Mike J Innes
Mike J Innes
The issue there is that it's still non-standard and specific to that one macro. If the syntax were actually added to the language that's a different issue of course. Maybe...
The macrotools `flatten` function might be useful here. That'd remove all the redundant begin blocks, and then the struct def will just be the first thing in the block.
There's no intentional reason why this couldn't be supported, but it may just be that we parse the unions before the types and then the types parser can't see inside...
Yes, this is just something we should document. MacroTools is more like a regex than a CSS selector; it doesn't have any concept of "better" matches and just takes the...
You can ask `prettify` not to alias like that via a kwarg. That's probably less necessary now that gensyms are printed in a sensible way.
You can do this by interpolating a match expression that you build by hand: ```julia @capture Expr(:$, :x) $(Expr(:$, :x_)) ``` Unfortunately this doesn't actually give you `x`. So to...
Unfortunately not: we have to scan the expression at compile time to figure out what should be bound to local variables, but if you interpolate like that then it can...
This is due to the fact that keyword arguments are stored as if they were the first argument to the function; when you ask for `xsym_`, you're really just asking...
Shell mode via `;` would be nice too, if that doesn't work already.
There are various reasons you might want to do this, for example if you want to make notebook-based tutorials for the package manager. I take the point about valid Julia...