[ET Trac] [Einstein Toolkit] #2106: Make ParseFile.c more robust

[Einstein Toolkit]

#2106: Make ParseFile.c more robust
-----------------------+----------------------------------------------------
  Reporter:  eschnett  |       Owner:                     
      Type:  defect    |      Status:  reopened           
  Priority:  unset     |   Milestone:                     
 Component:  Cactus    |     Version:  development version
Resolution:            |    Keywords:                     
-----------------------+----------------------------------------------------

Comment (by rhaas):

 Uff, I see. "usual arithmetic conversion" is really murky to me. So I try
 to err on the side of not relying on it.

 I think I would agree with this assessment if the conversion happened in
 in the if statement. However the actual code is:
 {{{
 static char *ReadFile(FILE *file, unsigned long *filesize)
 {
 ...
 *filesize = ftell(file);
 if (*filesize < 0)
 {
   // do stuff
 }}}
 so the result of ftell (a long int) is converted to unsigned long int when
 it is assigned to *filesize. Since {{{filesize}}} is known to be "unsigned
 long int" the optimization seems valid to me (namely an unsigned int
 comparison with <0 will always fail).

 The cast of -1 to {{{unsigned long int}}} that I added is just to avoid
 the warning about comparing signed and unsigned values.

 A test code:
 {{{
 void foo()
 {
   unsigned int bar = 1;
   if(bar < 0) {
     baz();
   }

   if(bar == -1) {
     baz();
   }
 }
 }}}
 gives me two warnings on gcc 7.3.0 when using -Wextra (but not eg -Wall):
 {{{
 rhaas at ekohaes8:.../rhaas/tmp$ gcc -Wall -Wextra -c  compare.c
 compare.c: In function ‘foo’:
 compare.c:4:10: warning: comparison of unsigned expression < 0 is always
 false [-Wtype-limits]
    if(bar < 0) {
           ^
 compare.c:8:10: warning: comparison between signed and unsigned integer
 expressions [-Wsign-compare]
    if(bar == -1) {
           ^~
 }}}
 Note the first warning about the comparison always being false.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/2106#comment:9>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit