When we moved from 1/3rd 2/3rd columns to 1/4th 3/4th columns we should
have excluded the tour page. The tour page needs the width of the 1/3rd
column to look right.
Users might be interested in international SMS. And when it’s fully
available, they’ll probably be able to control whether it’s on/off for
their service.
Until they point, the only way of getting it is to ask us. So let’s
make an in-the-meantime page that directs them to ask us, from the place
where they’d be able to do it themselves.
If a service can send internationally, our CSV validation should not
catch valid international phone numbers. This means calling through
to code added to utils in:
- [ ] https://github.com/alphagov/notifications-utils/pull/156
This has been removed from utils (so things will break if utils is
upgraded without this change isn’t made). I think it’s friendlier to
present the phone number as the user entered it anyway – because this is
what they think a ‘correct’ phone number representation looks like
anyway.
If you go back to this part of the registration flow then you get a 500
error, because we’re relying on something in the session. We clear the
registration info from the session after you’ve registered successfully.
Also there were no tests for the happy path of this page either
¯\_(ツ)_/¯
We’ve seen in research a user getting stuck playing with the
scheduler. They picked a day, but then didn’t want to choose one of the
options for that day. There’s no way to do this except pick a day and
then un-pick it.
What they ended up doing was clicking the grey back button, which took
them back to the previous page, making them upload their file again.
This commit adds a ‘back’ link for the scheduler. ‘Back’ seems like
sensible naming because that’s the thing that the user tried to click,
and the UI of a link matches the thing they clicked to get into this
situation.
Previously they were relative (ie percentages). This made sure that they
worked on mobile, when the letter might be narrower.
However it broke when the preview was more than one page, because
13% of the height of 2 pages is different to 13% of the height of one
pages.
This commit changes the positions to be pixel values, which match the
calculated percentage values when the preview is one page.
It was confusing because it didn’t do anything. We think the research
tomorrow will go more smoothly if we remove it. It should come back
in the same place when it actually works.
