Subject: Forward: Bug#181228: wrong assumption on char signedness
From: Junichi Uekawa (dancer_AT_netfort.gr.jp)
Date: Mon Feb 17 2003 - 03:05:15 EET
Hi,
I've received this message through Debian Bug Tracking System.
There is apparently a char where there should be an int.
attached mail follows:
Package: ecasound2.2
Version: 2.2.0-rel-4
Severity: important
Hi,
char is per default unsigned on arm, powerpc and s390. There is a warning
ecalength.c:99: warning: comparison is always true due to limited range of
data type
during build of your package that might cause a runtime error. Please check
for details.
In your case the variable curopt should be defined as int. Defining it as
char is even wrong on i386, since the char 0xff is then interpreted as EOF.
On the platforms mentioned above the EOF condition is never true, leading to
an infinite loop.
Thanks,
Gerhard
This archive was generated by hypermail 2b28 : Mon Feb 17 2003 - 02:56:29 EET