pg.ddx.io  pgsql-hackers@postgresql.org mailing list archive  

help / color / mirror / Atom feed

From: Alvaro Herrera <alvaro.herrera@2ndquadrant.com>
To: Boris Kolpackov <boris@codesynthesis.com>
Cc: pgsql-hackers@lists.postgresql.org
Subject: Re: Pipeline mode and PQpipelineSync()
Date: Wed, 7 Jul 2021 13:30:46 -0400
Message-ID: <202107071730.vkyej2goquhp@alvherre.pgsql> (raw)
In-Reply-To: <boris.20210707170157@codesynthesis.com>

On 2021-Jul-07, Boris Kolpackov wrote:

>     // Try to minimize the chance of blocking the server by first processing
>     // the result and then sending more queries.
>     //
>     if (FD_ISSET (sock, &rds))
>     {
>       if (PQconsumeInput (conn) == 0)
>         assert (false);
> 
>       while (PQisBusy (conn) == 0)
>       {
>         //fprintf (stderr, "PQgetResult %zu\n", rn);
> 
>         PGresult* res = PQgetResult (conn);
>         assert (res != NULL);
>         ExecStatusType stat = PQresultStatus (res);

Hmm ... aren't you trying to read more results than you sent queries?  I
think there should be a break out of that block when that happens (which
means the read of the PGRES_PIPELINE_SYNC needs to be out of there too).
With this patch, the program seems to work well for me.

***************
*** 94,112 ****
        while (PQisBusy (conn) == 0)
        {
          //fprintf (stderr, "PQgetResult %zu\n", rn);
  
          PGresult* res = PQgetResult (conn);
          assert (res != NULL);
          ExecStatusType stat = PQresultStatus (res);
  
-         if (stat == PGRES_PIPELINE_SYNC)
-         {
-           assert (wdone && rn == n);
-           PQclear (res);
-           rdone = true;
-           break;
-         }
- 
          if (stat == PGRES_FATAL_ERROR)
          {
            const char* s = PQresultErrorField (res, PG_DIAG_SQLSTATE);
--- 94,110 ----
        while (PQisBusy (conn) == 0)
        {
          //fprintf (stderr, "PQgetResult %zu\n", rn);
+         if (rn >= wn)
+         {
+           if (wdone)
+             rdone = true;
+           break;
+         }
  
          PGresult* res = PQgetResult (conn);
          assert (res != NULL);
          ExecStatusType stat = PQresultStatus (res);
  
          if (stat == PGRES_FATAL_ERROR)
          {
            const char* s = PQresultErrorField (res, PG_DIAG_SQLSTATE);
***************
*** 190,195 ****
--- 188,201 ----
          break; // Blocked or done.
        }
      }
+ 
+     if (rdone)
+     {
+       PGresult *res = PQgetResult(conn);
+       assert(PQresultStatus(res) == PGRES_PIPELINE_SYNC);
+       PQclear(res);
+       break;
+     }
    }
  
    if (PQexitPipelineMode (conn) == 0 ||
***************
*** 246,248 ****
--- 252,269 ----
      PQclear (res);
    }
  }
+ 
+ int main(int argc, char **argv)
+ {
+   PGconn *conn = PQconnectdb("");
+   if (PQstatus(conn) != CONNECTION_OK)
+   {
+     fprintf(stderr, "connection failed: %s\n",
+             PQerrorMessage(conn));
+     return 1;
+   }
+ 
+   test(conn);
+ }
+ 
+ 


-- 
Álvaro Herrera              Valdivia, Chile  —  https://www.EnterpriseDB.com/

view thread (41+ messages)  latest in thread

Message-ID: <202107071730.vkyej2goquhp@alvherre.pgsql>
Permalink:  ../202107071730.vkyej2goquhp@alvherre.pgsql/
Also on:    postgresql.org/message-id/202107071730.vkyej2goquhp@alvherre.pgsql

copy link · copy postgr.es

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: pgsql-hackers@postgresql.org
  Cc: alvaro.herrera@2ndquadrant.com, boris@codesynthesis.com, pgsql-hackers@lists.postgresql.org
  Subject: Re: Pipeline mode and PQpipelineSync()
  In-Reply-To: <202107071730.vkyej2goquhp@alvherre.pgsql>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox