Re: [xsl] Tracking entity references

["Peter Flynn peter@xxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

On 28/03/2023 11:23, Wegmann, Frank frank.wegmann@xxxxxxxxxxxxxx wrote:
> This may be a very silly question, but I couldnbt find an answer on
> it, with Saxon, or outside it, using tools relying on expat.
>
> Use case is a set of old XML documents (traditional, inhouse DTD)
> that make overly excessive use of entities of all kinds. While it is
> easy to get all entity declarations, and also to locate references of
> external entities, I could not locate references of text entities. By
> the time I see it the entity reference has already been resolved. How
> can I achieve that (ideally with line/column of the actual place in
> the original file or entity)?

I would edit or otherwise process a copy of the DTD (and any entity
files) and extract all the entity names to another file in the form of
system entity declarations to a non-existent file, then edit the DTD and
entity files to delete or comment out all the entity declarations.

(This does of course need robust checking because it's a non-SGML
operation, and DTD syntax is notorious for authors putting stuff in
weird places because the parser is super-tolerant of spacing. Entities
within inclusion or exclusion exceptions may need hand working :-)

If you then run (eg) onsgmls invoking the newly-made catalog of bogus
system entity declarations and set the max error number up high, you
should get a ton of error messages, one for each entity reference that
fails to resolve, giving the line and character number in the document
where it occurs.

Peter