Could someone transfer the following issues form the racket repository to the plot repository? Thanks, Alex.
I want to use errortrace when invoking a submodule, but when I do: racket -l errortrace -e '(require (submod "myprog.rkt" test))' it fails with: #%top-interaction: unbound identifier;
also, no #%app syntax transformer is bound
at: #%top-interaction
in: (#%top-interaction require (submod "myprog.rkt" test))
errortrace...: Same error if I start with the repl with racket -li errortrace then > ,en (submod "myprog.rkt" test) This is v8.4.0.3 [cs].
raco test is able to invoke submodules, but doesn’t seem to be able to instrument errortrace, AFAICT.
What am I doing wrong?
Done
Is there an easy way in the FFI to make use of a va_list past to a (Racket) callback?
what function(s) should I use to inspect a syntax object’s phase level?
syntax-debug-info?
Looks like it doesn’t return the phase-level
oh, it consumes phase-level
Is it meaningful to return a phase-level?
There could be multiple phase-levels associated with a syntax object, right?
Yes, you are right. multiple phase levels are fine.
I don’t know what happened to ,en, but for command-line options I use
racket -l racket/init -l errortrace -e '(require (submod "prog.rkt" test))' When there is a -l flag, Racket no longer loads racket/init by default. However, racket/init is a module that exports #%top-interaction, thereby enabling top-level expressions (and therefore -e).
(The complete description: * If -t/-l/-p/-u appears before the first -i/-e/-f/-r,
-n is added )
Given this help text I suspect racket -il errortrace is different from racket -li errortrace