Does run_testsuite() dissemble?



I was testing a proposed fix to bug 1692651. See

  http://sourceforge.net/tracker/index.php?func=detail&aid=1692651&group_id=4933&atid=104933


After loading the fix, Maxima gave an error break for rtest15. I don't know
the
cause of the error break, but it might be the test

  quad_qags (sin(y[1]), y[1], 0, %pi, tol);

where tol = 1e-8. From the command line, Maxima (5.11.0) reports an
internal error
for this test:

(%i4) quad_qags (sin(y[1]), y[1], 0, %pi, 1e-8);
Compiling C:/DOCUME~1/BARTON~1/LOCALS~1/Temp/gazonk_1708_2.lsp.
; (DEFUN COMPILER::CMP-ANON ...) is being compiled.
;;; The special declaration (SPECIAL (($Y SIMP ARRAY) 1) ERRORSW) contains
a non-symbol (($Y

Yeechs! run_testsuite() doesn't report this as an error. What is going on?
Why
doesn't run_testsuite() report quad_qags (sin(y[1]), y[1], 0, %pi, 1e-8) as
an
error? Why does run_testsuite() report an error break on rtest15 if I first
load compar.lisp (the file with my putative fix to bug 1692651)?

Does run_testsuite() dissemble?

Barton