[gs-bugs] [Bug 689794] gs gets /rangecheck in --initgraphics-- when compiled with -ffast-math

bugs.ghostscript.com-bugzilla-daemon at ghostscript.com bugs.ghostscript.com-bugzilla-daemon at ghostscript.com
Sat Apr 19 14:29:19 PDT 2008


http://bugs.ghostscript.com/show_bug.cgi?id=689794





------- Additional Comments From williambader at hotmail.com  2008-04-19 14:29 -------
Is transitivity the correct word?
Transitivity means A=B and B=C implies A=C, which involves three values.
Changing a condition F(x) from floor(x)==x to abs(floor(x+0.5)-x)<=eps affects a
function of only one variable, so transitivity is not relevant.
Changing a condition F(x,y) from floor(x)==y to abs(floor(x+0.5)-y)<=eps would
be a different matter.

My guess is that the original code was not for performance optimization but to
prevent small rounding errors from creating rotation matrices that cause
problems inverting.

Changing the condition causes differences only on rotations very near a multiple
of 90 like 89.9999999999999.  Would this ever be an issue except in test
patterns?  Printers can place at most on the order of 1000 dpi, so 6 or 7 digits
should be enough to position dots on the length of a page.  If two rotations
differ by 1e-15 degrees, how large a sheet of paper and how small dots would you
need before a dot is shifted to another position?



------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.



More information about the gs-bugs mailing list