Limited string size sample

[aplastiras]

Hi,

I am using hypothesis-jsonschema to generate test sample data :

I am using custom_formats for string data as shown below:

size_values = []

@settings(max_examples=10000)
@given(from_schema({"type": "string",
                    "format": "card",
                    "maxLength": 1000},
                   custom_formats={'card': text(printable)},))
def generate_samples(data):
    size_values.append(len(data))
    pprint([data, len(data)])
generate_samples()

print('Max sample length: %d'%max(size_values))

...
['KJK gh', 6]
['KXA', 3]
['KAA', 3]
['KKA', 3]
['p?0', 3]
['??0', 3]
['?00', 3]
['?0?', 3]
['00?', 3]
['00?0Q\rer', 8]
Max sample length: 46

The issue is that by having set maxLength = 1000 as part of json schema the generated strings have size 40-60 characters max even after generating 10000 samples.

Is there any configuration I am missing that is restricting the size of the string?
How can I stress the string size to its limit?

Thank you in advance
Apostolos

[Zac-HD]

This is a result of Hypothesis' underlying design, and unfixable without making it radically less effective at finding bugs on realistic problems with eg nested schemas. There's also no config that would touch this; the internal average_size param is, well, internal.

So my advice is the remarkably unhelpful "don't worry about it, it's fine" 😕