Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!lpi.liant.com!rcg From: rcg@lpi.liant.com (Rick Gorton) Newsgroups: comp.lang.fortran Subject: Re: xlf and RESOLVE on RISC 6000. what is it?!! Keywords: Link-time type checking Message-ID: <1991Jun3.144111.22513@lpi.liant.com> Date: 3 Jun 91 14:41:11 GMT Article-I.D.: lpi.1991Jun3.144111.22513 References: <1991May31.164029.86816@vaxc.cc.monash.edu.au> Organization: Language Processors Inc., Framingham MA Lines: 34 In article <1991May31.164029.86816@vaxc.cc.monash.edu.au> che358w@vaxc.cc.monash.edu.au writes: > > I have compiled a large fortran program with xlf on the RISC 6000. If I >us xlf -cC and then link everything is fine (except the program >doesn't give the right answers). However, if I use xlf -cC -qextchk >and then link I get an error number and then type mis matches >detected by RESOLVE. See error messages under RESOLVE for more information. > >What is RESOLVE ? > >Where can I find documentaion on it ? > >Can I make it give more specific information like where the type mismatch occurs >? > >Thanks in advance for any advice. > >Samantha Lane The -qextchk switch is the one that gets thrown when you want to detect link-time datatype mismatches. Which means to me that look at the interface to the routine RESOLVE (your code) and all the places it is called. Somebody probably called it with one kind of parameter, and it is really a different datatype. It could also be that it is a function, and is returning a datatype that conflicts with what is being specified by the left hand side of the expression: Example: CONJG(RESOLVE()) when RESOLVE returns INTEGER. Hope this helps. -- Richard Gorton rcg@lpi.liant.com (508) 626-0006 Language Processors, Inc. Framingham, MA 01760 USA Hey! This is MY opinion. Opinions have little to do with corporate policy.