Open rtoy opened 2 months ago
Imported from SourceForge on 2024-07-03 03:09:05 Created by billingd on 2023-09-28 04:19:34 Original: https://sourceforge.net/p/maxima/bugs/4186/#d9f5
contrib_ode testsuite issue addressed in commit [80b8f1]
Imported from SourceForge on 2024-07-03 03:09:08 Created by robert_dodier on 2023-11-03 19:50:15 Original: https://sourceforge.net/p/maxima/bugs/4186/#ef63
Imported from SourceForge on 2024-07-03 03:09:12 Created by robert_dodier on 2023-11-06 04:56:20 Original: https://sourceforge.net/p/maxima/bugs/4186/#941f
Imported from SourceForge on 2024-07-03 03:09:15 Created by robert_dodier on 2023-11-06 04:56:20 Original: https://sourceforge.net/p/maxima/bugs/4186/#d157
For the record, git bisect shows that commit [ 44d91a7 ] is the place where the behavior changed. The commit log says:
commit 44d91a796e3868de4524e173343e1e1844e95155
Author: Marius Gerbershagen <marius.gerbershagen@physik.uni-wuerzburg.de>
Date: Fri Nov 5 14:32:07 2021 +0100
Fix sign function with rational powers of complex arguments
In complex mode we cannot assume for instance that sqrt(-x)=0 for x <=
0. Anything non-positive or non-zero to the power of 1/(2*n) for
integer n has sign '$complex.
Example of a bug which is fixed by this commit:
(%i1) abs(sqrt(sin(theta) - 1));
(%o1) 0
I didn't investigate whether the changed behavior is a bug or not; probably not, I'm guessing, but it should be reviewed.
Imported from SourceForge on 2024-07-03 03:09:04 Created by billingd on 2023-09-28 03:19:23 Original: https://sourceforge.net/p/maxima/bugs/4186
There is a regression/change in askequal since version 5.47.0.
used to return "no" but now asks "Is sqrt(9 - a) - 1 equal to - 1?"
Now test kamke 2.405 in share/contrib/diffequations/tests/rtestode_kamke_2_5.mac fails. I can work around the issue by assuming a>9.
I will arrange a witch-hunt with git bisect to find the patch responsible, which may be an improvement and not a bug.
vs