re: [cgmo-webcgm] <object> interoperability issue

[Don Larson <dlarson@cgmlarson.com>]

Lofton,

 >  Implementors (et al) --
 >
 >  Stuart volunteered when I asked for a straightforward test case of <object
 >  data=">/<param>:" "Volunteer? (To turn it into a simple, minimal complete
 >  test case of the 3.4 Example?)"
 >
 >  I decided to modify his test to make the simplest possible <object> test
 >  case: one <object> element referencing a CGM, and no <param>. See attached.
 >
 >  Results: I tried 3 viewers (don't have access to Ulrich's right now).
 >
 >  PTC's very old Beta 6 passed. The other two did not.
 >
 >  Don, Forrest (, Ulrich) -- could you look into it please? This is a most
 >  basic test of referencing a CGM into an HTML document. It really should
 >  work, regardless of whether the viewer is capable of more advanced
 >  configurations (like loading via DOM), and regardless of whether IE
 >  sometimes has issues with <param> child elements.

 

I have done some work on our Viewer and I can now execute this test

successfully and get  the CGM via OBJECT DATA instead of PARAM SRC.

 

I now see the problem that Benoit encountered, corruption of the the SRC

PARAM in the DoPropExchange.  It's interesting to note however that

other PARAM'S are okay, HALIGN and VALIGN for example which is good.

Only SRC and another private keyword "CgmPath" we define for setting

a local CGM file path.

 

Regards,
Don Larson
Larson Software Technology

 >
 >  Regards,
 >  -Lofton.
 >