From: simon <simon@cda61777-01e9-0310-a592-d414129be87e>
Date: Thu, 19 Jul 2012 04:29:50 +0000 (+0000)
Subject: Rework the new type-check in sresize so that it doesn't cause a
X-Git-Url: https://git.distorted.org.uk/~mdw/sgt/putty/commitdiff_plain/f0f0ae7f27c50b0bc586d360da17f3cb05e9c467

Rework the new type-check in sresize so that it doesn't cause a
compile warning ('left-hand operand of comma expression has no
effect'), which of course becomes fatal under -Werror.

(This would have been instantly noticeable to people compiling with
the old-fashioned Makefile.gtk, which does include -Wall -Werror, but
those of us using the new autoconf makefile hadn't noticed.)


git-svn-id: svn://svn.tartarus.org/sgt/putty@9582 cda61777-01e9-0310-a592-d414129be87e
---

diff --git a/puttymem.h b/puttymem.h
index 6ef3f794..941aded3 100644
--- a/puttymem.h
+++ b/puttymem.h
@@ -34,10 +34,19 @@ void safefree(void *);
  * possible, in favour of these type-casting macros which ensure
  * you don't mistakenly allocate enough space for one sort of
  * structure and assign it to a different sort of pointer.
+ *
+ * The nasty trick in sresize with sizeof arranges for the compiler,
+ * in passing, to type-check the expression ((type *)0 == (ptr)), i.e.
+ * to type-check that the input pointer is a pointer to the correct
+ * type. The construction sizeof(stuff) ? (b) : (b) looks like a
+ * violation of the first principle of safe macros, but in fact it's
+ * OK - although it _expands_ the macro parameter more than once, it
+ * only _evaluates_ it once, so it's still side-effect safe.
  */
 #define snew(type) ((type *)snmalloc(1, sizeof(type)))
 #define snewn(n, type) ((type *)snmalloc((n), sizeof(type)))
-#define sresize(ptr, n, type) \
-    ((type *)snrealloc((sizeof((type *)0 == (ptr)), (ptr)), (n), sizeof(type)))
+#define sresize(ptr, n, type)                                           \
+    ((type *)snrealloc(sizeof((type *)0 == (ptr)) ? (ptr) : (ptr),      \
+                       (n), sizeof(type)))
 
 #endif