Opened 5 years ago

Last modified 3 years ago

#18697 new defect

any FP number in an Expression without symbol should trigger evaluation

Reported by: rws Owned by:
Priority: major Milestone: sage-6.8
Component: symbolics Keywords:
Cc: Merged in:
Authors: Reviewers:
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: #10035 Stopgaps:

Description (last modified by rws) 

sage: 0.1 * cos(pi/13)
0.100000000000000*cos(1/13*pi)

This is annoying because the 0.1 at once limits the precision of any other non-symbolic expression.

​https://github.com/pynac/pynac/issues/71

Change History (7)

comment:1 Changed 5 years ago by rws

  • Report Upstream changed from N/A to Reported upstream. Developers acknowledge bug.

comment:2 Changed 5 years ago by rws

  • Dependencies set to #10035

People will complain if they cannot change this behaviour, so we need a hold context.

comment:3 Changed 5 years ago by rws

  • Description modified (diff)

comment:4 Changed 5 years ago by rws

  • Report Upstream changed from Reported upstream. Developers acknowledge bug. to Fixed upstream, but not in a stable release.

comment:5 Changed 5 years ago by rws

  • Description modified (diff)
  • Report Upstream changed from Fixed upstream, but not in a stable release. to N/A

Partial "fix" of Sage code removed, it was actually an unrelated Pynac issue, which is fixed in 0.4.2.

comment:6 follow-up: Changed 3 years ago by rws

The problem with this ticket is shown by 

    sage: pi + RBF(1)

The expected output would be [4.141592653589793 +/- 5.61e-16], i.e. the same as with RBF(pi) + RBF(1). Constant conversions thus need to take the right parent into account. There may be multiple incompatible parents and that fact may be hidden deep in the expression tree.

Last edited 3 years ago by rws (previous) (diff)

comment:7 in reply to: ↑ 6 Changed 3 years ago by rws

Replying to rws:

There may be multiple incompatible parents and that fact may be hidden deep in the expression tree.

This is not a problem because we are evaluating depth-first. The only problem at all seems de-facto limited precision constants, in spite of #18255. Also this might be in the way: ​https://github.com/pynac/pynac/issues/311

Note: See TracTickets for help on using tickets.