Re: Small fix for inv_getsize

3 Ноябрь 2000 01:19, Tom Lane написал:
> Denis Perchine <dyp(at)perchine(dot)com> writes:
> > you wrote. Except VARATT_IS_EXTENDED check (is it neccessary, can I store
> > data and be sure that it is not toasted? I do not like this for BLOBs).
>
> Yes, it's necessary *and* appropriate. LO data won't be moved off,
> because pg_largeobject doesn't have a toast table (unless the user makes
> one), but it can be compressed.
>
> > All other seems the same... Please give me an example of this check...
>
> The loop only has to loop till it finds a valid tuple.

But my code do exactly this...
The only difference between your code and mine (except you search through all
tuples) is in toast handling... Isn't (tuple.t_data == NULL) a validity check?

My code:
while ((indexRes = index_getnext(sd, ForwardScanDirection))) {
tuple.t_self = indexRes->heap_iptr;
heap_fetch(obj_desc->heap_r, SnapshotNow, &tuple, &buffer);
pfree(indexRes);
if (tuple.t_data == NULL)
continue;
found++;
data = (Form_pg_largeobject) GETSTRUCT(&tuple);
lastbyte = data->pageno * IBLKSIZE +
getbytealen(&(data->data));
ReleaseBuffer(buffer);
break;
}

Your code:
while ((indexRes = index_getnext(sd, ForwardScanDirection)))
{
tuple.t_self = indexRes->heap_iptr;
heap_fetch(obj_desc->heap_r, SnapshotNow, &tuple, &buffer);
pfree(indexRes);
if (tuple.t_data == NULL)
continue;
found = true;
data = (Form_pg_largeobject) GETSTRUCT(&tuple);
datafield = &(data->data);
pfreeit = false;
if (VARATT_IS_EXTENDED(datafield))
{
datafield = (bytea *)
heap_tuple_untoast_attr((varattrib *)
datafield);
pfreeit = true;
}
thislastbyte = data->pageno * LOBLKSIZE +
getbytealen(datafield);
if (thislastbyte > lastbyte)
lastbyte = thislastbyte;
if (pfreeit)
pfree(datafield);
ReleaseBuffer(buffer);
}

--
Sincerely Yours,
Denis Perchine

----------------------------------
E-Mail: dyp(at)perchine(dot)com
HomePage: http://www.perchine.com/dyp/
FidoNet: 2:5000/120.5
----------------------------------