fix: hash request kwargs and headers correctly

[yetval]

Summary
Request fingerprint collisions: Request.update_fingerprint() previously hashed only kwarg names and lowercased header values. Distinct requests could collapse to the same fingerprint when fp_include_kwargs or fp_include_headers were enabled, which can silently break scheduler deduplication, cache replay, and checkpoint restore. Fix: hash kwarg names together with their values and preserve header values as-is.

Repro
from scrapling.spiders.request import Request

r1 = Request("https://example.com", timeout=1)
r2 = Request("https://example.com", timeout=2)

assert r1.update_fingerprint(include_kwargs=True) != r2.update_fingerprint(include_kwargs=True)

Files changed
scrapling/spiders/request.py — fingerprint kwargs/header handling

[samrusani]

I did a quick local check against this patch and the behavior matches the PR description:

from scrapling.spiders.request import Request

assert Request("https://example.com", timeout=1).update_fingerprint(include_kwargs=True) != Request("https://example.com", timeout=2).update_fingerprint(include_kwargs=True)
assert Request("https://example.com", headers={"X-Test": "A"}).update_fingerprint(include_headers=True) != Request("https://example.com", headers={"X-Test": "a"}).update_fingerprint(include_headers=True)

tests/spiders/test_request.py also passed locally for me after applying the patch (28 passed).

One small suggestion: it would be worth adding these as regression tests in tests/spiders/test_request.py, since the contribution guide asks bug fixes to include code that reproduces the bug. That would lock in both the kwargs-value collision and the header-value case handling.

[yetval]

Hey @samrusani, I have went ahead and added tests. Thanks for the suggestions!

[D4Vinci]

Using str(value) for kwarg values is fragile if someone passes a non-primitive.

[yetval]

Thanks! Just fixed this.

[D4Vinci]

LGTM

[D4Vinci]

Good catch as always @yetval
We will merge once the checks pass again.