Experimental Physics and
Industrial Control System

declares 'isnan' and 'isinf' as external functions but C99 demands
that they are macros and newlib-1.15 defines them as such.

Including RTEMS' epicsMath.h from a C++ or C99 source file
causes a compilation error.

In fact, using 'isnan/isinf' from code that claims to be ANSI compliant
(EPICS base is usually compiled (gcc) with -ansi) is wrong. ANSI-C code
should not assume these functions to be present.

*** src/libCom/osi/os/RTEMS/epicsMath.h.orig	2007-09-18 12:38:39.000000000 -0700
--- src/libCom/osi/os/RTEMS/epicsMath.h	2007-09-18 12:46:28.000000000 -0700
***************
*** 16,24 ****
  extern "C" {
  #endif
  
  int isnan(double);
- int finite(double);
  int isinf(double);
  
  #ifdef __cplusplus
  }
--- 16,29 ----
  extern "C" {
  #endif
  
+ /* T.S.; C99 demands that these are macros and newlib-1.15 defines
+  * them as such!
+  */
+ #if defined(STRICT_ANSI) && !defined(__cplusplus)
  int isnan(double);
  int isinf(double);
+ #endif
+ int finite(double);
  
  #ifdef __cplusplus
  }

Replies:

Re: ANSI source including epicsMath.h won't compile with newlib 1.15 (RTEMS) Eric Norum

Navigate by Date:

Prev: RE: Deadlock on epicsEvent Jeff Hill

Next: Re: ANSI source including epicsMath.h won't compile with newlib 1.15 (RTEMS) Eric Norum

Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  <2007>  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 

Navigate by Thread:

Prev: RE: Beginer problem Paul Joireman

Next: Re: ANSI source including epicsMath.h won't compile with newlib 1.15 (RTEMS) Eric Norum

Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  <2007>  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024