Closed RagnarGrootKoerkamp closed 2 years ago
Hi, Ragnar.
First, thanks for all your bug reports. I deeply appreciate your effort and contribution to the project. This helps me a lot!
WRT this problem. Can you double-check with the current development
branch? I have made some tests, and the current HEAD is returning 0. But, an older version/commit might perfectly behave as you indicate.
Let me know. Thanks,
This turned out to be because we forgot to set the alignment scope to only compute the score, which is initialized to some (negative) infinity when by default the path is returned.
When using BiWFA/ultralow memory mode and indel costs, aligning two empty sequences seems to return
-2147483648
as the cost. This only happens when calling the library directly (via rust), and not when runningalign_benchmark
aswhere
input
simply containsI.e., this could very well be a bug on my side in the rust-c interface.