`__getattr__` is called whenever an attribute error is raised.
This means that if something deep inside a property on a model raised
an attribute error, that error would be caught by `__getattr__`, which
would then raise an exception that looked like the property itself
didn’t exist. Very confusing.
The solution seems to be to override `__getattribute__` instead, which
handles _all_ attributes, not just those that aren’t explicitly defined.
We then only intervene if the desired attribute is one of the
`ALLOWED_PROPERTIES`, otherwise falling through to the built in methods
of the underlying `object`.
If we change our mind and decide whether a service should/should not be
counted in the list of live services then we should also drop the cache
which stores the count of how many live services there are.
This makes it consistent that an option which contains more options has
a hint about how many options it contains.
Also adds a formatter to get us ready for 1,000 services 🎉
Also add more tests for showing or not the cancel those letters link
Also check if all notifications already in database
Upgrade delete button text logic to handle more cases
Also corrections following review
If you can see a folder but not its parents we concatenate the
breadcrumb into one link.
This styles folder separators inside these links a bit differently to
make them do a bit less visual separation than the ones outside the
links.
It looks weird to have two different visual treatments for showing a
navigable hierarchy.
I reckon losing the slash won’t make things less folder like – Windows
for example uses chevrons as foler separators.
It was a bit inconsistent depending on whether there was/wasn’t a search
box or channel tabs on the page.
I found this just too complicated to do in pure CSS, so added a new
spacing class which gets toggled on and off.
Now that there’s a bit more stuff in the service name area at the top
of the page it looks a bit cramped. Moving the heading down gives it a
bit more space to breath, and associates the heading a bit more closely
with the content after it.
This commit aligns and spaces elements on the page to show which are
related to others.
This needs some adjustment now because we potentially have more things
on the page now – we need to make space for them.
Service names can be quite long. Organisation names can be quite long.
Together they can be very long. This isn’t great because:
- sometimes they overflow the width of the container, which looks broken
- even if they’re not that long they can make the UI look quite
cluttered
This commit restricts them to widths that should stop the above from
happening. In the case of the organisation name the width has
specifically been chosen to line up with the ¼ and ¾ column grid
used by the navigation.
Currently we set not-very-useful defaults for organisation type and
crown status when creating an organisation. This commit adds two field
to the form (in addition to the existing name field) to explicitly ask
for:
- organisation type
- crown status
We need these for all organisations before we can make any of their
services live.
This commit also records any new organisation as not having accepted the
data sharing and financial agreement, because if we don’t know about the
organisation already then they definitely won’t have signed it.
Rather than force us to write the decorators in a specific order let’s
just have one decorator call the other. This should make fewer lines of
code, and fewer annoying test failures. It also means that the same way
of raising a `401` (through the `current_app` method) is used
everywhere.
At the moment we mostly have `user_has_permissions` execute first. It
shouldn’t matter, but it feels right for us to check that a user is
logged in before we check their permissions to a service. Otherwise a
malicious user could (maybe) check if a service ID belongs to a real
service, and go on to do something malicious with that information.
This commit adds some extra test code to enforce that the order is
always the same.
N.B. decorators in Python execute from closest to furthest (from the
line on which the function is defined).
We accidentally miss these sometimes. This code adds a test which
inspects the code to automatically check that any function which:
- handles a route
- accepts a service_id
For each function it checks that each of these routes have the
permissions decorator we’d expect.
Most of the introspection/AST code is adapted from here:
https://mvdwoord.github.io/exploration/2017/08/18/ast_explore.html
A user might not have a guessable organisation type, even if the service
they’re working on does have an organisation set. This can happen for
users with @nhs.net email addresses, for example.
