RDROYA01@ULKYVX.BITNET (Robert Royar) (01/08/87)
I'm using the Alcyon compiler v 4.14 for this. I've been working on a cc program that compiles and links multiple files. I have it working fine using a startup file to get all its parameters. I run the compiler using Pexec() testing for a 0 return which means no error. Unfortunately, Alcyon's compiler always returns 0, even when there have been multiple errors in a compilation phase. I read somewhere in this group that version 4.14 of the compiler now passed a return code as it exitted, and that the code would tell whether there had been any errors. But this can't be true can it? How does the Make program distributed this summer tell whether or not to continue compilation? I looked at its Pexec() code and it's identical to the code I have. The only thing I can figure to do is to catch bios conout calls and test them for error markers (such as # for cp68 and * for c068) but then I might also abort on warnings. Any ideas? On another note. Earlier I wrote about problems with wildcard expansion. Well, it turns out that all of the devpack tools have the same problem. For example logged into drive C: with AR68.PRG on C:\BIN and a bunch of *.o files on drive E:\, try to create the library m:\lib.arc with the following command \bin\ar68 r m:\lib.arc e:\*.o a few whirring sounds and back to the command prompt or desktop. The brain-damaged .___open code strikes again. Can you say "class-action suit?" Seriously, I can't believe anyone would release tools as damaged as this and not bother even to check if they work, and remember a good deal of the ST OS was written with them, so some of the errors are bound to be embedded deeply in the system.