This document highlights the current differences from CommonMark:
-
Entities without a final semi-colon, such as
&
, are not required by CommonMark, but are valid HTML. The argument given in the spec, “because it makes the grammar too ambiguous”, seems quite a weird reason for me to build extra code which removes this functionality from the excellent entity parser remark uses. -
remark-html adds two extra entities when compiling to HTML, namely,
`
(tick) and'
(single quote), which are rendered as`
and'
. The first has to do with better support for IE8 (a tick can break out of an attribute or comment), the second for an optional future option to use single quotes as attribute quotes. -
remark-html ignores any markdown syntax in image
alt
attributes, whereas commonmark strips it from the alt attribute. Thus,![*foo*](./bar.png)
is rendered as<img alt="foo" src="/bar.png" />
in CommonMark, and as<img alt="*foo*" src="/bar.png" />
in remark-html.