Re: [REVIEW] pg_last_xact_insert_timestamp

On Thu, Sep 29, 2011 at 5:20 PM, Kyotaro HORIGUCHI
<horiguchi(dot)kyotaro(at)oss(dot)ntt(dot)co(dot)jp> wrote:
> Sorry for late to re-review.

Thanks!

>> > Nevertheless this is ok for all OSs, I don't know whether
>> > initializing TimestampTz(double, int64 is ok) field with 8 bytes
>> > zeros is OK or not, for all platforms. (It is ok for
>> > IEEE754-binary64).
>>
>> Which code are you concerned about?
>
> xlog.c: 5889
>>       beentry = pgstat_fetch_all_beentry();
>>
>>       for (i = 0; i < MaxBackends; i++, beentry++)
>>       {
>>               xtime = beentry->st_xact_end_timestamp;
>
>  I think the last line in quoted code above reads possibly
> zero-initialized double (or int64) value, then the doubted will
> be compared and copied to another double.
>
>>               if (result < xtime)
>>                       result = xtime;

I believe it's safe. Such a code is placed elsewhere in the source, too.
If it's unsafe, we should have received lots of bug reports related
to that. But we've not.

Regards,

--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center