Message of move

From: Kovacs Baldvin <kb136(at)hszk(dot)bme(dot)hu>
To: hannu(at)tm(dot)ee
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Message of move
Date: 2001-04-07 20:08:45
Message-ID: Pine.GSO.4.21.0104072201130.27305-200000@ural2
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi.

A few weeks (months?) ago I made a patch to the postgres
backend to get back the number of realized moves after
a MOVE command. So if I issue a "MOVE 100 IN cusrorname",
but there was only 66 rows left, I get back not only "MOVE",
but "MOVE 66". If the 100 steps could be realized, then
"MOVE 100" would come back.

I send this info to you, because I would like to ask you if
it could be OK to include in future versions. I think you
are in a beta testing phase now, so it is trivially not the
time to include it now...

The solution is 2 code lines into command.c, and then the
message of move cames with the number into for example psql.
1 other word into the jdbc driver, and this number is
available at update_count.

I made the patch to the latest (one day old) CVS version.

Here are the patches. Please look at them, and if you think
it's a good idea, then please let me know where and how should
I post them, and approximately when will you finish with the
beta testing, so it can be really considered seriously.

I included them also as an attachment, because my silly pine
insists to break the lines...

--- ./src/backend/commands/command.c.orig Fri Mar 23 05:49:52 2001
+++ ./src/backend/commands/command.c Sat Apr 7 10:24:27 2001
@@ -174,6 +174,12 @@
if (!portal->atEnd)
{
ExecutorRun(queryDesc, estate, EXEC_FOR, (long) count);
+
+ /* I use CMD_UPDATE, because no CMD_MOVE or the like
+ exists, and I would like to provide the same
+ kind of info as CMD_UPDATE */
+ UpdateCommandInfo(CMD_UPDATE, 0, estate->es_processed);
+
if (estate->es_processed > 0)
portal->atStart = false; /* OK to back up now */
if (count <= 0 || (int) estate->es_processed < count)
@@ -185,6 +191,12 @@
if (!portal->atStart)
{
ExecutorRun(queryDesc, estate, EXEC_BACK, (long) count);
+
+ /* I use CMD_UPDATE, because no CMD_MOVE or the like
+ exists, and I would like to provide the same
+ kind of info as CMD_UPDATE */
+ UpdateCommandInfo(CMD_UPDATE, 0, -1*estate->es_processed);
+
if (estate->es_processed > 0)
portal->atEnd = false; /* OK to go forward now */
if (count <= 0 || (int) estate->es_processed < count)

Here is the patch for the jdbc driver. >! I couldn't test it
with the current version, because it needs ant, and I didn't
have time and money today to download it... !< However, it
is a trivial change, and if Peter T. Mount reads it, I ask
him to check if he likes it... Thanks for any kind of answer.

--- ./src/interfaces/jdbc/org/postgresql/Connection.java.orig Wed Jan 31 09:26:01 2001
+++ ./src/interfaces/jdbc/org/postgresql/Connection.java Sat Apr 7 16:42:04 2001
@@ -490,7 +490,7 @@
recv_status = pg_stream.ReceiveString(receive_sbuf,8192,getEncoding());

// Now handle the update count correctly.
- if(recv_status.startsWith("INSERT") || recv_status.startsWith("UPDATE") || recv_status.startsWith("DELETE")) {
+ if(recv_status.startsWith("INSERT") || recv_status.startsWith("UPDATE") || recv_status.startsWith("DELETE") || recv_status.startsWith("MOVE")) {
try {
update_count = Integer.parseInt(recv_status.substring(1+recv_status.lastIndexOf(' ')));
} catch(NumberFormatException nfe) {

-------------------
(This last looks a bit complex, but the change is really a new
"|| recv_status.startsWith("MOVE")" only...)

Thank you for having read this,

Baldvin

p.s.: I read a page on your homepage, called "unapplied patches".
I would like to know if it means "still unapplied patches", or
it means "bad, and not accepted ideas".

Attachment Content-Type Size
move.patch text/plain 1.8 KB

Browse pgsql-hackers by date

  From Date Subject
Next Message Kovacs Baldvin 2001-04-07 20:10:45 The makefile of pgaccess (CVS)
Previous Message Tom Ivar Helbekkmo 2001-04-07 19:41:04 Re: Re: Call for platforms