Added localization demo #1481
Added localization demo #1481
Conversation
|
@bdarnell if you think we'd like to add this demo, I can do the initial code review. |
|
We do need an i18n demo, but i18n is a process, more than just "how do I use these functions". At a minimum, any demo needs to reasonably represent a real application (no hard-coded locales), and be ready for message extraction (which we unfortunately don't support yet. I consider Tornado's i18n support incomplete without either a solution for #622 or a non-gettext-based message extractor, although the only internationalized Tornado project I've actually worked on was Friendfeed, which got by without message extraction unless they had something that wasn't open-sourced). I had actually started a long time ago to make a demo based on http://google-opensource.blogspot.com/2011/06/pseudolocalization-to-catch-i18n-errors.html. I should probably dust off that branch: https://github.com/tornadoweb/tornado/compare/master...bdarnell:locale?expand=1 |
|
excited to see what comes afer ;) |
For unittesting I start and stop the server in setUp and tearDown. If I use Application.listen I don't have a reference to the HTTPServer and cannot close the socket. Binding it to the same address the next time will result in a address in use error. If the server is returned, calling close on it will allow the address/port to be reused.
If an HTTP response contains both Content-Length and Transfer-Encoding headers, flag this as an error as per RFC 7230, Section 3.3.3#3. Also added a unit test to validate the code.
Make test_root_static_path more robust.
Check for Content-Length and Transfer-Encoding
Return HTTPServer in Application.listen
Remove redundant parts of README; link to tornadoweb.org instead.
Current coverage is
|
I think people are looking for details on the localization;
https://groups.google.com/forum/#!topic/python-tornado/olR9J3ThH8I
I wrote a localization demo quickly. Maybe this could help better understanding.