Minor doc nits re bug 1262.

Update src comment to be clearer about why it's safe for "state of this transport" to affect other deliveries.
Mention change in externally observable state in README.UPDATING.
Reference bugzilla entry in ChangeLog.
Update Paul's credit in ACKNOWLEDGMENTS.

diff --git a/doc/doc-txt/ChangeLog b/doc/doc-txt/ChangeLog
index f9eaaf7a9dd6e7ccff4f43b6648807bd1b6f0812..a4c10d13afc515ea56cc73abd0aa990eabf7d146 100644 (file)
--- a/doc/doc-txt/ChangeLog
+++ b/doc/doc-txt/ChangeLog
@@ -69,7 +69,8 @@ JH/05 Permit multiple router/transport headers_add/remove lines.
 JH/06 Add dnsdb pseudo-lookup "a+" to do an "aaaa" + "a" combination.
 
 JH/07 Avoid using a waiting database for a single-message-only transport.
 JH/06 Add dnsdb pseudo-lookup "a+" to do an "aaaa" + "a" combination.
 
 JH/07 Avoid using a waiting database for a single-message-only transport.

-      Performance patch from Paul Fisher.
+      Performance patch from Paul Fisher.  Bugzilla 1262.
+

 
 Exim version 4.80
 -----------------
 
 Exim version 4.80
 -----------------

diff --git a/src/ACKNOWLEDGMENTS b/src/ACKNOWLEDGMENTS
index a611a291b1ed55fab6edf96becc6f325f4a6a57a..a66512f369e51e97c860fecb31b266332a0bf2c4 100644 (file)
--- a/src/ACKNOWLEDGMENTS
+++ b/src/ACKNOWLEDGMENTS
@@ -382,6 +382,7 @@ Maxim Dounin              Patch portability of accept() len
 Frank Elsner              Fixed build reliability by exporting LC_ALL=C
 Paul Fisher               Diagnosed smtp_cmd_buffer_size affecting GSSAPI SASL
                             initial response, raised buffer size
 Frank Elsner              Fixed build reliability by exporting LC_ALL=C
 Paul Fisher               Diagnosed smtp_cmd_buffer_size affecting GSSAPI SASL
                             initial response, raised buffer size

+                          Patch adjusting connection_max_messages wait-DB usage

 Oliver Fleischmann        Patches fixing compilation in older compilers
 Julian Gilbey             Helped improve userforward local_part_suffix docs
 Richard Godbee            Patch fixing usage fprintf
 Oliver Fleischmann        Patches fixing compilation in older compilers
 Julian Gilbey             Helped improve userforward local_part_suffix docs
 Richard Godbee            Patch fixing usage fprintf

diff --git a/src/README.UPDATING b/src/README.UPDATING
index b7406f43c177b2b85a9c7087c96db5b899350fe7..5742a64b9e29ea68d6256db0a8442ae9584c1403 100644 (file)
--- a/src/README.UPDATING
+++ b/src/README.UPDATING
@@ -32,6 +32,11 @@ Exim version 4.81
  * New option gnutls_enable_pkcs11 defaults false; if you have GnuTLS 2.12.0
    or later and do want PKCS11 modules to be autoloaded, then set this option.
 
  * New option gnutls_enable_pkcs11 defaults false; if you have GnuTLS 2.12.0
    or later and do want PKCS11 modules to be autoloaded, then set this option.
 

+ * A per-transport wait-<name> database is no longer updated if the transport
+   sets "connection_max_messages" to 1, as it can not be used and causes
+   unnecessary serialisation and load.  External tools tracking the state of
+   Exim by the hints databases may need modification to take this into account.
+

 
 Exim version 4.80
 -----------------
 
 Exim version 4.80
 -----------------

diff --git a/src/src/transports/smtp.c b/src/src/transports/smtp.c
index 36a053f2cdf0f774c13f5f8edae952bbcad36c81..dc24e69389cc401b340308e53c08a5c1a58a3e3f 100644 (file)
--- a/src/src/transports/smtp.c
+++ b/src/src/transports/smtp.c
@@ -3115,8 +3115,8 @@ for (addr = addrlist; addr != NULL; addr = addr->next)
 for which hosts to become available. For some message-specific errors, the
 update_waiting flag is turned off because we don't want follow-on deliveries in
 those cases.  If this transport instance is explicitly limited to one message
 for which hosts to become available. For some message-specific errors, the
 update_waiting flag is turned off because we don't want follow-on deliveries in
 those cases.  If this transport instance is explicitly limited to one message

-per connection, follow-on deliveries are not possible, and there's no need
-to create/update a waiting database. */
+per connection then follow-on deliveries are not possible and there's no need
+to create/update the per-transport wait-<transport_name> database. */

 
 if (update_waiting && tblock->connection_max_messages != 1)
   transport_update_waiting(hostlist, tblock->name);
 
 if (update_waiting && tblock->connection_max_messages != 1)
   transport_update_waiting(hostlist, tblock->name);