[Prev][Next][Index][Thread]

Re: HTML Tidy Wrapper for Functional Developer

Klemens Schmid <klemens.schmid@gmx.de> writes:

> I have problems with the compile of TidyCOM. I get lots of
> unresolved external I guess because the tidy.lib cannot be
> found. Can this come from the different configurations of Tidy and
> TidyCOM?

As far as I know you don't need the tidy.lib. All I did on a clean
machine to test the library was put the TidyCOM.zip files in a
directory, run regsvr32.exe on the TidyCOM.DLL and then built the
library in FD.

Have you got the COM libraries for FD available from the fun-o
website? I'm also using the MS linker if that makes a difference. If
this still doesn't work can you send me the log file of the link stage
or a list of the FD errors and I'll see what I can find.

> I have another problem when I pass the resulting XML file to
> e.g. XML notepad. He refuses to load because of duplicate
> attributes. Some special & expressions like &uuml; &middot; &copy;
> are not treated by tidy (not by your code) and therefore cause
> errors in XML parsers.  Do you know by heart where to treat this?

My tidycom dylan library only converts the HTML to well formed HTML,
not XML or XHTML. Things like fixing overlapping tags. I then parse
the code using a Dylan HTML parser. To convert to XML you need to use
some of the TidyCOM options. This involves doing something like:

  let s = "<html>...</html>"
  let tidy := make(<ITidyObject>, class-id: $TidyObject-class-id);
  ITidyOptions/OutputXML(tidy.ITidyObject/Options) := #t;
  let result = bstr-to-byte-string(ITidyObject/TidyMemToMem(tidy, s));

See the implementation of my html-tidy() method for more
details. Eventually I plan to extend the dylan wrapper so it has an
easy way to set the various HTML tidy options like XML output, etc.

Hope that helps,
Chris.
-- 
http://www.double.co.nz/dylan
Follow-Ups: References: