This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

[PATCH] Remove warning cleanups


Hi.

Here are two more patches which get rid of warnings seen in
the latest snapshot.

The first is for 'cccp.c' - the warnings are comparing signed
and unsigned values. To my (untrained) eye, the check performed
by the code seems silly, so removing the check removes the warnings.
Maybe I'm missing something obvious...
==========================================================
--- egcs-19990405/gcc/cccp.c.orig	Mon Apr  5 10:34:04 1999
+++ egcs-19990405/gcc/cccp.c	Wed Apr  7 13:26:31 1999
@@ -2234,8 +2234,6 @@
     /* Read a file whose size we can determine in advance.
        For the sake of VMS, st.st_size is just an upper bound.  */
     size_t s = (size_t) st.st_size;
-    if (s != st.st_size || s + 2 < s)
-      memory_full ();
     fp->buf = (U_CHAR *) xmalloc (s + 2);
     fp->length = safe_read (f, (char *) fp->buf, s);
     if (fp->length < 0) goto perror;
@@ -5327,8 +5325,6 @@
 
   if (S_ISREG (inc->st.st_mode)) {
     size_t s = (size_t) inc->st.st_size;
-    if (s != inc->st.st_size || s + 2 < s)
-      memory_full ();
     fp->buf = (U_CHAR *) xmalloc (s + 2);
     fp->bufp = fp->buf;
 
@@ -5450,8 +5446,6 @@
   if (S_ISREG (st->st_mode))
     {
       size_t s = (size_t) st->st_size;
-      if (s != st->st_size || s + 2 < s)
-	memory_full ();
       buf = xmalloc (s + 2);
       length = safe_read (pcf, buf, s);
       if (length < 0)
==============================================================

The second patch is for 'final.c' - on i386, the ASM_OPEN_PAREN
and ASM_CLOSE_PAREN are "", while every other platform I could
find they are "(" and ")". Assuming the macro as it stands
is correct, this fix gets rid of a zero-length format string
warning.

==============================================================
--- egcs-19990405/gcc/final.c.orig	Mon Apr  5 10:34:43 1999
+++ egcs-19990405/gcc/final.c	Wed Apr  7 15:12:41 1999
@@ -3653,9 +3653,9 @@
       if (GET_CODE (XEXP (x, 1)) == CONST_INT
 	  && INTVAL (XEXP (x, 1)) < 0)
 	{
-	  fprintf (file, ASM_OPEN_PAREN);
+	  fprintf (file, "%s", ASM_OPEN_PAREN);
 	  output_addr_const (file, XEXP (x, 1));
-	  fprintf (file, ASM_CLOSE_PAREN);
+	  fprintf (file, "%s", ASM_CLOSE_PAREN);
 	}
       else
 	output_addr_const (file, XEXP (x, 1));
==============================================================

I rebuilt the latest snapshot with these patches - no regressions
on i586-pc-linux-gnu.

Do patches this small require a ChangeLog entry? I'll resubmit
with one if desired.

Enjoy.

Art Haas


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]