| Commit message (Collapse) | Author | Age | Lines |
... | |
|
|
|
|
|
| |
This tests that the default string used by the old-style Open311 update
code if there is no incoming string and the report is marked fixed, is
translated.
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
bin/update-schema
locale/sv_SE.UTF-8/LC_MESSAGES/FixMyStreet.po
perllib/FixMyStreet/DB/Result/Body.pm
|
| |
| |
| |
| | |
This could be made cobrand-customisable in future.
|
| |
| |
| |
| |
| |
| | |
This commit reverses the "Remove unused Open311 GetUpdates code." commit.
FixaMinGata, the Swedish FixMyStreet cobrand, is still using
open311-update-reports.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Provide a button to download the details and text of reports
at the second "Look up" button on /dashboard.
Provides acknowledged, fixed, closed dates by parsing updates.
Fixes #692 (and mysociety/FixMyStreet-Commercial#399)
Email is not shown, and name honours anonymous flag: as /dashboard
is accessible to people who are from_body (who still do not see
anon-users name and email in the FMS web app.)
A basic test is provided as per Zurich's similar report.
|
| |
| |
| |
| |
| |
| |
| |
| | |
The tests have been commented out due to long-running failure. It looks
like the failures happen if working copy MAPIT_URL is not pointing at
live site. Have updated with an override_config setting this to live
site (which is obviously suboptimal, but seems to be widespread practice
across t/ at the moment.)
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Originally copied from template of Bromley code, substituting logo,
council ID, geolocation data, and renaming .scss variables etc.
(NB: this was a bad idea. As per notes/cobranding.txt, should have
started with less copying of files, as this required more cleanup.
Bromley's code has some particularities, like smaller map, split
First/Last name fields and so on.
The current branch has been rebased to present a slightly idealised
picture of the process.)
(Includes contributions by @davea and @stevenday)
* Some code cleanup and added is_two_tier
* Add Hart-specific links to footer
* Shift FMS nav out of the main header
* Update styling to follow design more closely
* Flip nav colours to better indicate links
* Make header banner entire width of the page
* Stretch footer bg to entire width of page
* Add top-level nav links
* Correct header height on mobile
* Correct wording of top menu item
* Remove FMS logo from top of page
* Shift top level nav to correct position
* Invert media query so IE8 displays header properly
* hide folding triangle, as makes no sense on this cobrand
* Some more renaming of Bromley to Hart
* Treat Hart the same as Bromley in the JS
* Whitespace/formatting fixes
* Basic test for cobrand
* Javascript #map_box fixes
* Make sure #map_box is not in desk-only
* ensure there is only single map (one .wrapper only)
* colour tweaks
* Fix example places (postcode, spelling)
* Basic FAQ text
* Set Zoom level 3
* CSS/HTML/JS tweaks removing Bromley customizations
* Switch templates to put navigation in footer, as FMS.com does
* remaining HTML/CSS changes for large map
* replace favicon
* move site-logo scss to one place
* make small logo transparent
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Fixes #677. The query to fetch the related alert sometimes:
* retrieves no result, causing an error on $alert->id
* retrieves multiple results, causing deprecation warnings
(e.g. which may at some point become future errors in DBIC)
This commit turns the resultset('DB::Alert')->find(...) into a
search(...) and loops over the results, to suppress them.
|
| | |
| | |
| | |
| | |
| | | |
The call to set_lang_and_domain needs to happen before
the first string is translated (the report state).
|
|/ / |
|
| |
| |
| |
| |
| |
| | |
Put alerts in default cobrand, so changes to other cobrand email
templates don't affect the test; and clean up at the end of the
test so that it can be rerun without issue.
|
| | |
|
| | |
|
| |
| |
| |
| | |
for Zurich tests, count how many reports already exist as a baseline.
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This value is now cached, but the template calls the function,
calculating it anew, so for old reports (closed within SLA) the
problem_row.html template showed these as overdue. Instead, we point
the template at a function which first checks the cached value.
(Attempting to roll this functionality into the .overdue() method itself
caused a fair amount of collateral test breakage, and it seems safest to
extract the caching function for now.)
Closes https://github.com/mysociety/FixMyStreet-Commercial/issues/462
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
See https://github.com/mysociety/FixMyStreet-Commercial/issues/461
Again, we just need to substitute the %params line (which elides
hidden reports) for the less restrictive %date_params.
This commit includes very basic tests for the previously untested
?export=1 functionality.
|
|/ /
| |
| |
| |
| | |
Test that the slash is added in correct place (the base URL should not
have its own trailing slash).
|
| |
| |
| |
| | |
So that tests pass on clean dataset such as travis...
|
| | |
|
| |
| |
| |
| |
| |
| | |
As per comments after testing mysociety/FixMyStreet-Commercial#459, the
business logic of the actual state changes is fine, but the stats do
also then need to include hidden records.
|
| |
| |
| |
| |
| | |
To allow us to infer additional stats changes requirements such
as in mysociety/FixMyStreet-Commercial#459 better in future.
|
| |
| |
| |
| |
| | |
This was causing previous test failures when config != Zurich, due
to expected logic not getting triggered.
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
First test that this page is only available to superuser.
Then test that the moderated and fixed stats are doing something that
is (hopefully) the spec in mysociety/FixMyStreet-Commercial#459
|
| |
| |
| |
| |
| |
| | |
- is set to 0 on state change within time-limit
- is set to 1 on state change after time-limit
- only marks overdue on first time
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When a body has been marked as deleted (currently only exposed in the
Zurich cobrand's admin) the categories for that body should not be
available on the report_new form.
Prior to this change there was an error running the
'zurich-overdue-alert' script, because it wasn't expecting the
'bodies_str' field to have any entries in the format 'x|y,z', this
prevents this problem from happening in the future (though it doesn't
fix any existing entries with this issue).
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If the ALLOWED_COBRANDS configuration variable only contains one entry
(and also work if it's a string rather than a list), always use that
cobrand, no matter what the hostname is.
The example Vagrantfile no longer needs the configuration changes at
all.
|
| |
| |
| |
| | |
Use a relative URL in the CSS in case it's proxied.
|
|\ \ |
|
| | | |
|
|\ \ \
| |/ /
|/| |
| | | |
https://github.com/andylolz/fixmystreet into andylolz-389-report-search-ref
|
| | | |
|
| | | |
|
| | |
| | |
| | |
| | | |
This fixes #527.
|
| | | |
|
| | | |
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The phone number was made mandatory in #541. To make this change
compatible with older versions of the mobile app I've had to drop
server side validation of the phone number for requests that come in
from mobile apps.
To determine if it's a mobile request I'm checking the 'service'
parameter in the request, which is only set in requests from the mobile
apps [1].
The phone validation will still happen client side for the newer
versions of the app, and the web app will continue to work with the
server side validations.
[1] https://github.com/mysociety/zurich_mobile/blob/1fec2bc21dad9326819105cdbf5d47183e7503ba/www/js/models.js#L75
|
| |
| |
| |
| | |
Fixes #586
|
| |
| |
| |
| |
| | |
If SET isn't given, the line means [% foo = ('bar' IF baz) %]
rather than [% ([SET] foo = 'bar') IF baz %].
|
| |
| |
| |
| | |
Closes mysociety/FixMyStreet-Commercial#410
|
| |
| |
| |
| | |
You only need to pass in the values you wish to be different.
|
| |
| |
| |
| |
| | |
This is for consistency as Sub::Override is used for another test and
also so we minimise the number of modules installed just for testing
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| | |
The first email count test in the abuse subtest was failing on Travis
but not on my machine because the sent email alert test was running on
my machine, as it has the mock module installed, and this was resetting
the bodies_str to one with a confirmed contact. Setting bodies_str on
both these tests stops this sort of thing happening again
|