Reader tests for the marCCD header-offset clamp and the 2-theta circle

Two cases on the existing byte-for-byte marCCD fixture: a file whose
end-angle block is zeroed while the detector sits on a 2-theta arm must
still name phi as the circle that turned, and a file whose tag 34710
names an offset far past its end must be refused rather than allocated
for. Also drops a bound in SweepLayout that the absolute jitter
threshold makes redundant.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013nW6FNRP1bBJJ8pfHiByAT
This commit is contained in:
leonarski_fandClaude Opus 5 committed 2026-09-20 19:23:32 +02:00
1 parent 1b9b6d7405
commit 65b9ae1259
2 files changed
+41 -4

No files matched your search

+1 -1
View File
@@ -142,7 +142,7 @@ Layout Place(const std::vector<Frame> &frames, const std::string &logger_name) {
double guess = 0;
for (size_t i = 1; i < frames.size(); i++) {
const double d = Fold(angle[i] - angle[i - 1]);
if (std::abs(d) > std::max(too_fine, 1e-6) && (guess == 0 || std::abs(d) < std::abs(guess)))
if (std::abs(d) > too_fine && (guess == 0 || std::abs(d) < std::abs(guess)))
guess = d;
}