| Age | Commit message (Collapse) | Author | 
|---|
|  |  | 
|  | Turns out this is necessary for the changelog. Wow. | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | "Clear float" that is! | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | Undefined isn't implied by empty() call so this is a behavioral
change, too. | 
|  |  | 
|  | Previous commits are A-OK for making html.isBlank() work properly
with [html.onlyIfSiblings] tags - and consequently, parent tags
with [html.onlyIfContent] - but if the parent *doesn't* have this
attribute, it needs to take responsibility in its own toString /
function. | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | This logic doesn't appear to be necessary to avoid overriding
compositional-infrastructure errors (or nested errors) when
reporting, really. | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | Most notably drops automatically setting ownTrackData.
Hilariously, this doesn't break any tests (presumably because
they were generally updated to handle a newer - albeit not the
newest - track section format). | 
|  |  |