This site is a static rendering of the Trac instance that was used by R7RS-WG1 for its work on R7RS-small (PDF), which was ratified in 2013. For more information, see Home. For a version of this page that may be more recent, see ArgumentOrder in WG2's repo for R7RS-large.


2015-06-01 04:44:11

This is a test of the order in which Scheme REPLs evaluate the arguments of a procedure call. It doesn't necessarily apply when arguments have different levels of complexity, nor does it apply to compiled code. The test is ((lambda x x) (display "l") (display "t") (display "r")). In some cases, side-effecting procedures other than display were used.

Displays ltr: Racket, Gauche, Gambit, Chicken, Bigloo, Scheme48/scsh, Kawa, SCM, Vicare, Ypsilon, Mosh, IronScheme, NexJ, STklos, KSi, SigScheme, Scheme 9, RScheme, S7, Rep, Schemik, Elk, UMB, Llava, Sizzle, FemtoLisp, Dfsch, Inlab, Foment, Sagittarius

Displays rtl: MIT, BDC, XLisp, SXM, Chibi

Unpredictable order: Chez, Larceny

Undeterminable order (no side effects): Owl Lisp

The following C code was also tested:

# include <stdio.h> int yotz(int a, int b, int c) { return a + b + c; } int main() { yotz(printf("l"), printf("t"), printf("r")); }

Displays ltr: Clang

Displays rtl: GCC