GitLab

Note the following: (from the API file)
  IMPORTANT: elvin uses timeouts internally. So, this function does
  NOT guarantee that when it returns, either an event has been
  recieved or your timeout has passed. This should not be much of
  a problem, but you have been warned!

The above is not really fixable, without hacking elvin. And it may
not be entirely fixable even then. In particular, the first call to
event_poll_blocking() will always return at once, since there are
leftover timers from connecting to elvind.

notification. Essentially a wrapper around the the elvin equivalent.

To be used by the forthcoming perl event system module.

hainv the sender usurp the host field in the tuple. The event library
sets this for the user.

system programs. Build two versions of the event library, one for non
threaded programs and another (libevent_r.a) for threaded versions.
Remove all that -pthread goo from all of the makefiles, except for the
scheduler (which is threaded); it gets linked as before, but with
-levent_r instead. If you try and use the threaded API without the
proper link, the library will print an error message and quit.

comma (or space) delimited list of tags for each member of the address
tuple. So, if you wanted to subscribe to LINK and TIME events, you
would set the address tuple like this:

	tuple->objtype   = "TRAFGEN,TIME";

and the subscription clause that would be passed to elvin would look
something like this:

	( OBJTYPE == "TRAFGEN" || OBJTYPE == "TIME" )

This should make it easier to subscribe to multiple types of events,
without having to get all of them. For example, the trafgen agent
could subscribe to TIME events and get the TIME START event that is
now sent once all nodes report ISUP

them into a private header file.

scheme that we discussed in email. Notifications and subscriptions now
take an "address_tuple" argument (I know, crappy name) that is a
structure that looks like this:

	char		*site;		/* Which Emulab site. God only */
	char		*expt;		/* Project and experiment IDs */
	char		*group;		/* User defined group of nodes */
	char		*host;		/* A specific host */
	char		*objtype;	/* LINK, TRAFGEN, etc ... */
        char		*objname;	/* link0, cbr0, cbr1, etc ... */
        char		*eventtype;	/* START, STOP, UP, DOWN, etc ... */

These can be a specific value, ADDRESSTUPLE_ANY if you are a
subscriber, or ADDRESSTUPLE_ALL if you are a producer. The reason for
the distinction is that you can optimize the match expression with the
extra bit of information, and the above structure can make for a
fairly lengthy match expression, which takes more time of course.
You should use address_tuple_alloc() and address_tuple_free() rather
than allocating them yourself. Note that host above is actually the
ipaddr of control interface. This turns out to be more convenient
since free nodes do not have virtual names.

Also added a new tbgen directly. This directory includes 3 programs in
the making:

tbmevd: Is the Testbed Master Event Daemon, to be run on boss and will
handle TBCONTROL events (reboot, reload, etc). It is just a shell of a
program right now, that takes the events but does not do anything
useful with them. Have not defined what the events are, and what DB
state will be modified.

tbmevc: Is the Testbed Master Event Client (akin to tmcc). It
generates TBCONTROL events which the tbmevd will pick up and do
something useful with. This program is intended to be wrapped by a
perl script that will ask the tmcd for the name of the boss (running
the event daemon).

sample-client: This is a little client to demonstrate how to connect
to the event system and use the address tuple to subscribe to events,
and then how to get information out of notifications.

Note that I have not created a proper build environment yet, so new
programs should probably go in the event dir for now, and link using
the same approach as in tbgen/GNUmakefile.in.

Also several changes for building on FreeBSD. I've left the older
Makefiles in place so Ian can continue to build in his environmant
(Linux?).

Fixed typo in API (s/event_notification/event_notification_t/).

Tell user (in README) to configure Elvin library with --enable-threads
to enable threaded API support.

Removed instructions on enabling server discovery feature (now optional).

Document event_register interface changes ("threaded" arguement, which
allows caller to choose between the thread-safe "threaded" API and the
simpler "sync" API).

Mention (in API) that event_notify and event_schedule don't deallocate
notification.

the threaded API; otherwise, use the sync API.  This option allows
multi-threaded programs to access the threaded API, which is thread-
safe.

to be scheduled at a later time. The interface to event_schedule is
identical to event_notify, except it takes an additional struct
timeval argument that specifies when the event should be fired.  We
assume time synchronization between nodes.

* Revamped the attribute interface. Rather than a single get and put
function that takes a union "type" argument, we now have separate
event_notification_get_<type> and event_notification_put_<type>
functions, where <type> is one of "double", "int32", "int64",
"opaque", "string". These changes should greatly simply the attribute
interface. The opaque attribute type is new, and allows
arbitrary data structures to be added to notifications as attributes.

* Added event_notification_remove, which deletes an attribute from a
notification.

* Modified the event notification callback to take "host" and "type"
parameters, which contain the "host" and "type" attributes from the
event notification, respectively.

the heap, not the stack, since event_subscribe's stack frame will
not exist when notify_callback is invoked.

Elvin server discovery protocol.

document.