+1 and good catch, Jon! I feel guilty for not spotting the visual changes in the docs previously.
The new test is nice. I wonder if we should add empty declarations to the style sheet for those unsettled classes, maybe even with a short comment about where it is used. That would make using these classes simpler. On the other hand it would inflate the style sheet a little bit. Hannes > Am 24.03.2020 um 23:46 schrieb Jonathan Gibbons <jonathan.gibb...@oracle.com>: > > I forgot to add the bug number into the @bug list in the test. I'll do that > in the next webrev (if one is needed) or before pushing. > > -- Jon > > > On 3/24/20 3:45 PM, Jonathan Gibbons wrote: >> Please review some simple fixes to the stylesheet to address some issues >> caused by a couple of recent changes: (1) to use "-page" as a suffix for the >> CSS class name for <body> elements, and (2) for the change to hyphenated >> naming. >> >> A significant new test case is added to the existing TestStylesheet.java >> class. The test determines the set of known CSS class names, based on the >> CSS class names defined in the stylesheet, and a set of known additional >> class names, and then reads the output generated from a small sample API, >> and verifies that the values of all the "class" attributes are in set of >> known CSS class names. >> >> A copy of the API documentation is provided. The primary kind of page to >> check is a regular type declaration page, such as for java.lang.Object. >> >> -- Jon >> >> JBS: https://bugs.openjdk.java.net/browse/JDK-8241544 >> Webrev: http://cr.openjdk.java.net/~jjg/8241544/webrev.00/index.html >> API: http://cr.openjdk.java.net/~jjg/8241544/api.00/index.html >>
