solve ignores conjugate

[rwst]

Results of solving with solve can be wrong
due to conjugate being ignored.

This is due to a Maxima bug.

sage: eq_a = conjugate(z) == I + 1
sage: eq_a
conjugate(z) == (I + 1)
sage: sol_a = solve(eq_a, z, solution_dict=True)
sage: sol_a
[{z: I + 1}]

This is obviously wrong.

sage: [eq_a.subs(s) for s in sol_a]
[(-I + 1) == (I + 1)]

Note that SymPy does not return a wrong answer,
but does not solve either:

sage: sol_aa = solve(eq, z, solution_dict=True, algorithm='sympy')
sage: sol_aa
ConditionSet(z, Eq(conjugate(z) - 1 - I, 0), Complexes)

Maybe solve should work around the Maxima conjugate bug.

Initial report and discussion:

• https://groups.google.com/forum/?hl=en#!topic/sage-support/wiwEbgILB14

Reported again at #30570.

Related:

• Search "conjugate" in Maxima bug tracker

CC: @slel @zimmermann6

Component: symbolics

Keywords: solve, conjugate, maxima, wrong result

Issue created by migration from https://trac.sagemath.org/ticket/18488

[slel]

Changed keywords from none to solve, conjugate, maxima, wrong result

[slel]

Description changed:

--- 
+++ 
@@ -1,7 +1,37 @@
-Reported in https://groups.google.com/forum/?hl=en#!topic/sage-support/wiwEbgILB14:
+Results of solving with `solve` can be wrong
+due to `conjugate` being ignored.
+
+This is due to a Maxima bug.
 
 ```
-z=var('z'); solve(conjugate(z)==1+i, z)
- [z == (I + 1)]
+sage: eq_a = conjugate(z) == I + 1
+sage: eq_a
+conjugate(z) == (I + 1)
+sage: sol_a = solve(eq, z, solution_dict=True)
+sage: sol_a
+[{z: I + 1}]
 ```
-Maybe solve should work around the Maxima conjugate bug.
+This is obviously wrong.
+
+```
+sage: [eq_a.subs(s) for s in sol_a]
+[(-I + 1) == (I + 1)]
+```
+
+Note that SymPy does not return a wrong answer,
+but does not solve either:
+
+```
+sage: sol_aa = solve(eq, z, solution_dict=True, algorithm='sympy')
+sage: sol_aa
+ConditionSet(z, Eq(conjugate(z) - 1 - I, 0), Complexes)
+```
+
+Maybe `solve` should work around the Maxima conjugate bug.
+
+Initial report and discussion:
+
+- https://groups.google.com/forum/?hl=en#!topic/sage-support/wiwEbgILB14:
+
+Reported again at #30570.
+

[slel]

comment:2

The "sage-support" discussion gives excerpts of posts
to the "maxima" mailing list.

A workaround discussed there consists in working with
real and imaginary parts and solving for those.

[slel]

Description changed:

--- 
+++ 
@@ -7,7 +7,7 @@
 sage: eq_a = conjugate(z) == I + 1
 sage: eq_a
 conjugate(z) == (I + 1)
-sage: sol_a = solve(eq, z, solution_dict=True)
+sage: sol_a = solve(eq_a, z, solution_dict=True)
 sage: sol_a
 [{z: I + 1}]
 ```
@@ -31,7 +31,11 @@
 
 Initial report and discussion:
 
-- https://groups.google.com/forum/?hl=en#!topic/sage-support/wiwEbgILB14:
+- https://groups.google.com/forum/?hl=en#!topic/sage-support/wiwEbgILB14
 
 Reported again at #30570.
 
+Related:
+
+- [Search "conjugate" in Maxima bug tracker](https://sourceforge.net/p/maxima/bugs/search/?q=conjugate)
+

[mkoeppe]

comment:4

Moving to 9.4, as 9.3 has been released.