Live services shouldn't be able to request to go live again. Once a
service is live we remove the option to go live from the Settings page,
but we still link to the page to request to go live from other places
e.g. the 'Get started' page. As a result, we've seen some services make
another request to go live when their service has already been live for
months - this change will stop that from happening.
We look for `original_file_name` in the metadata now. Initially we were
still checking the query string too, but now that the change to add the
filename to the metadata has been deployed for a while there shouldn't
be any cases of the filename still being in a query string.
Since the `original_file_name` is not being added to the metadata in
`.check_messages` (it has happened earlier in the process) a few tests
are no longer needed.
The code was looking for `original_file_name` in the metadata for a
contact list, or the query string if it wasn't in the metadata. Now that
the change to use the metadata for the file name has been deployed for a
while e can stop looking in the query string for the
`original_file_name`.
When sending from an uploaded CSV `.send_messages` now puts the filename
in the metadata. It previously used the query string to pass the
filename to `.check_messages`, where it can be lost.
The `.send_from_contact_list` function redirected to `.check_messages`
with `original_file_name` in the query string. Contact lists already
have `original_file_name` as part of their metadata, so we can stop
sending it in the query string and use the metadata instead.
We were passing `original_file_name` from the `.upload_contact_list`
view function to the `.check_contact_list` view function as a query
param. We now store it in the metadata instead. `.check_contact_list`
still checks for `original_file_name` in the query string if it's not in
the metadata - this is necessary until the code has been deployed for a
few days and we can be sure that there are no contact lists that are
mid-way through the upload stage.
If a subclass of `JSONModel` defines a property then we shouldn’t try
to override it with the value from the underlying dictionary.
Rather than silently fail we should raise an exception because it will
help keep our list of `ALLOWED_PROPERTIES` nice and tidy.
I'm mainly making this change because it's useful
for the CSS that styles the hint text when the
link is focused for the link to have no parent
container.
That being said, there isn't really enough content
underneath these headings to justify them as it is.
I've wrapped them in a list instead because:
- they're structured like a list
- we already called them a `template-list`
This commit also replaces the `message-type` class
on the paragraph below where the headings went,
for consistency. It also removes the CSS for that
class as I couldn't find anywhere else that used
it now.
When the list of areas is restricted to half the width of the page it
starts to look pretty higgledy-piggledy when you have lots of areas or
areas with very long names.
To do this I’ve ripped out the table markup in favour of headings,
paragraphs and lists. Probably pros and cons for each, but it was really
hard to do the layout with the content in a table.
For emails and text messages we sort by the time the user (or API) sent
them.
This makes sense for broadcasts too, since most users will receive the
alert within seconds of it being broadcast.
For alerts that haven’t started yet we can sort by `updated_at`, which
is when the user preparing the broadcast submitted it for approval.
Now that pending alerts aren’t in their own section there’s nothing to
label them as pending. So this commit replaces the extra metadata we
show for a pending alert (the name of the person who created it, which
was only ever a reckon) with an explicit label that says it’s waiting
for approval.
Splitting the dashboard into multiple sections was confusing, and people
sometimes mistook the headings as labels, especially when a section was
empty. It just wasn’t clear what the hierarchy of the page was.
This commit combines the current and pending broadcasts into one list
on the dashboard. Previous broadcasts have already moved to their own
page.
If you refresh the page on a current broadcast while someone has
cancelled it you’ll see the wrong navigation item selected. This commit
adds redirects to take you to the correct endpoint in these edge cases.
Once a broadcast has been submitted for approval it either lives on the
‘Current alerts’ or ‘Previous alerts’ page, depending on where it is
in its lifecycle.
Therefore when clicking into a broadcast from one of those pages the
same navigation item should remain selected.
Because we select the navigation items based on the request endpoint,
this means we need an endpoint for each navigation page, even if the
content of the pages will be the same in both cases.
This commit adds the two new end points, removes the old, single
endpoint and updates links to point to the new endpoint.
We were trying to delete the old 'template-{template-id}' keys but
should have been deleting the new keys which have the service id as part
of the key name. This was causing the cache to not be correctly purged
when we did things like update sender names or set defaults. This should
fix it.
We wrote custom `__getattr__` and `__getitem__` implementation to make
it easier to find code that was accessing fields in the dictionary that
we weren’t aware of. See this commit for details:
b48305c50d
Now that no view-layer code accesses the service dictionary directly we
don’t need to support this behaviour any more. By removing it we can
make the model code simpler, and closer to the `SerialisedModel` it
inherits from.
It still needs some custom implementation because:
- a lot of our test fixtures are lazy and don’t set up all the expected
fields, so we need to account for fields sometimes being present in
the underlying dictionary and sometimes not
- we often implement a property that has the same name as one of the
fields in the JSON, so we have to be careful not to try to override
this property with the value from the underlying JSON
The most important part of the broadcast is what content was sent where
(and when).
This commit reduces the priority of the ‘meta’ information, like who
prepared and approved the broadcast. I also think that the ‘end’ time is
a lot less important than the start time, since most people will receive
the alert at or near to the start time.
The idea was that this would be a place to document all the design
patterns used in Notify. However it hasn’t been kept up to date, and,
looking at the `git blame`[1] no new patterns have been added for 5
years.
I think it’s better to get rid of it than have to keep maintaining
something which is inaccurate.
1. 64aa0d359c/app/templates/views/styleguide.html
Our style for areas is pale blue background with black keylines or bold
black text.
This commit makes the display of area names on the dashboard consistent
with that visual style.
This also means that we’re not truncating the list of areas, which is
appropriate because no one area is more important than any of the
others.
Broadcast services only have broadcast templates. But we show the
template type under the name of the template. This is redundant. It
would be better to preview the content of the template instead.
This then makes the templates page consistent with the dashboard.
Depends on:
- [ ] https://github.com/alphagov/notifications-api/pull/2996
The dashboard for normal services is quite general, because it tells
you a bit about channels, templates and spend.
What is now the dashboard for broadcast services is much more specific,
therefore less like a dashboard. We can reflect this by giving it a more
specific name. This should reduce the amount of navigation surfing
people need to do in order to find the thing they’re looking for.
Previous alerts are much less important than ones that are live or
waiting for approval.
Therefore we can make the dashboard more focused by moving previous
alerts to their own page.
This prefixes everything to do with areas/the map with `area-list`, so
from looking at one element you know which `.scss` file will contain the
relevant code.
It’s not clear where 103 is coming from unless you know that the code is
doing the slightly unintuitive thing of displaying the max of all return
values.
We don't need these anymore as all users will use the `one-off/step`
routes.
This has mostly involved tidying up the tests which are still a little
disorganised and not as good as I'd like but it's a step in the right
direction.
More refactoring is still possible to the routes, it may come in a later
PR if I have time.
