Changes for version v0.011_003

  • use 5.8.5: perlver says that the only requirement for 5.8 is the constant hash
    • v0.011_002: attempt to lower to 5.6.0, and change from constant hash to multiple `use constant` lines
    • v0.011_003: edit Makefile.PL to also 'use 5.6' instead of '5.8.5'
  • Test suite for :raw754 now switches to '==' comparison for numeric arguments, and check for 64bit NV
    • bug @ <https://rt.cpan.org/Public/Bug/Display.html?id=116004>; still exists in v0.011_001
    • v0.011_002: change test suite to use a '%+24.16e' got eq expect, rather than raw got == expect
    • v0.011_002: issue did not appear during CPAN Testers, so believed fixed
  • nextup():
    • bug @ <https://rt.cpan.org/Public/Bug/Display.html?id=116006>,
    • could not replicate; but adding diagnostics showed that when LSB underflows (goes to -1), the hexification is too many F's; perhaps on some systems, that was being misinterpreted when the two were being recombined, so masked the MSB and LSB before merging
    • v0.011_001: issue did not appear during CPAN Testers, so believed fixed
    • v0.011_002: issue did not appear during CPAN Testers, so believed fixed
  • Test suite for :floatingpoint now converts SNAN to expect /[SQ]NAN/, because it doesn't really matter that much whether Perl silences a signaling NAN
    • bug @ <https://rt.cpan.org/Public/Bug/Display.html?id=116005>
    • v0.011_001: issue did not appear during CPAN Testers, so believed fixed
    • v0.011_002: issue did not appear during CPAN Testers, so believed fixed

Modules

Various tools for understanding and manipulating the underlying IEEE-754 representation of floating point values