this is a bit premature where I’m at, but can someone point me towards any code that racket provides to do free-variable analysis at macro expansion or run time? I’m wanting to expand a system I have to allow for free variables to be treated symbolically (a la Mathmatica) instead of erroring out as undefined.

I’m still working through BR so this is a bit ahead of myself. Concrete examples (that are understandable) would be nice

@zenspider if you only wish to change the behavior of undefined variables, might I recommend looking at the #%top macro? it’s inserted as a wrapper around all unbound identifiers and you can override it to change what happens when an unbound identifier occurs.

zenspider: (require syntax/free-vars)

I’ve run out of immediately available test failures for the rtkio refactoring, so I’m going to push it to the main repo’s master branch. I’m still interested in build reports, though.

@mflatt here’s a build failure: https://travis-ci.org/racket/racket/jobs/244521015

mflatt: I like the log message :)

IRC-slack bridge is malfunctioning again

@stamourv It looks like the contract-profile commits on June 1 caused the documentation build for contract-profile.scrbl to take substantially more memory. At over 1GB, it’s now the two peak memory-use points during a build (at “running” and “rendering”) as shown on the graph http://build-plot.racket-lang.org/ . Do you know offhand what might have triggered the memory use and whether it would be easily avoidable?

@mflatt: Yes, I increased the size of the examples in the docs. The original input sizes were too small to show anything.

But I should find a better solution; that kind of memory use is not good.

@notjack that might actually be perfect. thank you.

@ben I’d like to be a contributor to racket/rackunit in order to label / organize issues and not deal with keeping my own fork in sync, who hath the authority to bestow this power upon me?

@samth can you give jack push/issue-labeling access to racket/rackunit?