This is the mail archive of the egcs@egcs.cygnus.com mailing list for the EGCS project. See the EGCS home page for more information.

[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index] [Subject Index] [Author Index] [Thread Index]

Re: partly bracketed initializer ?

> What is a partly bracketed initializer warning?

The allocator template uses a pthread_mutex_t object
__node_allocator_lock. On 2.5.1, this is defined as

typedef struct  _pthread_mutex {
        struct {
                uint8_t         pthread_mutex_flag[4];
                uint32_t        pthread_mutex_type;
        } pthread_mutex_flags;
        union {
                struct {
                        uint8_t pthread_mutex_pad[8];
                } pthread_mutex_lock64;
                uint64_t pthread_mutex_owner64;
        } pthread_mutex_lock;
        uint64_t pthread_mutex_data;
} pthread_mutex_t;

Sun uses the construct {0,0,0} to initialize such a thing; this
doesn't reflect the data structure very well. The correct
initialization is {{{0,0,0,0}, 0}, {pthread_mutex_owner64:0}, 0}

> Should I be concerned?

Somewhat. For the current initialization, I get (in a reduced example)

b:
        .byte   0
        .byte   0
        .byte   0
        .skip 1
        .skip 4
        .skip 16

whereas for the complete initialization, I get

b:
        .byte   0
        .byte   0
        .byte   0
        .byte   0
        .uaword 0
        .uaword 0
        .uaword 0
        .uaword 0
        .uaword 0

So the Sun code only partially initializes the object.

Since we are talking about a static member, it is zero-initialized
anyway, and therefore it happens to work. I'm not sure about the
constraints on mutex_t objects - if you can make them automatic
variables, you might run into problems.

> Is the a -Wno-bracketed-initializer to turn them off if I'm 
> not concerned?

Yes, it's called -Wno-missing-braces.

Regards,
Martin