In action=parse, the documentation states that "contentmodel - Content model of the input text. Default is the model of the specified title, or wikitext if title is not specified" long ago. However, https://git.wikimedia.org/commitdiff/mediawiki%2Fcore.git/1104392eba1b7701582054fed11d353868614d62 discourages doing so by showing a warning that there should be a contentmodel if title is not specified. Hence, now the code and the documentation are inconsistent. To make them consistent, either suppressing the warning or changing the documentation to discourage input without contentmodel must be done. I don't know, though, that which way is better.
Version: unspecified
Severity: normal