From: SourceForge.net <no...@so...> - 2010-09-30 17:27:03
|
Bugs item #3078851, was opened at 2010-09-30 18:44 Message generated for change (Comment added) made by jphartmann You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=684730&aid=3078851&group_id=119701 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Interpreter Group: v4.0 >Status: Closed Resolution: Invalid Priority: 5 Private: No Submitted By: John P. Hartmann (jphartmann) Assigned to: Nobody/Anonymous (nobody) Summary: Incorrect result with exposed variables. Initial Comment: On Fedore 13, This is a paraphrase of one of Mike's original test cases in rexdiag exec. If it is any consolation, another of the tests breaks on current CMS. Maybe rexdiag should be rolled into the regression test? The following program says 1, but the correct answer is 6. Looks like the assignment of the default value does not make it to the exposed stem or whichever way round this is handled. x.=1 call x1 say x.1 exit x1: procedure expose x.1 x.=5 x.1=x.1+1 return ---------------------------------------------------------------------- >Comment By: John P. Hartmann (jphartmann) Date: 2010-09-30 19:27 Message: User error. See Rick's comment. ---------------------------------------------------------------------- Comment By: Rick McGuire (bigrixx) Date: 2010-09-30 19:02 Message: This is one fundamental difference between classic Rexx and Object Rexx. In Object Rexx, a compound variable assignment of a default value creates a totally new stem object that is assigned to the stem variable in the local context. This can sever the connection between the variable in the subroutine context and the calling context. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=684730&aid=3078851&group_id=119701 |