Before this was only possible for some specific ones, like onBeforeWrite.
This excludes any callbacks with augment*() or update*() naming,
since these are assumed to be on extension only, with a corresponding
base method available on the class itself (e.g. "updateCMSFields()"
vs "getCMSFields()").
- Refactored SiteTreeURLSegmentField to render controls in template
rather than JS for better clientside performance, and cleaner behaviour.
- Added dynamic ellipsis to start of URL, to retain most relevant
part of the URL (the last bits)
- Added "suffix" setting to field, which defaults to ?stage=Stage
- Removed prefix from edit view to leave more room for URL
Thanks to @sunnysideup for getting this started in
https://github.com/silverstripe/silverstripe-cms/pull/269
We'll need to fix the "no space left on device" issue,
most likely caused by Postgres keeping too much of a query log,
or somehow creating a history of past data.
For now, having a Postgres build breaking the whole
build process (incl. MySQL builds) does more harm than good.
If the redirect is to an internal page, the preview iframe
will automatically load this new URL, causing the CMS edit
form to switch as well - effectively making it impossible
to edit a redirector page in the CMS as soon as it is saved
with an internal redirection target.
Utilise the new features provided by the framework to get richer
interface:
* save buttons that highlight the current state of the page
* minor actions in a drop-up
* embed last publishing and saving information
Moving to another page in the GridField for the root assets folder in AssetAdmin
caused all files to be displayed. This was caused by the functionality that tries
to fix some edge cases for detail view kicked in and didn't filter the list properly.
Fixed by making the edge case detection a little more specific (i.e. checking if
a numeric ID was sent as a request variable).
Used to return ArrayList, but since 9d74c99e08 its no
longer using keys, which they functionality using these return
values rely on. Hence the breakage of ReportAdmin,
since no report was being detected as "current" by looking
through the array keys.
This reverts commit a63a9f0e74.
Broken by merging in 0be51a93213, which works on 3.0,
but failed once we put those two patches together in 3.1.
They fix the same problem on different ends, and assume
the respective other behaviour still exists.
