Date:	Fri, 6 Sep 1996 13:04:31 GMT
From:	"<\"R. Steven Regan, Hydrologist, Reston, VA \""<rsregan AT usgs DOT gov>
Message-Id:	<199609061304.NAA12912@hassrvares.er.usgs.gov>
To:	djgpp AT delorie DOT com
Subject:	version 2 problem
Cc:	"David L Parkhurst, Hydrologist, Denver, CO "<dlpark AT usgs DOT gov>

Hello again, I noticed there were two copies of math.h (one in include and the
other in include/libm).  I mv'd the one in libm into include and recompiled and
was able to get the program working.  However, the results of the model are
different under version 2 than under version 1 for some of the test cases.  This
is very disturbing.  Also, I get a message "Extended error 183" occassionally.
By this I mean, I will run the program on one test and get the message, then
run it on another test and not get the message, then run the first test again
and not get the message OR I could run one test and not get the message then
run it again and get the message.  Note, I am using a Gateway 2000 P5-120 under
Windows 95, using a DOS window to run the geochemical model.

I noticed there are other duplicate header file names in the include directory
structure.

file.h  libc/file.h    sys/file.h
pc.h    inlines/pc.h
dir.h   sys/dir.h
errno.h sys/errno.h
time.h  sys/time.h

Apparently, the correct math.h was not getting included, so I am wondering if
the correct versions of other header files are getting included.  I assume
I should put back the original math.h, but how do I make sure the math.h in
libm is used?



**************************************************************************
  R. Steven Regan  rsregan AT usgs DOT gov  phone 703-648-5896  fax 703-648-6802
    USGS, 12201 Sunrise Valley Dr, MS 437, Rm 5B406, Reston, VA  22092
        workstation: hassrvares.er.usgs.gov internet: 130.11.51.203
**************************************************************************

- Raw text -