Commits on Sep 2, 2023

1. Resolve integer overflow when running on a large 3D image …

   In a computation like x+y*nx+z*(nx*ny) if the sub-computation y*nx or
   z*(nx*ny) overflows the size of a 32-bit integer, we will get a bogus
   result even though the final destination is a 64-bit integer.
   The solution is to promote at least one operand to a bigger-sized
   integer. Since long can also be 32-bit on 32-bit architectures, we just
   use explicit named types here.
   This patch also fixes a few cases of comparisons of integers with
   different signs.
   
   Resolves: stardist#235

   

   ximion committed Sep 2, 2023
   Configuration menu

   • View commit details
   • Copy full SHA for 1f17f25
   • Browse repository at this point

   Copy the full SHA

   1f17f25 View commit details

   Browse the repository at this point in the history