fredbi / uri / 6212471233

89%

push

github

enhancement: replaces regular expressions by explicit parsing following (#20)

the RFC grammar.

Performance
-----------
* Removed all usage of regexp's.
* Replaced IPv6 parsing by standard lib functionality.

Overall, the benchmark shows a speed improvement of x5 when compared to
v0.1. This is still 2x slower than URL parsing with the standard libary.

RFC strictness
-----------
This PR fixes a few strictness issues:
* identifies wrong URIs with empty authority and path starting with a
double '/'
* identifies invalid IPv6 addresses with an empty zone
* strictier check of DNS names
* strictier check of URL-encoded escape sequences
* allow escape sequences in IP addresses
* strictly follows the RFC rule: "if not a valid IP, then consider a
registered name": hence
* strictier IP kind distinction: IPv6 must be between [], IPv4 must not

Caveats: IPv6 relies on the standard library and is therefore not very
strict regarding valid IPv6.

Fixes
-----------
* fixes #14

Signed-off-by: Frédéric BIDON <fredbi@yahoo.com>

184 of 184 new or added lines in 1 file covered. (100.0%)

613 of 627 relevant lines covered (97.77%)

252.44 hits per line