Re: Pseudo Code or Compiled

johnl@informix.com (Jonathan Leffler) wrote: >}From: "Billy Wheeler" <billy.wheeler@pixie.co.za> >} >}Perhaps Jonathan Leffler or someone could tell us if there is any way >}of switching off the function resolution on loading. Failing that, what say you >}all about putting it in as a feature request? >Although the p-code interpreter does check that there is a definition for >every function which might be called, it does not fix up the p-code >references to any function until the code is executed. Before it starts >executing, FGLGO (or FGLDB) should definitely check that all the functions >it might need are available to it. People get angry if they've been >entering data for half-an-hour and the program suddenly stops and says "Hi, >I can't go on -- you've called a non-existent function!". The reason that they would get angry is that the program fell over - the cause is academic. They would, quite rightly, be upset with the quality control that let the program through testing. How does this situation differ from faulty SELECT statements, or functions called with the incorrect number of parameters ? And, furthermore, if the generation of the 4gi did the check then there would not be any necessity to check at EVERY run time. (Even setting one BIT in the 4gi to say that function resolution has been done would be sufficient!). >You can decide to make it a feature request. I don't think anything >will happen, though. This is Informix after all. :-) Cheers, Brent Jackson. brent@advgroup.co.nz Phone : 64-9-3610613 Fax : 64-9-3764210