Cached Code? (OpenInsight 32-bit Specific)
At 13 MAR 2002 01:58:46PM David Kafka wrote:
Since I am working to convert an ARev app, I started by bringing some of my ARev tables into OI. It?s a long story, but I have been trying to maintain compatibility with some 1.16 deployments so the ARev tables still have tons of dots. I knew OI wouldn?t tolerate this, so I upgraded the tables and ran a program I wrote years ago to intelligently convert dots to underscores in the code, before adding the tables to OI. Unfortunately, there are still some stray dots left hanging around in the code for symbolics and procedures.
I copied a couple of my routines from BP to SYSPROCS and compiled them using the System Editor.
I created a form that included a symbolic that refers to a function which does XLATEs. The function still had dots for some table and row names. The debugger told me that it could not load ?DOC.DATE? from the dictionary. Fine, I went back into SYSPROCS and repaired the references in the function. But, then, no matter what I did, it continued to give me the same error, even though DOC.DATE no longer existed in the source. The source displayed by the debugger had DOC_DATE, but it was still complaining about DOC.DATE. (Similar to when ARev buffers object code after a program has been recompiled.) I could not get that to go away no matter what I did.
I finally gave up and created a new form from scratch, and the problem went away. Any ideas why/
David
dkafka@supmat.com
At 13 MAR 2002 02:18PM [url=http://www.sprezzatura.com" onMouseOver=window.status=Click here to visit our web site?';return(true)]The Sprezzatura Group[/url] wrote:
Could speculate but sometimes it's better to just forget it! This MIGHT be dict.mfs tryign to be kind and convert . to _ for you - ultimately confusing all parties!
World Leaders in all things RevSoft
