The problem is that some versions of awk, including the one in the cactus docker image (mawk) return scientific notation for numbers > 2Gb (32bits), and this trips up Cactus which casts it to int.
This PR is a little patch that specifies float output, which seems consistent across awk flavours and number sizes and returns output readable by the existing code.
In no way does this mean that cactus-pangenome will now work well on giant genomes... it's just this one particular crash that should be fixed.
cactus-graphmap-splitusesawkat some point to make a contig size table.The problem is that some versions of
awk, including the one in the cactus docker image (mawk) return scientific notation for numbers > 2Gb (32bits), and this trips up Cactus which casts it to int.This PR is a little patch that specifies float output, which seems consistent across
awkflavours and number sizes and returns output readable by the existing code.In no way does this mean that
cactus-pangenomewill now work well on giant genomes... it's just this one particular crash that should be fixed.Resolves #1125