that `f(zero(T)) == zero(f(_::T))` (or...

`iszero` cannot catch every case. Consider `f(x) = fill(x,1,1,1)`. If we only check that `iszero(f(zero(T)))` then we will determine that the `Diagonal`-result map is legal. But It isn't. ```julia julia>...

Would changing this error text to something like `unexpected "end" or incomplete expression` be sufficiently helpful? Obviously, this has changed nothing but it may help to direct the person to...

We can't special-case on `end` alone. There are a number of situations where this is relevant: ```julia [x + ] (x + ,) x + ; y ``` Related, but...

I agree that a `SingularException` is often useful - and that sidestepping it could cause plenty of chaos. However, I don't think it's terribly consistent that this works for all...

It's slightly more verbose, but `Base.Fix2(getproperty,sym)` is equivalent to your proposed `getproperty(sym)` definition and has some compiler advantages - plus you can use it right now. The `Base.Fix1` and `Base.Fix2`...

Before we canonize `reduce(vcat,itr)` and friends as the proper way to do these things, should we consider whether this is actually something we want to do? I've always despised these...

Technically, removing these performance shortcuts would be non-breaking since the fallback is still operable. That said, reversion to the fallback would impact performance very negatively and for that reason I...