Hello Michael, On Sun, Sep 23, 2012 at 12:41:02AM +0100, Michael Tautschnig wrote: > Package: asclock > Version: 2.0.12-23 > > While building your package using our research compiler infrastructure we > noticed the following conflicting declarations: > > ./asclock.h:extern char exec_str[MAX_PATH_LEN]; > ./asclock.c:char exec_str[] = "echo no program has been specified > >/dev/console";
In wich sense are they conflicting? I just checked:
asclock.c:
==========
...
#include "asclock.h"
--> extern char exec_str[MAX_PATH_LEN];
--> Variable is defined with maximum length (see below for reason)
…
char exec_str[] = "echo no program has been specified >/dev/console";
--> Variable is initialized to this value (see below for a reason)
int main(int argc,char *argv[])
…
parseArgs(argc, argv);
--> strcpy(exec_str, argv[i]);
strcat(exec_str, " &");
--> If option "-e" is provided, this is read as it is from the
command line, the program specified could be fully qualified,
hence it could be up to max path length long.
…
system(exec_str);
At this point, either the default string (no program) or the one
provided on the command line are used.
> This may cause undefined behaviour (likely a segfault) if any part of the
> program relies on the fact that exec_str is of size MAX_PATH_LEN, while only
> 50
> bytes are actually allocated.
I don't see this, see above. Further the only usage of this variable
is for calling "system", i.e. no further internal usage.
Hence there is no bug.
Greetings
Helge
--
Dr. Helge Kreutzmann deb...@helgefjell.de
Dipl.-Phys. http://www.helgefjell.de/debian.php
64bit GNU powered gpg signed mail preferred
Help keep free software "libre": http://www.ffii.de/
signature.asc
Description: Digital signature
