Package: hhsuite
Version: 3.0~beta2+dfsg-1
Severity: minor
Tags: upstream patch
User: reproducible-bui...@lists.alioth.debian.org
Usertags: randomness
Dear Maintainer,
hhblits.cpp uses the %i printf(3) conversion to format a 'double' value:
src/hhblits.cpp:333: printf(" -pre_evalue_thresh max E-value
threshold of Smith-Waterman prefilter score (default=%i)\n",
par.prefilter_evalue_thresh);
Formally that's undefined behaviour, although in practice it simply
prints a garbage value. That garbage value is also embedded into the
hhblits_omp(1) man page.
Furthermore, hhsuite-userguide.tex specifies the default value as '0'
whereas the code uses '1000'.
The attached patch should fix both issues.
Cheers,
Daniel
P.S. ViterbiRunner::calculateEarlyStop() puts the return value of log(3)
into a 'float' variable, even though that function returns double. I'm
not sure whether that's intentional or what effect it may have.
