home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.aix
- Path: sparky!uunet!newsgate.watson.ibm.com!yktnews!admin!yktnews!prener
- From: prener@watson.ibm.com (Dan Prener)
- Subject: Re: xlf -qextchk
- Sender: news@watson.ibm.com (NNTP News Poster)
- Message-ID: <PRENER.92Jul27025457@prener.watson.ibm.com>
- In-Reply-To: ron@sarah.lerc.nasa.gov's message of Mon, 27 Jul 1992 04:36:57 GMT
- Date: Mon, 27 Jul 1992 07:54:57 GMT
- Disclaimer: This posting represents the poster's views, not necessarily those of IBM
- References: <1992Jul27.043657.22336@eagle.lerc.nasa.gov>
- Nntp-Posting-Host: prener.watson.ibm.com
- Organization: IBM T.J. Watson Research Center, Hawthorne, New York
- Lines: 56
-
- In article <1992Jul27.043657.22336@eagle.lerc.nasa.gov> ron@sarah.lerc.nasa.gov (Ron Gaug) writes:
-
- >Does anyone know how to get some useful diagnostics from the xlf compiler's
- >extchk option? Here's an excerpt from the xlf manual page:
-
- > extchk Perform procedure interface checking as well as
- > detection of mismatched common blocks.
-
- [lots of stuff deleted]
-
- > [hilbert]:514% xlf -v -qextchk -qsource extchk_bug.f
- > exec: /usr/lpp/xlf/bin/xlfentry(xlfentry,extchk_bug.f,/tmp/F8FNsCkR,extchk_bug.lst,xlfsmsg.cat,xlfmsg.cat,EXTCHK,SOURCE,NULL)
- > ** foo === End of Compilation 1 ===
- > ** xyzzy === End of Compilation 2 ===
- > 1501-510 Compilation successful for file extchk_bug.f.
- > 0706-316 Error: Type mismatches detected. References dumped to load map.
- [more stuff deleted]
-
- >And the references to the type mismatches that were detected were dumped
- >to the load map. Now if I could just get a load map. :-) And where's my a.out?
-
- In order to get the loadmap, add the option
-
- -bloadmap:<filename>
-
- and the loadmap will appear as the file you name in place of "<filename>".
- Well, actually it isn't a loadmap, but rather an inappropriately named
- log of linker actions and messages.
-
- The reason you didn't get an a.out is that there was an error: your
- type mismatch. Now, in case you are wondering about the remaining
- bunch of puzzling details, such as what were all those files being
- unlinked, and how would a log of linker actions and messages help if
- it never invoked ld, here is the crucial fact. When xlf is given a
- file containing more than a single program or subprogram, it compiles
- them into multiple files, the strangely named files that were
- unlinked, in your posting, and then quietly feeds that collection of
- files to ld to do a partial link that will produce a single .o file.
- It was that hidden ld step that failed, causing the real ld, which
- would link in the libraries and produce an executable, never to be
- executed. It was also the hidden ld whose "loadmap" would contain
- the messages you were looking for.
-
- This explanation is given only to explain the details. You should not
- have to concern yourself with the distinction between the hidden
- invocation of ld, and the usual one, e.g., when multiple files are
- compiled. In both cases the messages will be found in the file you
- specify with the
-
- -bloadmap:<filename>
-
- option.
-
-
- --
- Dan Prener (prener@watson.ibm.com)
-