Re: Weird compiler problem with g77

[Joseph Mildenberger <Joe dot Mildenberger at triumf dot ca>]

Hello,
    I have put together a self-contained "package" which demonstrates
the bug described below. Just to give you some background: I work at
TRIUMF, a particle physics lab in Vancouver. The code is a fairly large
data analysis program that reads data records from a cernlib-produced
ntuple file, does various hocus-pocus stuff to the data, and then writes
out another ntuple file. There are a lot of different routines (as you
will see...), and a whole raft of necessary calibration files, included
in a subdirectory. It is all contained in a gzipped tar file, which I
hesitate to include as an attachment, since it is 8.5 MB (32 MB
unzipped). I will await your suggestion as to how to send it to you.
     As I said, I have done my best to make it "self-contained", and as
such, the largest components of the archive (about 18 MB) are 2 cernlib
object libraries. Almost every item necessary to make the executable is
contained in the archive, along with a sample input file and run script.
The only thing _not_ there are 3 libraries which I assume are standard
system stuff (I'm a physicist, not a programmer...), designated in the
Makefile by -lcrypt , -ldl , and -lnsl . I will send instructions on
compiling and running the code when I hear back from someone about
sending the tar file. Please let me know if you have any more questions,
or if there is something else I can do to make this simpler for you. I
appreciate your assistance.
                                                 Sincerely,
                                                    Joe Mildenberger,
TRIUMF

Toon Moene wrote:
> 
> Joseph Mildenberger wrote:
> 
> > Can anyone suggest ideas on how to track down the root of this problem?
> 
> Well, you *could* sent us a complete, compilable and runnable example of
> the code showing this problem (gcc-bugs@gcc.gnu.org) and we will point
> out your error :-)
> 
> [ This is only slightly facetious:  The last time we were able to track
> a
>   NaN to a compiler error in a released version of g77 is about 5 years
>   ago - but, agreed, it could happen again .... ]
> 
> --
> Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
> Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
> Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
> Join GNU Fortran 95: http://g95.sourceforge.net/ (under construction)