Re: pgstattuple extension for indexes

Bruce Momjian <bruce(at)momjian(dot)us> writes:
> Patch applied. Thanks.

>> Also, I added an experimental feature for btree indexes. It checks
>> fragmentation factor of indexes. If an leaf has the right link on the next
>> adjacent page in the file, it is assumed to be continuous (not fragmented).
>> It will help us to decide when to REINDEX.

This was done in an entirely unacceptable way, to wit

ereport(NOTICE,
(errmsg("%.2f%% fragmented",
100.0 * (stat.forward + stat.backward) /
(stat.continuous + stat.forward + stat.backward)),
errhint("continuous=%llu, forward=%llu, backward=%llu",
stat.continuous, stat.forward, stat.backward)));

The really serious problem with reporting this info via NOTICE is that
there's no way for a program to get its hands on the info. The output
tuple format needs to be extended instead.

The lesser problem that drew my attention is that %llu is unportable
(and in fact draws gcc warnings for me; did you ignore that?). But
using UINT64_FORMAT instead would create a headache for translation
because the string would vary across platforms.

I won't bother correcting the violations of message style guidelines,
because this code simply has to go away.

regards, tom lane