Browse Prior Art Database

Method for Determining Error Source in a Large OS/2 Application

IP.com Disclosure Number: IPCOM000107509D
Original Publication Date: 1992-Mar-01
Included in the Prior Art Database: 2005-Mar-21
Document File: 2 page(s) / 58K

Publishing Venue

IBM

Related People

Johnson, SH: AUTHOR

Abstract

Disclosed is a method for determining the source of a known error situation when multiple error traps are included in code being developed. Modern optimizing compilers often group calls to subroutines into one area, making it difficult to determine exactly where the subroutine was called from the mainline. (Image Omitted)

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 96% of the total text.

Method for Determining Error Source in a Large OS/2 Application

       Disclosed is a method for determining the source of a
known error situation when multiple error traps are included in code
being developed.  Modern optimizing compilers often group calls to
subroutines into one area, making it difficult to determine exactly
where the subroutine was called from the mainline.

                            (Image Omitted)

      When debugging a large application, it usually becomes
necessary to include breakpoints in the code that cause the
application to stop at a predetermined point. Unfortunately, calls to
the breakpoint subroutine are usually optimized to a single place,
making it difficult to determine the location from which the
breakpoint was called. This method provides a simple way to make that
determination.

      The tagging of each breakpoint location takes a two-level
approach.  First, each source file in the application is given a
unique index in the range OxOO to OxFF.  Second, each breakpoint
within each source file is given a unique index in the same range.
The combination of these two indices at the time of the breakpoint
gives a unique error index that can be easily located within the
application being debugged.

      Specifically, the method being described uses two of the 80386
registers to indicate the error index and the location of the
application's global data area.  The accompanying technique shows h...