inconsistent behaviour of solve

[fchapoton]

I have found the following :

sage: w,z=var('w','z')
sage: solve([z-4,w-3],[z,w])
[[z == 4, w == 3]]
sage: solve([z-4],z)                                       
[z == 4]
sage: solve([z-4,z-2],z)
[]
sage: solve([z-4,z-2],[z])
[]
sage: solve([z-4],[z])
TypeError: [z] is not a valid variable.

The last line is a really bad behaviour ! It forces to distinguish the case when there is only one equation and one variable. Compare also with the first solve.

Component: symbolics

Keywords: solve, symbolic

Author: Punarbasu Purkayastha

Reviewer: Frédéric Chapoton

Merged: sage-5.3.beta2

Issue created by migration from https://trac.sagemath.org/ticket/13286

[kcrisman]

comment:1

At first I thought this was part of #10750, but apparently not. Thanks for this report.

On the plus side, the behavior isn't that bad; it doesn't return a mathematically wrong result, gives a readable error message (this doesn't always happen in Sage!), and by reading solve? one can see the proper syntax. It isn't a good thing, but isn't horrendous.

[ppurka]

comment:2

Added a simple patch.

[ppurka]

Author: Punarbasu Purkayastha

[fchapoton]

comment:4

Thanks for the ticket. Are you sure of the syntax of the links to trac ? According to #12490, it seems rather to be

:trac:`13077`

with no # inside.

[ppurka]

Attachment: trac_13286-allow_list_in_solve.patch.gz

Apply to devel/sage

[ppurka]

comment:5

Replying to @fchapoton:

Thanks for the ticket. Are you sure of the syntax of the links to trac ?

Thanks for checking this. Indeed, it was incorrect. Have corrected the patch now.

[fchapoton]

comment:6

ok, positive review