summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2010-10-08More cleanup of session termination process.Martin Sustrik
Fix in termination of session attached to a REP socket. Several fixes of session termination in case of disconnection. Signed-off-by: Martin Sustrik <sustrik@250bpm.com>
2010-10-08Scalability improvements for large amounts of connectionsMartin Lucina
Add signaler_sndbuf_size option to config.hpp which allows the user to increase the size of the send buffer used by the signalling socketpair. Implement random backoff for reconnection attempts using a primitive pseudo-random generation to prevent reconnection storms. Add wait_before_connect option to config.hpp to allow the user to enable random delay even on initial connect. Default is false for low latency. Signed-off-by: Martin Lucina <mato@kotelna.sk>
2010-10-08Clean-up of session termination processMartin Sustrik
Specifically, shutdown of child objects is initiated *before* termination handshake with socket object. Signed-off-by: Martin Sustrik <sustrik@250bpm.com>
2010-10-02make sure that reader does not terminate in the middle of a multipart messageNir Soffer
2010-10-01issue 92 -- Assertion failed: inpipe && outpipe (pair.cpp:86) -- fixedMartin Sustrik
2010-10-01Merge branch 'master' of github.com:zeromq/zeromq2Martin Sustrik
2010-10-01assert when xrep socket gets reconnected in the middle of the shutdown -- fixedMartin Sustrik
2010-09-30zmq_poll(): Fix busy-loop if timeout is zeroMartin Lucina
Fix a case where zmq_poll() (poll-based version) could go off into a busy-loop if no revents are returned and the timeout passed in is zero.
2010-09-30Merge branch 'maint'Martin Sustrik
* maint: zmq_msg_move called on uninitialised message in xrep_t::xrecv -- fixed
2010-09-30zmq_msg_move called on uninitialised message in xrep_t::xrecv -- fixedMax Wolf
2010-09-30fix in XREP's fair queueing mechanismMartin Sustrik
2010-09-30minor whitespace fixMartin Sustrik
2010-09-30FreeBSD doesn't define MSG_ERRQUEUE -- now it's defined in 0MQSteven McCoy
2010-09-30more fixes to (e)pgm transportSteven McCoy
2010-09-30 * Add assertions to check for OpenPGM calls with invalid parameters.Steven McCoy
* Assertion to check that pgm_getaddrinfo is actually returning something. * Missing pgm_connect call. * Typo on TOS causing immediate abort. * Placeholder calls for timeouts whilst continuing spin loop functionality. * OpenPGM v5 now supports reference counting so remove init checks. * Duplicate UDP unicast port setting, requires one unicast and one multicast. * Incorrectly set socket rcvbuf size with sndbuf. * Replace std::lexicographical_compare of TSI's with long word integer comparisons. * pgm_socket_t::receive returns -1 on no data.
2010-09-28Upgrade to OpenPGM-5.0.78Steven McCoy
2010-09-28ZMQ_TYPE socket option addedGonzalo Diethelm
2010-09-28Merge branch 'maint'Martin Sustrik
* maint: crash when closing an ypipe -- fixed
2010-09-28crash when closing an ypipe -- fixedDhammika Pathirana
2010-09-28REQ socket can die when reply is delivered on wrong unerlying connection -- ↵Martin Sustrik
fixed
2010-09-27MSVC build fixedMartin Sustrik
2010-09-27latency optimisation for the case where there are no timersMartin Sustrik
2010-09-26timers properly implementedMartin Sustrik
2010-09-26common base for all pollers created; the only thing it handles at the moment ↵Martin Sustrik
is 'load'
2010-09-26new interfaces for timers; the implementation is not changed yetMartin Sustrik
2010-09-26minor beautificationsMartin Sustrik
2010-09-26clock_t class addedMartin Sustrik
2010-09-26Merge branch 'maint'Martin Sustrik
* maint: if TSC jumps backwards (in case of migration to a different CPU core) latency peak may occur -- fixed Conflicts: src/app_thread.cpp
2010-09-26if TSC jumps backwards (in case of migration to a different CPU core) ↵Martin Sustrik
latency peak may occur -- fixed
2010-09-21minor error in comment fixedMartin Sustrik
2010-09-21The flag in the socket has clear semantics now -- it tracks whether ↵Martin Sustrik
corresponding context was closed, it doesn't track whether zmq_close was called on the socket itself
2010-09-20zmq_poll (select version) now correctly assumes that ZMQ_FD is edge-trigerredMartin Sustrik
2010-09-20zmq_poll honours ZMQ_FORCE_POLL and ZMQ_FORCE_SELECT optionsMartin Lucina
2010-09-20minor problem in zmq_poll (select version) fixedMartin Lucina
2010-09-20problem with engine being attached to session while it's being terminated fixedMartin Sustrik
2010-09-19zmq_init_t destroyed zmq_engine_t before plugging it out from the poller ↵Martin Sustrik
first -- fixed
2010-09-19minor code reorganisation in ctx_t::terminateMartin Sustrik
2010-09-19synchronisation problem fixed in ctx_tMartin Sustrik
2010-09-19two races fixed in session creationMartin Sustrik
2010-09-19termination of pipe via delimiter message could stuck when no data were read ↵Martin Sustrik
from the pipe (because connection wasn't active) -- fixed
2010-09-19obsolete 'active' flag removed from session_tMartin Sustrik
2010-09-18single term ack counting mechanism for every socket (no separate mechanisms ↵Martin Sustrik
for fq_t and lb_t)
2010-09-18more details on launch_sibling in commentsMartin Sustrik
2010-09-18potential duplicate termination of pipes removedMartin Sustrik
2010-09-17reconnection process fixed when failure occurs during init phaseMartin Sustrik
2010-09-17bug in REQ+multipart fixedMartin Sustrik
2010-09-15Merge branch 'maint'Martin Sustrik
* maint: values of RATE, RECOVERY_IVL and SWAP options are checked for negative values
2010-09-15values of RATE, RECOVERY_IVL and SWAP options are checked for negative valuesMartin Sustrik
2010-09-13zmq::writer_t: Add missing test for swapMartin Lucina
2010-09-10lefover assert removedMartin Sustrik