#4603 Control stack regression with abs_integrate / 5.48.0

[Michael Orlitzky]

https://sourceforge.net/p/maxima/bugs/3144/

This works in 5.47.0:

Maxima 5.47.0 https://maxima.sourceforge.io
using Lisp ECL 24.5.10
Distributed under the GNU Public License. See the file COPYING.
Dedicated to the memory of William Schelter.
The function bug_report() provides bug reporting information.
(%i1) integrate(sin(k*x)/x*erf(x^2),x,0,inf);
                            inf
                           /                  2
                           [    sin(k x) erf(x )
(%o1)                      I    ---------------- dx
                           ]           x
                           /
                            0
(%i2)  load(abs_integrate);  
(%o2) /usr/share/maxima/5.47.0/share/contrib/integration/abs_integrate.mac
(%i3) integrate(sin(k*x)/x*erf(x^2),x,0,inf);
                            inf
                           /                  2
                           [    sin(k x) erf(x )
(%o3)                      I    ---------------- dx
                           ]           x
                           /
                            0

But has regressed in 5.48.1:

Maxima 5.48.1 https://maxima.sourceforge.io
using Lisp ECL 24.5.10
Distributed under the GNU Public License. See the file COPYING.
Dedicated to the memory of William Schelter.
The function bug_report() provides bug reporting information.
(%i1) integrate(sin(k*x)/x*erf(x^2),x,0,inf);
                            inf
                           ⌠                  2
                           ⎮    sin(k x) erf(x )
(%o1)                      ⎮    ──────────────── dx
                           ⎮           x
                           ⌡
                            0
(%i2) load(abs_integrate);
(%o2) /usr/share/maxima/5.48.1/share/contrib/integration/abs_integrate.mac
(%i3) integrate(sin(k*x)/x*erf(x^2),x,0,inf);

Maxima encountered a Lisp error:

 BINDING-STACK overflow at size 10240. Stack can probably be resized.
Proceed with caution.

Automatically continuing.
To enable the Lisp debugger set *debugger-hook* to nil.

[Jaime E. Villate]

Well, I wouldn't say it worked in 5.47.0. It just gave up immmediately quoting the input. In 5.48.0 it now seems to be trying to do something but leading to an error.

[Robert Dodier]

Looks like this is triggered by a bug in limit. See bug report #4619.

[Barton Willis]

Fixed by Commit [0c11d9] master. The result is an integrate nounform, but it nolonger gives a stack overflow.

Append test to rtest_abs_integrate. Closing this ticket. Thanks to Robert who identified the reason for this bug.

[Michael Orlitzky]

Thank you both!