feature: support rfc-compliant bracketed ipv6 addresses in URLs.

[andrewsuperlegit]

@robinst
fix for issue 78

Updates URL parsing to support RFC 3986 and RFC 6874 bracketed IPv6 addresses within a valid URL scheme.

• replaces the for loop in find_authority_end with a while let loop to handle ipv6 extraction which apparently allows rust to use a jump table
• when an open bracket is encountered, standard domain validation is suspended. the parser validates ipv6 valid characters-- including zone identifiers aka percentage signs until the closing bracket is found
• maintains near zero performance regression on standard plaintext by ensuring normal characters bypass the ipv6 branch stuff
• adds dedicated ipv6 benchmark cases and a substantial number ipv ipv6 unit tests
• centralizes some testing utilities into tests/common/mod.rs to prevent duplicating test logic/functionality.
• clippy remains undefeated

NOTE: this adds ipv6 addresses within valid schemes-- schemeless bracketed IPs are intentionally excluded to prevent performance regressions on standard text.

[andrewsuperlegit]

I think i'm going to do a little bit of a refactor to clean up the contribution a little bit. i mean let me know what you think of it overall but i'm going to extract my portion into something thats a little easier to read.

[andrewsuperlegit]

full disclosure, after the refactor i noticed the following performance regressions:

---

Gnuplot not found, using plotters backend
no_links time: [13.742 ns 13.762 ns 13.784 ns]
change: [-2.2038% -1.7054% -1.2243%] (p = 0.00 < 0.05)
Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
2 (2.00%) high mild
1 (1.00%) high severe

some_links time: [142.94 ns 143.55 ns 144.27 ns]
change: [+12.596% +12.965% +13.367%] (p = 0.00 < 0.05)
Performance has regressed.
Found 5 outliers among 100 measurements (5.00%)
1 (1.00%) high mild
4 (4.00%) high severe

heaps_of_links time: [517.40 ns 521.46 ns 526.98 ns]
change: [+7.2029% +9.0657% +11.216%] (p = 0.00 < 0.05)
Performance has regressed.
Found 7 outliers among 100 measurements (7.00%)
1 (1.00%) high mild
6 (6.00%) high severe

some_links_without_scheme
time: [181.52 ns 183.25 ns 185.82 ns]
change: [+5.8928% +6.8380% +7.7802%] (p = 0.00 < 0.05)
Performance has regressed.
Found 5 outliers among 100 measurements (5.00%)
1 (1.00%) high mild
4 (4.00%) high severe

---

that said i'm on a laptop so those might not be real performance hits- by all means check; just wanted to call it out. but if they are real its up to you whether you want me to get rid of the refactor. I think it's cleaner and easier to read the way i have it but it's not my codebase.

[robinst]

Some comments, but overall looks good. Note that the rustfmt check failed.

I don't think the benchmarks would be affected that much because of an additional case for [ in the loop, so it might just be noise. But I'll run the benchmarks on my side when I have some time.