File acq_private.h¶
FileList > acq > acq_private.h
Go to the source code of this file
Shared internals for acq_core.c and its test. More...
#include "doppler/acq/acq_core.h"#include <math.h>#include <stddef.h>
Public Static Functions¶
| Type | Name |
|---|---|
| int | acq_rate_better (size_t h, float stat, float best) Does hypothesis h's statisticstat replace the best so far,best ? The rule the rate search picks its winner by. |
| size_t | acq_resolve_tile_alias (const dp_acq_state_t * st, size_t row, size_t col) |
Detailed Description¶
Not part of the public API: nothing here is exported, and nothing here is promised to a caller. It exists so a rule that decides a result (the rate search's pick, the tile-edge alias's hop) can be unit-tested as a rule (test_acq_core.c includes it), the pattern detector/det_private.h follows.
Public Static Functions Documentation¶
function acq_rate_better¶
Does hypothesis h's statisticstat replace the best so far,best ? The rule the rate search picks its winner by.
Hypothesis 0 always wins the first place, whatever it reads: a dwell whose every statistic is NaN (a power that overflows float is inf/inf) would otherwise fail stat > best at every hypothesis, pick none, and leave the search restoring the PREVIOUS dwell's winning surface. After it, a hypothesis must beat the best strictly, so a NaN never displaces one and ties keep the earlier.
Unit-tested as a rule on NaN / inf / finite values, because the end-to-end route to a NaN statistic (a float power overflowing) depends on the compiler's code generation and differs between Release and Debug.
function acq_resolve_tile_alias¶
The documentation for this class was generated from the following file native/inc/doppler/acq/acq_private.h