Flake8 Bugbear checks for some extra things that aren’t code style
errors, but are likely to introduce bugs or unexpected behaviour. A
good example is having mutable default function arguments, which get
shared between every call to the function and therefore mutating a value
in one place can unexpectedly cause it to change in another.
This commit enables all the extra warnings provided by Flake8 Bugbear,
except for the line length one (because we already lint for that
separately).
It disables:
- _B003: Assigning to os.environ_ because I don’t really understand this
- _B306: BaseException.message is removed in Python 3_ because I think
our exceptions have a custom structure that means the `.message`
attribute is still present
The data flow of other bits of our application looks like this:
```
API (returns JSON)
⬇
API client (returns a built in type, usually `dict`)
⬇
Model (returns an instance, eg of type `Service`)
⬇
View (returns HTML)
```
The user API client was architected weirdly, in that it returned a model
directly, like this:
```
API (returns JSON)
⬇
API client (returns a model, of type `User`, `InvitedUser`, etc)
⬇
View (returns HTML)
```
This mixing of different layers of the application is bad because it
makes it hard to write model code that doesn’t have circular
dependencies. As our application gets more complicated we will be
relying more on models to manage this complexity, so we should make it
easy, not hard to write them.
It also means that most of our mocking was of the User model, not just
the underlying JSON. So it would have been easy to introduce subtle bugs
to the user model, because it wasn’t being comprehensively tested. A lot
of the changed lines of code in this commit mean changing the tests to
mock only the JSON, which means that the model layer gets implicitly
tested.
For those reasons this commit changes the user API client to return
JSON, not an instance of `User` or other models.
Since we added template folders the templates page has had a ‘medium’
sized heading, where other pages have stuck with a ‘large’ size.
This commit rationalises the decision around which pages have which
heading size:
- ‘navigation’ pages (eg templates, team members, email reply to
addresses) have medium sized headings
- transactional pages (ie ones which have a green button) keep the
larger heading size
The Design System has standardised on back links being at the top of the
page, decorated with a small text-coloured arrow.
I think this makes more sense than having them at the bottom, because it
suggests, in some way, being able to go back before commiting to any of
the forms on the page. Whereas the things at the bottom of the page
should be performing actions on what’s in the page.
The reason for making this change now is that it de-clutters the area
around the green buttons. This was presenting a design challenge where
multiple levels of interaction were happening in the same form. Moving
these back links to the top of the page should mean that, in these
complicated forms, there’s one fewer thing to compete for the user’s
attention.
I’ve componentised this into a `page_header` macro so that the change is
easier to roll out and maintain.
It:
- saves repetetive boilerplate code
- does some extra checks (eg checking for a `200` response)
- makes the codebase less confusing to consistently do the same thing in
the same way
Currently the order of API keys seems to be non-deterministic:
d46caa184e/app/dao/api_key_dao.py (L32-L39)
Generally we sort things alphabetically, unless there’s a good reason to
do otherwise.
This commit sorts the API keys alphabetically.
It’s redundant to make two API calls here, one to get all keys and one
to get a single key. Since the API calls are sequential we can speed
things up by getting the one key from the list of all keys.
Once all our users have upgraded to the latest clients they won’t need
this. The latest clients only use the combined key and service ID.
Discuss: when can we safely remove it?
Pytest is deprecating the direct calling of fixtures. One fixture that
we call directly quite a lot is `fake_uuid`. Since it just returns the
value of `sample_uuid()` we can either call that instead (where we need
a fixed value) or generate a new UUID each time (where a fixed value is
not needed).
It’s useful to be able to see what the email or text message looks like,
especially if you’ve sent it with a test API key (so it isn’t in your
inbox or on your phone). We already have the page for this, so we just
need to link to it.
This is better than just keying into the JSON because it means you get
an exception straight away when looking up a key that doesn’t exist
(which via mocking you could ordinarily miss).
We’ve had a user who’s said:
> Seems configured callbacks cannot be removed once they’re set as the
> fields have a presence check. Is that intentional?
This means it’s not working as they expect. Rather than have to go and
change stuff in the database for them, let’s make it work as they’d
expect.
Only lets you clear the form if you remove both the token and the URL.
we were seeing isort produce different outputs locally and in docker -
this was due to it having different opinions about whether the tests
module (ie all our unit tests) is a first party (local) or third party
(pip installed) import. It's a first party import, so by defining this
in the setup.cfg isort settings, we can force it to be consistent
between environments.
Note: I don't know why it was different in the first place though
test_api_keys.py and test_api_integration.py were almost identical
files with only a few lines difference between them. By moving one
test we can now delete test_api_keys.py
Precompiled letters can now have two additional states:
* pending-virus-check
* virus-scan-failed
Both new states should show in the notifications dashboard, and
virus-scan-failed should appear as an error state, with a descriptive
message. You should not be able to preview a letter in one of the two
new states, so the preview link has been removed for precompiled letters
in these states.
Done using isort[1], with the following command:
```
isort -rc ./app ./tests
```
Adds linting to the `run_tests.sh` script to stop badly-sorted imports
getting re-introduced.
Chosen style is ‘Vertical Hanging Indent’ with trailing commas, because
I think it gives the cleanest diffs, eg:
```
from third_party import (
lib1,
lib2,
lib3,
lib4,
)
```
1. https://pypi.python.org/pypi/isort
Both `<button type='submit'>Submit<button>` and
`<input type='submit' value='Submit'>` can be used to submit a form.
We have historically[1] used `<input>` because it’s better-supported by
IE6 in that:
- the `submit` attribute is mandatory on `<button>`, not on `<input>`
- the `innerHTML` of a button will be submitted to the server, not the
value (as in other browsers)
Reasons to now use `<button>` instead:
- IE6/7 support is no longer a concern (especially with deprecation of
TLS 1.0 on the way)
- Because an `<input>` element can’t have children, the pseudo-element
hack[2] used to ensure the top edge of the button is clickable doesn’t
work. We’re seeing this bug[3] affect real users in research.
1. We inhereted our buttons from Digital Marketplace, here is me making
that change in their code: 8df7e2e79e (diff-b1420f7b7a25657d849edf90a70ef541)
2. 24e1906c0d (diff-ef0e4eb6f1e90b44b0c3fe39dce274a4R79)
3. https://github.com/alphagov/govuk_elements/issues/545
