enable abs_integrate package from maxima

[kcrisman]

This is the followup to #8624, as suggested in #10434.

We want to do more integrals! There are lots at the patch at #8624 which it would be great to add.

Depends on #12094

CC: @burcin @jasongrout

Component: calculus

Author: Michael Orlitzky, Jason Grout

Reviewer: Karl-Dieter Crisman

Merged: sage-5.0.beta3

Issue created by migration from https://trac.sagemath.org/ticket/11483

[kcrisman]

comment:1

This would fix #11594.

[kcrisman]

comment:2

See this sage-support thread for another example.

[kcrisman]

comment:3

See this sage-devel discussion for someone working on this!

Also, see #11966 for a newer spkg to base upon.

[orlitzky]

comment:4

Here's my shot at this. There's one regression (a Maxima bug) that's fixed by Maxima 5.24.0, thus the dependency.

[orlitzky]

Dependencies: #12094

[kcrisman]

Reviewer: Karl-Dieter Crisman

[kcrisman]

Author: Michael Orlitzky

[kcrisman]

comment:6

Ok, this looks good, and is the proper way to deal with this. All tests pass.

Should we add Jason as an author due to his work at #8624? Up to him, probably.

[kcrisman]

comment:7

Scratch that - Michael, can you update and give a proper commit message etc.? Also, the documentation is incorrectly formatted with double colons etc. - including the zXXX comment, which should have the double colon for EXAMPLES made single and then have a double colon after "subject to change".

[kcrisman]

Work Issues: commit message, formatting

[orlitzky]

Updated patch addressing reviewer comments.

[orlitzky]

comment:8

Attachment: sage-trac-11483.patch.gz

That should be a little bit better.

Jason should get author credit too. I unfortunately didn't realize there was a patch at #8624, since it's closed: fixed, until after I had created my patch. Once I saw his, I went back and added some of his doctests to mine.

If he doesn't get author credit, then it looks like I just stole everything =)

[kcrisman]

Changed author from Michael Orlitzky to Michael Orlitzky, Jason Grout

[kcrisman]

comment:9

In retrospect, I realize that this file (maxima_lib) isn't even in the reference manual, and it might be nice to have some of these examples in more places (note that x.integrate? is horrible, though it points to a better place). Anyway, this is all fine now.

[jdemeyer]

Changed work issues from commit message, formatting to none

[jdemeyer]

Merged: sage-5.0.beta3

[dimpase]

comment:12

abs_integrate can't do easy things like integrate(abs(cos(x)),x) right.

(%i8) integrate(abs(cos(x)),x);

log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
log: encountered log(0).
(%o8) 2*((1/2-sin(x)/((cos(x)+1)*(sin(x)^2/(cos(x)+1)^2+1)))
        *signum(sin(x)/(cos(x)+1)-1)
        +1)*signum(sin(x)/(cos(x)+1)+1)

and the resulting expression is wrong, as it is 1 at 0, and not 0. As a result, we get integrate(abs(cos(x)),x,0,pi)==-1.
Needless to say, same happens in Sage:

sage: integrate(abs(cos(x)),(x,0,pi))
-1

This is with maxima 5.26, as well as with 5.29, see #13364.

[rwst]

comment:13

Please don't add to this closed ticket. The current ticket about abs_integrate is #12731.