I'm working on automating the extraction of the D grammar from the docs I have got it down to one sed script and about 6 calls to grep. However a good bit of the sed script is fixing typos. For example

A number of rules are missing the ':'

FunctionLiteral
Operands
AsmPrimaryExp
DotIdentifier
SuperInterfaces
TemplateParameterList
VersionSpecification
DebugSpecification
DeclarationBlock

A number of other rules have odd formatting:

(In attribute.html)
const
override
static
auto
scope

If these were fixed I think I can programmatically extract the grammar. I'll see if I can put together a patch for the sources in dsource/phobos

BCS wrote:
> 
> If these were fixed I think I can programmatically extract the grammar. I'll see if I can put together a patch for the sources in dsource/phobos
> 
> 

attached is the patch... Er... 3 patches:

typo_docs.patch has mostly simple edits adding ':' in a few places and what not

opts.patch moves optional parts of NewAnonClassExpression into a sub rule

and "gname.patch" replaces the $(I ) macro with the $(GNAME ) macro

However, because I wasn't thinking well at the time "gname.patch" also has all of the other 2 patches and might even have the another edit (I don't remember)

with the patches I can strip out the D grammar from the .html doc using sed.