rfc3986

1.4.0 -- 2020-04-07

Security

• Prevent users from receiving an invalid authority parsed from a malicious URL. Previously we did not stop parsing the authority section at the first backslash (\\) character. As a result, it was possible to trick our parser into parsing up to the first forward-slash (/) and thus generating an invalid authority.

  See also GitHub pr-64 and the blog post that sparked this change

Bug Fixes and Features

• Add from_uri to URIBuilder to allow creation of a URIBuilder from an existing URI.

  See also GitHub pr-63

• Fix a typographical error in our documentation.

  See also GitHub pr-61

1.3.2 -- 2019-05-13

• Remove unnecessary IRI-flavored matchers from rfc3986.misc to speed up import time on resource-constrained systems.

  See also GitHub #55

1.3.1 -- 2019-04-23

• Only apply IDNA-encoding when there are characters outside of the ASCII character set.

  See also GitHub #52

1.3.0 -- 2019-04-20

• Add the IRIReference class which parses data according to RFC 3987 and encodes into an URIReference.

  See also GitHub #50

1.2.0 -- 2018-12-04

• Attempt to detect percent-encoded URI components and encode % characters if required.

  See also GitHub #38

• Allow percent-encoded bytes within host.

  See also GitHub #39

• Correct the IPv6 regular expression by adding a missing variation.

• Fix hashing for URIReferences on Python 3.

  See also GitHub !35

1.1.0 -- 2017-07-18

• Correct the regular expression for the User Information sub-component of the Authority Component.

  See also GitHub #26

• ~rfc3986.validators.Validator.check_validity_of to the ~rfc3986.validators.Validator class. See Validating URIs documentation for more information.

1.0.0 -- 2017-05-10

• Add support for 6874 - Zone Identifiers in IPv6 Addresses

  See also issue #2

• Add a more flexible and usable validation framework. See our documentation for more information.

• Add an object to aid in building new URIs from scratch. See our documentation for more information.

• Add real documentation for the entire module.

• Add separate submodule with documented regular expression strings for the collected ABNF.

• Allow None to be used to eliminate components via copy_with for URIs and ParseResults.

• Move release history into our documentation.

0.4.1 -- 2016-08-22

• Normalize URIs constructed using ParseResult.from_parts and ParseResultBytes.from_parts

0.4.0 -- 2016-08-20

• Add ParseResult.from_parts and ParseResultBytes.from_parts class methods to easily create a ParseResult

• When using regular expressions, use [0-9] instead of \d to avoid finding ports with "numerals" that are not valid in a port

0.3.1 -- 2015-12-15

• Preserve empty query strings during normalization

0.3.0 -- 2015-10-20

• Read README and HISTORY files using the appropriate codec so rfc3986 can be installed on systems with locale's other than utf-8 (specifically C)

• Replace the standard library's urlparse behaviour

0.2.2 -- 2015-05-27

• Update the regular name regular expression to accept all of the characters allowed in the RFC. Closes bug #11 (Thanks Viktor Haag). Previously URIs similar to "http://http-bin.org" would be considered invalid.

0.2.1 -- 2015-03-20

• Check that the bytes of an IPv4 Host Address are within the valid range. Otherwise, URIs like "http://256.255.255.0/v1/resource" are considered valid.

• Add 6 to the list of unreserved characters. It was previously missing. Closes bug #9

0.2.0 -- 2014-06-30

• Add support for requiring components during validation. This includes adding parameters require_scheme, require_authority, require_path, require_path, require_query, and require_fragment to rfc3986.is_valid_uri and URIReference#is_valid.

0.1.0 -- 2014-06-27

• Initial Release includes validation and normalization of URIs