Make scan type list rely on Dynamic parser infra.

[damiencarol]

Simplify code around scan type lists and parser loading/validating making some serializer/forms don't rely on DB but on code directly.

[Maffooch]

Why is there such an urgency to remove the test_type table from the DB?

[damiencarol]

@Maffooch Sorry the text was confusing. I will change the text of the PR. The goal is to have Test_type less "fixed.
I mean, as Test_type could depends on new parsers loaded dynamically, it doesn't make sense to rely on the data in the DB for the UI.

[Maffooch]

Okay that sounds better. I was under the impression the goal was to remove the entire table from the database.

[Maffooch]

Things look good to me

[StefanFl]

@damiencarol I will approve this merge request if you do me one favour 😃: The list of parsers at the import page is not sorted. Please use get_choices_sorted() for the scan types in dojo/engagement/views.py/import_scan_results.

[damiencarol]

I tried but it's not that simple. To do that we need to modify TestForm to fake the link between scan_type and the test_type attribute of a Test.
I failed to found an elegant way to do it. Will try again today.

[StefanFl]

Not sure we talk about the same thing. I meant using get_choices_sorted() in line

django-DefectDojo/dojo/engagement/views.py

Line 664 in b44af80

'scan_types': get_choices(),

That does the trick.

[damiencarol]

@StefanFl done.