database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-07-12 21:01:52 +03:00
Code Activity

From: Tom Lane <tgl@sss.pgh.pa.us>

Browse Source 
Making PQrequestCancel safe to call in a signal handler turned out to be
much easier than I feared.  So here are the diffs.

Some notes:
  * I modified the postmaster's packet "iodone" callback interface to allow
    the callback routine to return a continue-or-drop-connection return
    code; this was necessary to allow the connection to be closed after
    receiving a Cancel, rather than proceeding to launch a new backend...
    Being a neatnik, I also made the iodone proc have a typechecked
    parameter list.
  * I deleted all code I could find that had to do with OOB.
  * I made some edits to ensure that all signals mentioned in the code
    are referred to symbolically not by numbers ("SIGUSR2" not "2").
    I think Bruce may have already done at least some of the same edits;
    I hope that merging these patches is not too painful.
This commit is contained in:
Marc G. Fournier
1998-07-09 03:29:11 +00:00
parent 8bf61820f0
commit a0659e3e2c
20 changed files with 597 additions and 287 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
src/bin/psql/psql.c
View File

@ -7,7 +7,7 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/bin/psql/Attic/psql.c,v 1.146 1998/06/16 07:29:38 momjian Exp $
* $Header: /cvsroot/pgsql/src/bin/psql/Attic/psql.c,v 1.147 1998/07/09 03:28:53 scrappy Exp $
*
*-------------------------------------------------------------------------
*/
@ -291,15 +291,27 @@ PSQLexec(PsqlSettings *pset, char *query)
* If interactive, we enable a SIGINT signal catcher that sends
* a cancel request to the backend.
* Note that sending the cancel directly from the signal handler
* is safe only because the cancel is sent as an OOB message.
* If it were inline data, then we'd risk inserting it into the
* middle of a normal data message by doing this.
* (It's probably not too cool to write on stderr, for that matter...
* but for debugging purposes we'll risk that.)
* is safe only because PQrequestCancel is carefully written to
* make it so. We have to be very careful what else we do in the
* signal handler.
* Writing on stderr is potentially dangerous, if the signal interrupted
* some stdio operation on stderr. On Unix we can avoid trouble by using
* write() instead; on Windows that's probably not workable, but we can
* at least avoid trusting printf by using the more primitive fputs.
*/
static PGconn * cancelConn = NULL; /* connection to try cancel on */
static void
safe_write_stderr (const char * s)
{
#ifdef WIN32
fputs(s, stderr);
#else
write(fileno(stderr), s, strlen(s));
#endif
}
static void
handle_sigint (SIGNAL_ARGS)
{
@ -307,11 +319,13 @@ handle_sigint (SIGNAL_ARGS)
exit(1); /* accept signal if no connection */
/* Try to send cancel request */
if (PQrequestCancel(cancelConn))
fprintf(stderr, "\nCANCEL request sent\n");
{
safe_write_stderr("\nCANCEL request sent\n");
}
else
{
fprintf(stderr, "\nCannot send cancel request:\n%s\n",
PQerrorMessage(cancelConn));
safe_write_stderr("\nCannot send cancel request:\n");
safe_write_stderr(PQerrorMessage(cancelConn));
}
}