Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fixes a segfault
|
|
The did when this was all done with threads, but with separate processes each worker has its own copy of
the structure already. It can be updated after every image and freed at the end.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This was needed before, but not any more.
|
|
|
|
|
|
|
|
This is almost completely benign, and happens when profiling.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Previously, they were just used to filter out reflections outside the resolution range, and the bins would
be calculated from the resolution limits of whatever was left. Now, the given values are used exactly
for calculating the bins.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|