Обсуждение: OO and tuples of different types

Came across this comment in execTuples.c...
*In the implementation of nested-dot queries such as*"retrieve (EMP.hobbies.all)", a single scan may return tuples*of
manytypes, so now we return pointers to tuple descriptors*along with tuples returned via the tuple table.  This
means*wenow have a bunch of routines to diddle the slot descriptors*too.  -cim 1/18/90
 

Hmm. Another few years of reading the source and I'll probably
understand all this :-) :-(.

Can anybody provide any insight? Like What EMP.hobbies.all meant in
postquel?

-- 
Chris Bitmead
mailto:chris@bitmead.com

Chris <chris@bitmead.com> writes:
> Can anybody provide any insight? Like What EMP.hobbies.all meant in
> postquel?

Not a lot of insight here, but there are a couple of examples in the
"misc" regress test; also see the discussion in TODO.detail/function.
        regards, tom lane

Chris wrote:
> 
> Came across this comment in execTuples.c...
> 
>  *In the implementation of nested-dot queries such as
>  *"retrieve (EMP.hobbies.all)", a single scan may return tuples
>  *of many types, so now we return pointers to tuple descriptors
>  *along with tuples returned via the tuple table.  This means
>  *we now have a bunch of routines to diddle the slot descriptors
>  *too.  -cim 1/18/90
> 
> Hmm. Another few years of reading the source and I'll probably
> understand all this :-) :-(.
> 
> Can anybody provide any insight? Like What EMP.hobbies.all meant in
> postquel?

I guess you can still pind the original Postgres 4.2 (the last (?) 
one with postquel) somewhere on the net. I used to have them around 
somewhere but I'm not sure where they are ...

---------------
Hannu

Chris wrote:
> 
> Came across this comment in execTuples.c...
> 
>  *In the implementation of nested-dot queries such as
>  *"retrieve (EMP.hobbies.all)", a single scan may return tuples
>  *of many types, so now we return pointers to tuple descriptors
>  *along with tuples returned via the tuple table.  This means
>  *we now have a bunch of routines to diddle the slot descriptors
>  *too.  -cim 1/18/90
> 
> Hmm. Another few years of reading the source and I'll probably
> understand all this :-) :-(.
> 
> Can anybody provide any insight? Like What EMP.hobbies.all meant in
> postquel?

You can get the latest sources for postgres 4.2 (from 30.06.1994) 
from http://db.cs.berkeley.edu/postgres.html 

IIRC they have a small manual inside the tarball

They should even compile ;), I remember having compiled them (on the 
then-current version of solaris) and played around a little with postquel

----------------
Hannu

On Fri, Jan 28, 2000 at 09:03:34PM +0200, Hannu Krosing wrote:
> Chris wrote:
> > 
> > Came across this comment in execTuples.c...
> > 
> >  *In the implementation of nested-dot queries such as
> >  *"retrieve (EMP.hobbies.all)", a single scan may return tuples
> >  *of many types, so now we return pointers to tuple descriptors
> >  *along with tuples returned via the tuple table.  This means
> >  *we now have a bunch of routines to diddle the slot descriptors
> >  *too.  -cim 1/18/90
> > 
> > Hmm. Another few years of reading the source and I'll probably
> > understand all this :-) :-(.
> > 
> > Can anybody provide any insight? Like What EMP.hobbies.all meant in
> > postquel?
> 
> You can get the latest sources for postgres 4.2 (from 30.06.1994) 
> from http://db.cs.berkeley.edu/postgres.html 
> 
> IIRC they have a small manual inside the tarball
> 

If you can read academese, a bunch of white papers and technical reports
are available at:

http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/papers/

Ross
-- 
Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> 
NSBRI Research Scientist/Programmer
Computer and Information Technology Institute
Rice University, 6100 S. Main St.,  Houston, TX 77005