Changes between Initial Version and Version 1 of Ticket #6200, comment 26


Ignore:
Timestamp:
2021-09-30T12:03:04Z (4 years ago)
Author:
Karim Adbdelhak

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #6200, comment 26

    initial v1  
    55The main rationale was that i still use actual zero comparison even for real valued coefficients (not with an epsilon, just flat on equality). The odd thing is that it triggered for this case. The only reason that comes to my mind is either another epsilon used in the generated C-code, or we hit our precision threshold and it is rounded in such a way that it actually ends up being zero.
    66
    7 Either way, i added the config flag `-d=realASSC` which defaults to false.
     7Either way, i added the config flag `--realASSC` which defaults to false.