Hi Dave, Yeah, something seems squirrely with the v16 compiler. I just posted about an error on a component which compiles fine in 15r5, but now generates a "4d method and variable with same name" error (and of course you can't find out the name of said conflict, which btw does not exist in reviewing all code and variable names). Oh and generating typing in this component crashes 4D
Not nice On Mon, May 15, 2017 at 6:22 PM, David Adams via 4D_Tech < 4d_tech@lists.4d.com> wrote: > For those of you following my tedious reports on components, the latest > chapter! > > I've consolidated my two components into one because, well, it's a whole > lot easier that way. I'm putting that into a host database where it makes > life simpler. Instead of my zillions of methods, you see just a few. > > But then I compile. And I get this sort of error: > > The argument list for the method named 'ValueCheck_LongintRange' > in the component "MessageHub" is incompatible with the database > "MessageHubDemo.4DB" > > Hmm. No it isn't. I got this error from a few methods, seemingly randomly. > I tried: > > * Rebuilding all of the constants in both systems. > > * Double-checking all calls. > > * Various other bits of voodoo. (Retokenize all methods, compile with > different options, throw salt over my left shoulder, light a candle to Alan > Turing, etc.) > > * All with constant quit-reopen-recompile-build-quit cycles to be sure > nothing weird was hanging around. > > No dice. Finally, I tried generating a typing file over in the source of > the component and...there's an extra parameter. I'm not using that > parameter, I'm not declaring that parameter (with all vars typed)...but the > Compiler thinks it exists. No idea why. I've declared the phantom parameter > and now things work. Weird. > > A note: You can also get this error for good reason. Like, for example, > you've got a bad parameter list in your host! In other words, this is a > useful error that may be 100% accurate. But when it isn't look for the > weird thing that I noticed. > > P.S. It would make more sense for the compiler to detect this when > compiling the host rather than having it only appear when you try to open > the compiled host. > > P.P.S. It would be better if the compiler would detect bad argument lists > (too long, for example) all of the time. This feature request still needs > more love: > > Compiler/Typing Improvements: Detect some runtime errors in advance > http://forums.4d.fr/Post/EN/19107688/1/19107689 > > Even if you don't care, vote for it for the sake of your brothers and > sisters who do. We'll owe you one! > > P.P.P.S. Is it just me, or is there something going on with the compiler? > I'm in 16.1 and I keep noticing errors that the compiler isn't finding. > Hard to reproduce, but I'm noticing it a lot. The compiler has gotten > stricter in recent versions (good!), so I hope I'm imagining something > here. > ********************************************************************** > 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:4d_tech-unsubscr...@lists.4d.com > ********************************************************************** -- Bob Finnerty Exodus Software ********************************************************************** 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:4d_tech-unsubscr...@lists.4d.com **********************************************************************
