Apparently the reason I can't just tag a "class" attribute onto a closxp and call that an iteror is.... ESS is fucking with me. If I evaluate a chunk in ESS, it is altering the environment of the resulting closxp. it comes back with the enclosing environment set to the package namespace, instead of the environment that really enclosed it and has its variables. It's weird.
Aargh, something is happening in ess-mode that makes emacs lock up waiting for a response when it's in the debugger. Can be triggered by ess-whatever-completions mode. Is triggered by just having point in an R buffer for a few seconds.
ESS output window keeps saying "Disabling output delimiter because CMD failed to parse."
This is likely because all my variable names in the compiled generators are fucky.
(i.e. ended up with the names and translated functions.) One way to do that might be to not bother slapping context on the names of nodes in the special "first" context encountered.)
make a custom node style for handlers, or maybe an out-and-back edge? An edge to a node in the middle of a trampoline edge might work too. Hey maybe rewrite graphAsync() in collector style?
I fixed a bug in break_cps where it wasn't forcing "cont" and this caused semicolons to be screwy. Seems like I could statically check that context constructors are forcing all the continuation args they use. Extend the static check for handler-passing to check for argument forcing...
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.