Changes between Initial Version and Version 1 of Ticket #29156, comment 4


Ignore:
Timestamp:
10/12/20 02:40:56 (13 months ago)
Author:
gh-bmlivin
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #29156, comment 4

    initial v1  
    1 This is something that might be a bug in GiNaC, or it might be that Sage is using GiNaC incorrectly. Here is an excerpt from GiNaC's `numeric::power` in `numeric.cpp`, which is what ends up being called in this example:
     1This is something that might be a bug in pynac, or it might be that Sage is using pynac incorrectly. Here is an excerpt from pynac's `numeric::power` in `numeric.cpp`, which is what ends up being called in this example:
    22
    33{{{
     
    2222Here, `v._bigint` is the absolute value of the exponent's numerator, 21111111111. That happens to be a bit larger than 32 bits, so the if statement doesn't execute, and `throw_conversion_error` is called.
    2323
    24 I haven't combed through GiNaC's documentation to see whether they point out that `numeric::power` shouldn't be called with negative rational exponents whose numerators are greater than 32 bits, but I can't see why they would want to do this and not do the same with positive numerators or denominators greater than 32 bits. So I ''think'' this is a bug in GiNaC.
     24I haven't combed through pynac's documentation to see whether they point out that `numeric::power` shouldn't be called with negative rational exponents whose numerators are greater than 32 bits, but I can't see why they would want to do this and not do the same with positive numerators or denominators greater than 32 bits. So I ''think'' this is a bug in pynac.