few more questions, todo updates, and changelog clarifications

CHANGELOG.md

@@ -1,19 +1,26 @@
 # Hyperlink Changelog
 
-## 17.0.0
+## "dev"
+
+* Python 2.6 support
+
+## 17.1.0
 
-* Improved error on invalid scheme, directing users to URL.from_text
-  in the event that they used the wrong constructor
 * Correct encoding for username/password part of URL (userinfo)
 * Dot segments are resolved on empty URL.click
 * Many, many more schemes and default ports
 * Faster percent-encoding with segment-specific functions
 * Better detection and inference of scheme netloc usage (the presence
-  of // in URLs)
-* IPv6 support
+  of `//` in URLs)
+* IPv6 support with IP literal validation
 * Faster, regex-based parsing
 * URLParseError type for errors while parsing URLs
 * URL is now hashable, so feel free to use URLs as keys in dicts
+* Improved error on invalid scheme, directing users to URL.from_text
+  in the event that they used the wrong constructor
+* PEP8-compatible API, with full, transparent backwards compatibility
+  for Twisted APIs, guaranteed.
+* Extensive docstring expansion.
 
 ## Pre-17.0.0
 

TODO.md

@@ -1,7 +1,6 @@
 # hyperlink TODO
 
-* CI
-* Get Mark to fix the git history
+* Polish logo
 * Optimize percent decoding
 
 ## Complete
@@ -17,12 +16,19 @@
 * pct encode userinfo
 * `__hash__`
 * README
+* CI
+* Get Mark to fix the git history
+* Python 2.6 support
+* Full read the docs
 
 ## Questions
 
 * What's the deal with twisted.python.urlpath?
-* What's the deal with sip.URL
-* Do we need a separate .normalize()?
+* What's the deal with sip.URL?
+* Should we take the alternative IPv6 literal approach and just
+  interpret anything with colons in the host as requiring square
+  brackets? Or should we try to validate with a regex?
+* Do we need a separate .normalize()? Resolves all dots, lowercases scheme and host.
 * Is whitespace really allowed in IRIs?
 * Can we drop the ':' when password isn't displayed?
 * Why is the constructor's default scheme "http" when host is not None?