Re: New 4GL runtime errors

>From: "Fergus C Hayne" <fergus_hayne@virgin.net> >Date: 15 May 1997 11:42:44 GMT >X-Informix-List-Id: <news.37933> > >I have recently upgraded my 4GL tools from 6.00 to 6.05 and I have had to >amend several of my applications because of runtime errors caused by me >ignoring the results returned from functions. It was a bug that ever allowed this to happen. The bug was fixed. The release notes (which you probably didn't read) documents both the change and the backwards compatability workaround. *** READ THE RELEASE NOTES -- THEY CONTAIN USEFUL INFORMATION *** >Why have Informix changed their compiler - this has really pissed me off. Because there were bugs in it which needed to be fixed. >While I am bitching about the compiler why is status not set in r4gl when a >record is not found during a fetch from a cursor, SQLCA.SQLCODE is set and >when compiled with i4gl status is set. That sounds like a new bug. Basically, any discrepancy in behaviour between p-code and c-code which is not due to using uninitialized variables is a bug in one or the other of the products. That's why the ignoring returned results was a bug -- the p-code system always complained about that, and the c-code system didn't. Yours, Jonathan Leffler (johnl@informix.com) #include <disclaimer.h>