mirror of
https://github.com/opencv/opencv.git
synced 2025-01-19 15:04:01 +08:00
16ea1382f7
In that function, the floats are cast to int to be compared to 0. But a float can be -0 or +0, hence define CHECK_NZ_FP(x) ((x)*2 != 0) to remove the sign bit. Except that can trigger the sanitizer: runtime error: signed integer overflow: -1082130432 * 2 cannot be represented in type 'int' Doing everything in uint instead of int is properly defined by the standard. |
||
|---|---|---|
| .. | ||
| 3rdparty/SoftFloat | ||
| cmake/parallel | ||
| doc | ||
| include/opencv2 | ||
| misc | ||
| perf | ||
| src | ||
| test | ||
| CMakeLists.txt | ||