Bug in plugin/object/obj40/obj40_repair.c

Bug in plugin/object/obj40/obj40_repair.c

[David T Hollis]

I was having trouble compiling reiser4progs due to warnings being 
treated as errors and came across this function that GCC reported 'info 
possibly being used uninitialized'.  In looking at it, it does look 
bad.  Info is never assigned to anything before it is being accessed. I 
don't know what it is supposed to point to so I can't provide a patch at 
this time, but I thought I'd bring it up.


/* Checks that @obj->info.start is SD of the wanted file.  */
errno_t obj40_stat(obj40_t *obj, stat_func_t stat_func) {
        object_info_t *info;
        errno_t res;
                                                                               
 
        aal_assert("vpf-1200", obj != NULL);
                                                                               
 
        if (!info->start.plug) {
^^^^^^^^^^^^^^^^^^^^^
                if (!obj->core->tree_ops.valid(info->tree, &info->start))
                        return RE_FATAL;
                                                                               
 

Re: Bug in plugin/object/obj40/obj40_repair.c

[Vitaly Fertman]

Hi David,

On Thursday 28 August 2003 15:27, David T Hollis wrote:
> I was having trouble compiling reiser4progs due to warnings being
> treated as errors and came across this function that GCC reported 'info
> possibly being used uninitialized'.  In looking at it, it does look
> bad.  Info is never assigned to anything before it is being accessed. I
> don't know what it is supposed to point to so I can't provide a patch at
> this time, but I thought I'd bring it up.
>
>
> /* Checks that @obj->info.start is SD of the wanted file.  */
> errno_t obj40_stat(obj40_t *obj, stat_func_t stat_func) {
>         object_info_t *info;
>         errno_t res;
>
>
>         aal_assert("vpf-1200", obj != NULL);
>
>
>         if (!info->start.plug) {
> ^^^^^^^^^^^^^^^^^^^^^
>                 if (!obj->core->tree_ops.valid(info->tree, &info->start))
>                         return RE_FATAL;

thank you for the report, I am fixing it now.

-- 
Thanks,
Vitaly Fertman