[PATCH 0/1] MR2230: webservices: Avoid uninitialized variable warning in format_double

[Gerald Pfeifer (＠gerald)]

GCC 13 was going to diagnose
dlls/webservices/writer.c:960:18: warning: 'mag2' may be used uninitialized

[PATCH 1/1] webservices: Avoid uninitialized variable warning in format_double

[Gerald Pfeifer]

From: Gerald Pfeifer

GCC 13 was going to diagnose
dlls/webservices/writer.c:960:18: warning: 'mag2' may be used uninitialized
---
 dlls/webservices/writer.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dlls/webservices/writer.c b/dlls/webservices/writer.c
index a4cfc94f08e..6923035f00c 100644
--- a/dlls/webservices/writer.c
+++ b/dlls/webservices/writer.c
@@ -896,7 +896,7 @@ static ULONG format_double( const double *ptr, unsigned char *buf )
     static const double precision = 0.0000000000000001;
     unsigned char *p = buf;
     double val = *ptr;
-    int neg, mag, mag2, use_exp;
+    int neg, mag, mag2 = 0, use_exp;

if (isnan( val ))
     {

[Hans Leidekker (＠hans)]

It's a false positive. GCC doesn't see that mag2 is only used when use_exp evaluates to true, in which case it has been set earlier.

[Gerald Pfeifer (＠gerald)]

We generally do silence warnings like this (by initializing the variable), So that's what I did here, too. Luckily it's about the only one left.

[Hans Leidekker (＠hans)]

Sure but wouldn't it be better to wait with such fixes until GCC 13 is released?

[Gerald Pfeifer (＠gerald)]

GCC 13 has been released two weeks ago and this warning is still there.

Are you good to approve this now, Hans?

[PATCH 0/1] MR2230: webservices: Avoid uninitialized variable warning in format_double - approved

[Hans Leidekker (＠hans)]

This merge request was approved by Hans Leidekker.