This fixes further issues on win64 with casts from the streams.
Sockets/descriptors handling was unitized. This has an impact only on
win64, php_socket_t otherwise can be feed back to int datatype.
- New functions (each accepts a pgsql $connection resource):
. pg_connect_poll
. pg_socket
. pg_consume_input
. pg_flush
- Modified functions
The following functions now additionally return zero if the
underlying socket is set to non-blocking mode and the send
operation does not complete immediately. Previously these
functions returned only boolean TRUE/FALSE and blocked
execution while polling until all data was sent:
. pg_send_execute
. pg_send_prepare
. pg_send_query
. pg_send_query_params
- New constants
Used with pg_connect() to initiate an asynchronous connection
attempt:
. PGSQL_CONNECT_ASYNC
Used with pg_connection_status() to determine the current state
of an async connection attempt:
. PGSQL_CONNECTION_STARTED
. PGSQL_CONNECTION_MADE
. PGSQL_CONNECTION_AWAITING_RESPONSE
. PGSQL_CONNECTION_AUTH_OK
. PGSQL_CONNECTION_SSL_STARTUP
. PGSQL_CONNECTION_SETENV
Used with pg_connect_poll() to determine the result of an
async connection attempt:
. PGSQL_POLLING_FAILED
. PGSQL_POLLING_READING
. PGSQL_POLLING_WRITING
. PGSQL_POLLING_OK
. PGSQL_POLLING_ACTIVE
- Polling via returned pg_socket() stream
pg_socket() returns a read-only socket stream that may be
cast to a file descriptor for select (and similar) polling
operations. Blocking behavior of the pgsql connection socket
can be controlled by calling stream_set_blocking() on the
stream returned by pg_socket().
Switch to SHA1, which match internal openssl hardcoded algo.
In most case, won't even be noticed
- priority on user input (default_md)
- fallback on system config
- fallback on this default value
Recent system reject MD5 digest, noticed in bug36732.phpt failure.
While SHA1 is better than MD5, SHA256 is recommenced,
and defined as default algo in provided configuration on
recent system (Fedora 21, RHEL-7, ...). But the idea is to
keep in sync with openssl internal value for PHP internal value.
The FPM change didn't make it into 5.5.10 but will instead be in 5.5.11; http://php.net/Changelog-5.php, the release and the NEWS file in the release accurately reflect this, but the 5.5 branch NEWS doesn't (as the release was made and tagged from the last RC, which didn't contain that change yet).
