Arnaud and Kirk, thanks for your answers. This: > I began using "all variables are typed" several years ago and the overall > quality of my databases (though not necessarily individual methods) > improved greatly.
and this > And it can be (I stress 'can be') convenient to make Compiler_ methods for > specific modules or areas of the database, eg. Compiler_scheduling, > Compiler_vendors, Complier_web. ... is that we do since years and it helped and are helping still to get a good code quality. But therefore I'm a bit surprised that we are not able to detect such a simple array dimensions problem before runtime. And this error message isn't very specific: "Argument types are incompatible." Yes, it isn't wrong, but ... The error is easily detectable as you can see if you switch the compiler settings. Maybe we have to adapt our guidelines for creating new builds of our software: Do at least two compiler passes manually, one with setting "all variables are typed" and one with "type the variables". Hmmm ... @Arnaud: Nice method you provide. Especially as you already said for resizing of arrays. I like assertions, it really helps to avoid errors made by the developers. But in our case it was not a "real" resize of that array. The declaration in the Compiler_ method was done to fulfill the requirements for compiling in mode "all variables are typed". The second call in code is an one-time initialization of the array, it serves as a constant array and is not changed again. Now I'm thinking about strategies to avoid the reported problem. One could be the following: Since we use this kind of arrays as constant arrays it could be possible to put the declaration and the initialization in one place and not in two different methods. And because we use the compiler setting "all variables are typed", this place could be the Compiler_ method. Is this a good idea? Any drawbacks? Thanks again Lutz ********************************************************************** 4D Internet Users Group (4D iNUG) FAQ: http://lists.4d.com/faqnug.html Archive: http://lists.4d.com/archives.html Options: http://lists.4d.com/mailman/options/4d_tech Unsub: mailto:[email protected] **********************************************************************
