Re: [INTERFACES] Convert PGconn, PGresult to opaque types?

>>>>> "TL" == Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:

(sorry, attributions are lost here)

>>>>> Basically this would force applications to use the accessor functions
>>>>> as recommended in the documentation, and not touch fields of a PGconn
>>>>> object directly. (Ditto for PGresult.)
>>>
>>> I am scared about external stuff like php. If they use it, and we
>>> release something that doesn't work with their stuff, we are cooked
>>> until they upgrade.
TL>
TL> But if they are using any direct references to fields of the PGconn
TL> struct, their stuff *already* won't work with 6.4. Admittedly it'd
TL> most likely only take a recompile to fix, and not code changes
TL> (however trivial). But if they'd been using only the documented API,
TL> ie using the accessor functions and not directly touching the struct,
TL> then a new shared library or DLL could be plopped right in without even
TL> a recompile of calling applications.
TL>
TL> Is the PHP source code available? It wouldn't take much work to check
TL> whether it will compile without a definition for struct pg_conn.

The PHP source is available from http://www.php.net/. From a quick
look through it, it does access the PGconn structure directly. Stuff
like (this is from the file php3.0.2a/functions/pgsql.c):

lo_read((PGconn *)pgsql->conn, pgsql->lofd, buf, 8192))

However, the whole PostgreSQL-specific stuff is only 1400 lines worth,
and the PHP guys are reputed to very active, so I don't think a change
should pose too much of a problem if they are forewarned.

[I have CCed the PHP/PostgreSQL module developers]

--
Eric Marsden
emarsden @ mail.dotcom.fr
It's elephants all the way down