GitLab

see #3762

see #3206

see #3206

Following the fix on min_cover_representative.
The representative on small datasets may be longer.

Thus we need to update those. There also some side effects:
- V genes are less ambiguous (hence less warnings)
- Consensus is longer (hence less warnings too)
- Positions changes of end of V, J, CDR3…

Use the new option names.
Remove also '--trim 0' options used in some tests.

See #3295.

Since IGHV4-59*09 can be found instead of IGHV1-8*01, we can replace it in this
test file.

For more informations see #3298.

See #2611.

This will help testing the effect of some parameter on all the functional tests.

This will be more flexible.

The tests using Stanford-S22 mostly test things on the extracted windows,
sometimes on the representatives, but almost never on the fine segmentation.
We thus add '-y 0', '-z 0', or similar options to these tests.

On some laptop, running all *.should_get tests now takes 2'34 instead of 3'48.

Additionally, one test has a different output due to a previously forgotten '-d' option.