[Beepwg] [ANN] Vortex Library 1.1.11 'Man of simple pleasures' is ready!
Francis Brosnan Blazquez
francis en aspl.es
Mie Ene 4 11:59:40 CET 2012
## Vortex Library 1.1.11 ##
## "Man of simple pleasures" ##
Advanced Software Production Line is proud to announce Vortex
Library 1.1.11 "Man of simple pleasures" release availability: an
open source "BEEP Core" protocol implementation written in C, mapped
Vortex Library is a stable and robust application protocol
development toolkit, with active support for Microsoft Windows and
Check out the following links to get ready with Vortex Library:
[Vortex Library homepage]
[Turbulence BEEP server homepage]
[Sources and Win32 binaries]
[BEEP Core community web site]
[Axl Library, XML 1.0 implementation]
This release in short
PYVORTEX: Several updates to make the binding more stable in the
long run (also controlling memory consumption).
General PyVortex update to allow running a handler
watcher that can warn the user that some particular
handler is taking too long to finish and potentially may
cause performance degradation.
TLS: now it is possible to provide a handler that is called
when a TLS handshake fails.
ENGINE: Complete vortex sequencer rewrite to allow handling all
outgoing sending jobs in a fair manner (handling all of
them in rounds, giving all channels the possibility to
send content without being starved by highly interactive
Several fixings and updates to make the product more
- Phil Black-Knight (reported and fixed problem at
vortex_connection_set_profile_mask where the function was
returning a wrong value at the first call. Also reported an API
clash between TLS versions used in debian/ubuntu and redhat).
Changes since 1.1.10
* [fix] Several changes to vortex-regression-client.c to make it
compilable in win64.
* [fix] Added additional headers to fix xml-rpc-support.c under win64.
* [fix] Several modifications to PyVortex to avoid reusing python
references which do not reduce consumption significantly and makes
the binding less maintainable.
* [fix] Updated py channel frame received code to not store the
PyVortexChannel reference as a parameter to the frame received
handler to avoid a crash when the frame received is fired having the
python channel collected (but not its C representation).
* [fix] Updated vortex channel pool module to nullify its reference
inside VortexChannel object to ensure the channel do not access the
channel pool after the pool was uninstalled from the channel. Added
regression test to reproduce the bug and check fix introduced
* [fix] Updated PyVortex regression test to include a check to create
a channel with a frame received handler forcing the PyVortexChannel
reference to be garbaged to check then to receive frames over that
channel. Reg test (py test-10-e).
* [fix] Version files updates..
* [fix] Updated TLS API to allow setting a failure handler that will
be called when an error is found. Really useful at listener side to
get error reporting that until now were passing silently. API added:
Updated documentation to explain how to use this.
* [fix] Updated PyVortexCtx internal register to use internal
VortexCtx to avoid relying on having always available the python
vortex.Ctx that was used.
* [fix] Updated PyVortex listener side frame received, close channel
and start channel handling to avoid passing as reference the python
vortex.Ctx that was used to install those handlers.
* [fix] Updated PyVortex ctx.new_event to return the handle id of the
event installed. Also added ctx.remove_event to allow removing an
event installed providing its unique id. Updated documentation to
explain how to use it.
* [fix] Updated vortex-file-transfer-server.c to also register a
profile that echos all content received..
* [fix] Added new regression test (test-03f) that exposes a bug at
channel pool close when all channels are available.
* [fix] Updated vortex sequencer internal reference checking to use
connection reference from channel instead of the connection
reference directly (channel reference is owned by the sequencer and
connection is not). Updated internal queue priority to acquire a
reference to the channel to avoid wrong dealloc when a connection
failure is found in the middle of a transfer.
* [fix] Fixed wrong header for vortex_thread_pool_stats.
* [fix] Reorganized vortex internal debug log to use a single shot to
write the log (to avoid missing log lines under heavy logging). Also
updated to print a time stamp for each log.
* [fix] Removed flag not required to get access to vsnprintf
* [fix] Making automatic thread pool resize to be also checked on
thread pool wait timeout.
* [fix] Fixed reg test 00-c2 which wasn't removing all threads created
causing test 00-c1 to fail.
* [fix] Added additional check to channels internal reference inside
* [fix] Updated PyVortexChannel set_frame_received to allow removing
handler installed when no parameter is received. Updated
documentation and checked regression tests..
* [fix] Updated vortex sequencer module to avoid doing a priority
queue for resequence operations (when a SEQ frame was received) to
avoid giving priority to large send operations that are served fast
and also acked fast by the receiving side (causing starvation for
the rest of pending operations). Also updated resequence signal
code to check first if there are already other previous notification
(to avoid filling the queue with notifications).
* [fix] Updated vortex thread API to include a new function that
allows doing search operations over a queue looking it during the
search. API added:
* [fix] Fixed pop_channel_error example inside connection
* [fix] Fixed vortex.Connection documentation about getting channel
* [fix] Complete vortex sequencer rewrite to allow handling all
outgoing sending jobs in a fair manner (treating them in rounds,
giving all channels the possibility to send content without being
starved by highly interactive transfers). This rewrite will also
open the possibility to implement channel sending priorities. All reg
* [fix] Removed old code used to drop messages for broken connections
(inside vortex sequencer) which is no longer needed.
* [fix] Several fixings to make new vortex sequencer design to work
with microsoft windows.
* [fix] Fixed automatic thread pool removal which wasn't checking for
running threads causing to remove threads from the pool under heavy
* [fix] Making vortex channel send msg and send rpy operations to
acquire a reference to the channel object during the operation to
avoid races with new vortex sequencer implementation.
* [fix] Applied several fixings according to valgrind's hellgrind
information (awesome tool!)...
* [fix] Code cleanups at vortex channel module (removed internal
variables last_reply_delivered, last_msg_delivered,
* [fix] Updated vortex_set_conf API to include a new key to allow
controlling if context finalization should wait for threads inside
the thread pool to finish. API added:
Added regression test (test_00-e) to reproduce a situation where
threads from the pool are blocked and to check that even in those
situations vortex_exit_ctx finishes right. Also updated
* [fix] Version updates..
* [fix] Fixed PyVortex start channel start data reference (it was
passed in properly).
* [fix] General PyVortex update to allow running a handler watcher
that can warn the user that some particular handler is taking too
long to finish and potentially may cause performance degradation.
* [fix] Updated PyVortex internal exception handling to report and
report exceptions to better cancel/close current operations rather
just ignoring it.
* [fix] Updated PyVortex to allow registering a handler that is called
when it is detected a python handler is taking too long to finish.
* [fix] Removed unnecessary lock/unlock pair from vortex channel
module. Moved log debug which may cause wrong memory access under
* [fix] Making vortex.Ctx.new_event to flag event period with "l"
instead of "i" to avoid OverflowError: signed integer is greater
* [fix] Updated internal channel add handler to report a boolean value
with the operation result so channel resources can be released after
* [fix] Making vortex pull cleanup to also clear event masks.
* [fix] Making vortex thread pool to acquire the VortexCtx reference
before starting the pool thread (to avoid races where thread isn't
still started and VortexCtx was finished when having skip thread
pool wait flag enabled).
* [fix] Moved debug log before connection registering at
vortex_listener module to avoid wrong mem. access under some cases..
* [fix] Making vortex ALIVE implementation to check vortex_channel_ref
result before continuing. Also making the implementation to grab the
reference before setting the channel used by the check loop (fixed
race exposed by PyVortex).
* [fix] Fixed memory leak at PyVortex caused by a circular reference
built by vortex connection reference counting and python reference
counting. Added initial file.sup to better track all memory consumed
* [fix] Included file.sup into make dist result.
* [fix] Fixed memory leak inside PyVortex handler tracker: it was
removing handler tracking object without calling destroy func: oops!
* [fix] Phil Black-Knight reported and fixed a problem at
vortex_connection_set_profile_mask where the function was always
returning an error code for the first handler installed (although it
was indeed properly installed).
* [fix] Removed code that checks for reference counting at internal
channel termination when detected pending messages to be sent (new
design doesn't requires this).
* [fix] Fixed API clash between debian/ubuntu and RHEL where
SSL_CTX_new and method created returns and receives const SSL_METHOD
*. Updated the code to avoid declaring variables with that type
passing directly the value returned from method creation into
SSL_CTX_new. Bug reported by Philip Black-Knight..
Advanced Software Production Line provides GNU/Linux support and
consulting services to help organizations to introduce GNU/Linux
inside its process.
Contact us, using English or Spanish, to get commercial support
and/or BEEP based development services.
You can reach us:
http://www.aspl.es - info en aspl.es
Enjoy Vortex Library!
Francis Brosnan Blázquez - francis en aspl.es
Advanced Software Production Line - http://www.aspl.es
4th jan 2012, Madrid (Spain)
Más información sobre la lista de distribución Beepwg