Closed zmaril closed 9 years ago
Oh interesting -- the problem isn't that gensyms are inherently bad to capture this way, but since the test.chuck code calls clojure.core/destructure itself it just ends up with different gensyms than the ones that eventually get used in the course of normal destructuring.
I'm going to ponder whether there's an easy enough clean way to do this without guessing based on symbol names; thanks for the PR!
No problem! I'm using this fix locally for now but I have no idea if it will come back and break something else later on. All the other tests run just fine, but I don't know enough about how for
works to say if there is a cleaner way to do this.
Okay I think I figured out a different approach. I'll merge this to get the test, add my changes on top, then make a release.
Thanks again.
Version "0.1.15"
is out
This seems to work for me! Thanks for the quick release.
When trying to generate code to destructure the results of generators with for-all, the macro was getting caught up by gen-syms. By filtering out symbols with the prefixes "map" and "vec", the macro no longer tries to generate code that references the generated symbols and works as expected (or at least how I expected it to work). A simple test was added as well.