>>>I suppose that a correct fix involves doing MAXALIGN(VARDATA(evec)),
>>>but I do not know what places need to change to support this.
>
>
>>Its only union and picksplit user-defined methods in contrib modules.
>
>
> If I recall correctly, we decided to go with the present hack because we
> found the problem just before a release date and there wasn't time to do
> it more cleanly. It seems to me that there is time to fix it right for
> 7.5 ...
Yes, you are right.
I suggest to replace bytea by struct
typedef struct {int32 n; /* number of GISTENTRY */GISTENTRY vector[1];
} GistEntryVector;
#define GEVHDRSZ (MAXALIGN(sizeof(int32))
so, allocation will be:
evec = palloc( GEVHDRSZ + sizeof(GISTENTRY)*n );
MAXALIGN guarantee that allocated memory will be no less than required (it may
be greater for 4 bytes).
And change interface to user defined structures from
Datum union(bytea *entryvec, int *size)
Datum picksplit(bytea *entryvec, GIST_SPLITVEC *v)
to
Datum union(GistEntryVector *entryvec, int *size)
Datum picksplit(GistEntryVector *entryvec, GIST_SPLITVEC *v)
In this function it's need to use entryvec->n and entryvec->vector
We can do even
Datum union(int32 n, GISTENTRY *entryvec, int *size)
Datum picksplit(int32 n, GISTENTRY *entryvec, GIST_SPLITVEC *v)
It seems to me that first case is clearer. Of course, I change all contrib
modules to new interface.
What do you think?
--
Teodor Sigaev E-mail: teodor@sigaev.ru