In cases where CLI tidy fails to recognise HTML with errors or
unknown HTML5 elements, it doesn't output any content and
CSSContentParser fails.
This adds the --force-output parameter so that it always outputs
the input content, even if it wasn't completely fixed by tidy.
It seems that tidy class in PHP already does this, so this just
fixes the CLI use of tidy in CSSContentParser to be consistent.
This fix can be observed on the HHVM travis builds, as tidy isn't
compiled into HHVM as an extension.
Versioned is not writing Version to _version tables for subclasses of Version dataobjects which have their own DB fields
- Fix disjoint of ID / RecordID (which should be the same)
- Fix calculation of new record version
- Fix use of empty vs !isset to check for existing version
The text on the button is hidden from view via CSS already. This allows the button to have text in the event that the GridField does not provide icons (and likely is better for your screen readers).
saveInto() assumed you were either saving into a hasOne or if you wanted to create a new object.
This made it impossible to have a FileField on an object that you wanted to upload to. This adds a check so that if the local object is a file, save to that.
When sorting a DataQuery over a relation, the SQLQuery automatically included the sort column. The issue with the implement is that potentially the joined record has a field with the same name as the source record causing it to be overridden.
In the attached test case, without the patch the title will be set to 'Bar' rather than 'Foo'.
This patch aliases the sort column. Alternativally a patch would be to
It breaks logic flow, e.g. when
Its called by BasicAuth:requireLogin() when basic auth is enabled,
before any controller logic kicks in (on every HTTP request).
This means you can't use session-based BackURLs with basic auth enabled,
breaking flows like redirection after Facebook logins.
I can't see why a clear() was necessary here, looks like a overly
cautious way to prevent infinite loops? Can't see how those
would be caused by requireLogin() though.
Been there since all the way back in 2007: a377a67e54
BUG Fixes missing i18n translation in Date::TimeDiffIn
BUG Fixes Date::TimeDiffIn not respecting mocked SS_Datetime::now
This provides less vague date periods. I.e. "36 days" has a lot more relevance that "1 month"
Reduced duplication of time period calculation code
(ref: CWPBUG-141)
These should be avoided because they undermine the process of
peer review and merging in github, we should strive to have
zero open pull requests, as opposed to treating it as a stage
for work in progress. Intermediary code review can happen in github forks instead.
Also remove some checklist items which were based on the Trac bugtracker,
e.g. its not longer possible to assign yourself to issues because
of github's limited permission abilities.
Example: you have a site in a sub-directory off the webroot, you call
->Link() on a SiteTree record, which returns "/[sitedir]/my-page", and
you pass this URL to Director::test(). It's a valid URL, but
Director::test() will throw a 404.
Director::test() should be ensuring that all URLs passed to it are
properly made relative, not just in the case where it thinks the URL
is absolute.
