bcftools stats --samples-file give wrong results if sample order is different than vcf

[niederhuth]

I'm using bcftools v1.21

If I run bcftools stats with --sample-files, I've noticed that the per-sample statistics will be incorrect if the sample order in the input list of samples does not match the order of the samples in the input vcf. The statistics in the output file will keep the original order in the vcf file, but the order of the names will match the order in the input file.

[pd3]

I was not able to reproduce this behavior. Can you provide a small test case, the file and the exact command you are using, highlighting where is the output different from your expectation?