- 10 Nov, 2017 1 commit
-
-
John Hume authored
-
- 21 Oct, 2017 1 commit
-
-
John Hume authored
-
- 16 Oct, 2017 4 commits
-
-
John Hume authored
-
John Hume authored
Previously: heroku create app-name --buildpack https://github.com/emk/heroku-buildpack-rust.git
-
John Hume authored
-
John Hume authored
-
- 03 Aug, 2017 1 commit
-
-
John Hume authored
-
- 01 Aug, 2017 2 commits
- 29 Jul, 2017 2 commits
-
-
John Hume authored
So Todo no longer needs to do double-duty: it always represents a saved item. Now url needn't be optional.
-
John Hume authored
so the web API has a clean interface to interact with and doesn't know how Todos are stored. Meanwhile, the TodoList is ALMOST totally decoupled from the web. (It at least doesn't know about Rocket.)
-
- 28 Jul, 2017 5 commits
-
-
John Hume authored
-
John Hume authored
-
John Hume authored
This gets us the last 2 tests: "can PATCH a todo to change its order" and "remembers changes to a todo's order"
-
John Hume authored
This gets us the first test in "tracking todo order:" "can create a todo with an order field"
-
John Hume authored
This gets us the next test: "can delete a todo making a DELETE request to the todo's url"
-
- 26 Jul, 2017 3 commits
-
-
John Hume authored
This gets us through two more tests: "can change the todo's completedness by PATCHing to the todo's url" and "changes to a todo are persisted and show up when re-fetching the todo"
-
John Hume authored
which gets us the next test: "can change the todo's title by PATCHing to the todo's url" LOTS of duplication at this point, though most of it is an artifact of the in-memory state, which is hopefully a temporary hack anyway.
-
John Hume authored
which gets us the next two tests: "each new todo has a url, which returns a todo" and "can navigate from a list of todos to an individual todo via urls" which are kind of redundant.
-
- 25 Jul, 2017 3 commits
-
-
John Hume authored
No comment on the quality of that URL. This passes the next test: "each new todo has a url"
-
John Hume authored
This passes the next test: "sets up a new todo as initially not completed" Using the same type for requests and proper TODOs is starting to get weird.
-
John Hume authored
So much for tiny changes! This passes the first of the "storing new todos by posting to the root url" tests: "adds a new todo to the list of todos at the root url" Several questionable decisions here.
-
- 24 Jul, 2017 1 commit
-
-
John Hume authored
This cheap trick gets us through the last of "the pre-requisites:" "after a DELETE the api root responds to a GET with a JSON representation of an empty array"
-
- 23 Jul, 2017 5 commits
-
-
John Hume authored
but has no real side-effect. This passes the third test "the api root responds successfully to a DELETE"
-
John Hume authored
This passes the second test. "the api root responds to a POST with the todo which was posted to it"
-
John Hume authored
-
John Hume authored
This passes the first test on http://www.todobackend.com/specs/index.html?http://localhost:8000 "the api root responds to a GET (i.e. the server is up and accessible, CORS headers are set up)"
-
John Hume authored
-
- 21 Jul, 2017 3 commits