- Currently the test is taking the response date
which is in UTC and passes it through moment()
which by default parses input as local time. We
then add 12 hours to this now local time
and when compared against the response set-cookie
header the time spread is wrong.
- To fix we’re parsing the response date with
moment.utc which parses the date in UTC.
closes#1661
- disables save button at beginning of image upload
- enable save button at image upload fail, at image upload and download success and also at dropzone init
issue #1789
- we can do this step a bit earlier and it aids with neatness
- also means for tests that config.load() is enough to get config & paths initialised
closes#1470
issue #1405
- added cache control middleware
- added defaults for all routes, assets, etc
- updated asset helper to add a query string with a timestamp hash to all assets
- added unit tests for asset and ghostScriptTags helpers
- added cache-control checks to route tests
fixes#1800
- changed permalink regex to require 4/2/2/slug
- changed url helper to enforce the same
- changed permalink toggle to set a specific state, this means the
functional tests are independent again
- chnaged permalink toggle to wait for the settings page to load
- change as many frontend tests to not login as possible
closes#1757 and #1773
- switches routes.frontend for posts and pages
to use a regex with two capturing groups. This removes
the need to dynamically remove an express route at a
later point, leaving the decision making to frontend
controller.
- added unit tests for all routing conditions that
can arise for posts and pages.
- updated functional tests to also test for same thing
in unit tests
- removes old code from server/api/index that used
to fix this issue, but is no longer needed
- removed some un-needed require statements in routes/admin
closes#1777
- Uses coreHelpers.url to resolve post url instead of assuming it is siteUrl + post.slug
- Functional feed tests now check for content instead of just rss tags
closes#1681
- import doesn't override user credentials
- import doesn't override theme
- import doesn't kill session
- import does refresh the settings cache
- updated tests, they now use a fixture instead of a generated export
- tests check to ensure import is safe
Closes#1770
- Previously, the middleware would check that the route on the stack was an empty string, which worked when there was no subdirectories
- When subdirectories were added, the proper route was only set when
updating the theme
- Because it was only set when updating, this explains themes working on
initial load, since the stack location was looking for an empty
string, which is what the middleware was initialized with
- However, once a new theme was set, it was still look for an empty
string, which would never exist, which caused the issue
- Now, the route is properly set on initialization of the middleware,
and then the `config.paths().subdir` property is used for the check
issue #1754
- remove path (it was only used once, and not needed)
- change webroot to subdir
- add unit tests for config.paths
- various other cleanup
- renamed client-side ghostRoot to subdir
- added url helper for client
fixes#1696
- this is a temp workaround until full fledged support
is added directly to bookshelfjs
- when importing we use the import json blob’s timestamps
as the value that’s set in the DB
- added tests for this change
fixes#1749
- pass config().url into theme.update
- rename paths functions to match theme function
- adds tests for theme config
- We should probably try to eliminate passing config.url around
Reported in issue #1475
- `String.localeCompare` can be more reliable for keys that do not relate to cursor movement
- Adds a third key handler (`keypress`) that contains the character code rather than physical key
- `COMMA` key constant no longer required (unless `,` char should be a constant?)
