SGML cleanup
This patch fixes various minor issues with the SGML docs:
- SGML markup police: <command> for SQL commands, <productname> for the
formal name "PostgreSQL", <type> for type names, and so on
- untabify release.sgml and runtime.sgml (I'll do a more thorough
untabification later)
- add a section "Deprecated Features" to the release notes, and move the
to_char(interval) notice to it. Also added a notice about the
planned-in-8.1 change of default_use_oids to false by default. Does
anyone know of any more deprecated features in 8.0 that we have active
plans to get rid of?
- other minor improvements
Barring any objections, I intend to apply this patch within the next 48
hours or so.
-Neil
Attachments:
pg_80_relnote_improv-6.patchtext/plain; name=pg_80_relnote_improv-6.patch; x-mac-creator=0; x-mac-type=0Download+670-642
Neil Conway wrote:
This patch fixes various minor issues with the SGML docs:
Woops, I forgot we renamed default_use_oids to default_with_oids.
Updated patch attached.
-Neil
Attachments:
pg_80_relnote_improv-7.patchtext/plain; name=pg_80_relnote_improv-7.patch; x-mac-creator=0; x-mac-type=0Download+670-642
On Sun, Aug 22, 2004 at 02:48:23PM -0700, Neil Conway wrote:
While you are at it ...
<listitem>
<para>
A new version of the <application>plperl</> server-side language now
! supports a persistent shared storage area, triggers, returning records
! and arrays of records, and SPI calls to access the database.
</para>
</listitem>
"arrays of records" seems inaccurate; how about "set of records"
instead?
<para>
! The SQL2003 standard specifies that the keyword
! <literal>SAVEPOINT</> is mandatory. <productname>PostgreSQL</> and
! <productname>Oracle</> allow the <literal>SAVEPOINT</literal>
! keyword to be omitted. SQL2003 allows only <literal>WORK</>, not
Why are we mentioning Oracle behavior at all?
--
Alvaro Herrera (<alvherre[a]dcc.uchile.cl>)
"El n�mero de instalaciones de UNIX se ha elevado a 10,
y se espera que este n�mero aumente" (UPM, 1972)
Alvaro Herrera wrote:
<para>
! The SQL2003 standard specifies that the keyword
! <literal>SAVEPOINT</> is mandatory. <productname>PostgreSQL</> and
! <productname>Oracle</> allow the <literal>SAVEPOINT</literal>
! keyword to be omitted. SQL2003 allows only <literal>WORK</>, notWhy are we mentioning Oracle behavior at all?
One of the influences on the SAVEPOINT/ROLLBACK syntax was avoiding
gratuitous incompatibility with Oracle; doesn't it make sense to
document this?
-O
Alvaro Herrera wrote:
"arrays of records" seems inaccurate; how about "set of records"
instead?
Okay, I'll check this in separately this evening/tomorrow.
Why are we mentioning Oracle behavior at all?
I don't have a strong opinion on this -- is there a consensus we don't
want to do it?
-Neil
Neil Conway wrote:
Woops, I forgot we renamed default_use_oids to default_with_oids.
Updated patch attached.
Applied to HEAD.
-Neil
Neil Conway <neilc@samurai.com> writes:
Alvaro Herrera wrote:
Why are we mentioning Oracle behavior at all?
I don't have a strong opinion on this -- is there a consensus we don't
want to do it?
I think the existing mention is fine. It's silly to pretend that these
decisions are not motivated by Oracle compatibility, especially when we
are deliberately deviating from the spec syntax.
regards, tom lane