not print sharg version into help menu

[feldroop]

I feel like printing the sharg version into the help menu does not really help the user of an application. It might actually be confusing.

[smehringer]

Hi @feldroop,

not sure if I understood you correctly. Do you think

VERSION
    Last update:
    Eat-Me-App version:
    Sharg version: 1.1.2-rc.1

is confusing and it should be

VERSION
    Last update:
    Eat-Me-App version:

?
And can you elaborate on what is confusing?

[feldroop]

Yes, that's pretty much what I think.

I feel like people who use command line tools are generally not interested in the version of the command line parser their tool uses. Especially not biologists who use bioinformatics tools. Most people will just ignore it, but it's still unnecessary noise. In the worst case, desperate users, who struggle with debugging their workflows, might think: "What is this sharg thing in the help menu? Do I have the wrong version? Maybe that's why it doesn't work?". Most likely sharg is not the reason for their problems and it's not like they can change the sharg version independently of the tool version either.

I would argue that it's an unnecessary leakage of an implementation detail.

[eseiler]

It's not so uncommon:

$ samtools --version
samtools 1.16.1
Using htslib 1.16

SeqAn2 does it too:

$ mason_genome --version
mason_genome version: 2.0.10 [77c2180]
SeqAn version: 2.5.0

We even print the seqan3 version if it's used

$ fasta_to_fastq --version
fasta_to_fastq
==============

VERSION
    Last update:
    fasta_to_fastq version:
    Sharg version: 1.0.1-rc.1
    SeqAn version: 3.3.0-rc.1

Another thing is visibility (of the sharg lib), and IMO --version is quite an unintrusive way to do that.

[feldroop]

htslib and SeqAn are core dependencies of the tools. E.g. samtools even writes on their github that they are built upon htslib. I feel like it is completely different with sharg. For example, the htslib version might be useful for troubleshooting the compatibility with other tools. When would the sharg version help a user?

The second point is valid, but IMO an argument parsing library should do its job silently.