da85d39c0c
codec: const-ify socket
...
Avoid possible data race of changing _channel
before worker threads are joined.
Was just using this as a boolean flag for a method
which is already guarded in BlockingTCPTransportCodec::close()
from being called twice.
2018-04-02 10:41:30 -07:00
38965bb847
codec: collapse internalDestroy() into internalClose()
2018-04-02 10:24:35 -07:00
6c2d20353a
minor, avoid extra string copys
2018-03-28 13:59:33 -07:00
a6d86d2a3c
hide MockTransportSendControl
2018-03-28 13:49:21 -07:00
735dc22e2d
collapse SimpleChannelSearchManagerImpl into ChannelSearchManager
...
Virtual base with only one sub-class. So avoid some unnecessary
indirection.
2018-03-28 13:37:42 -07:00
1a100a0955
client fix init order leading to *NULL
...
Avoid a race between the main thread alloc of SimpleChannelSearchManagerImpl
the the UDP socket worker trying to use the same.
Split allocation and initialization of search manager
around socket creation and worker start.
Also in-line initializeUDPTransport() to only call site.
2018-03-28 13:21:46 -07:00
aec1c7c9d7
client search: cleanup context pointer handling
2018-03-28 11:25:54 -07:00
8d363f9fe7
server responseHandlers annotate OVERRIDE FINAL
2018-03-22 14:03:57 -07:00
3b53e81f2f
avoid destroy() w/ lock
2018-03-22 13:39:04 -07:00
e9cc8298a6
server monitor cleanup
...
ServerMonitorRequesterImpl::getChannelMonitor()
seems like no possibility of deadlock atm.
2018-03-22 13:34:55 -07:00
4fcbdff580
missing includes
2018-03-19 15:36:21 -07:00
93196e6314
drop rtemsConfig.c
...
no longer needed.
2018-03-19 09:29:25 -07:00
8d262fc9fe
simpler test harness main()
2018-03-13 18:55:49 -07:00
9687555c98
Move epicsExit() call into pvAccessAllTests()
...
Needed on VxWorks to display the test summary.
2018-03-13 13:46:56 -05:00
802efbd6c1
Rename vxTestHarness -> pvaTestHarness
...
Integrate with Michael's similar changes for RTEMS.
2018-03-13 13:46:14 -05:00
0e61565308
rename rtemsTestHarness -> pvaTestHarness
...
avoid name clash with rtemsTestHarness
from pvDataCPP.
2018-03-12 09:13:59 -07:00
af2adf9248
pvlist remove comma
...
some shells will include the trailing ',' in selection
when double clicking. Users find this annoying.
2018-03-08 11:43:36 -05:00
acab4cd408
fix potential *NULL from Get2PutProxy
2018-03-08 11:42:56 -05:00
23c2ed3484
spamme respond to all spam
...
match the prefix "spam" to allow
clients to subscribe eg. "spam1"
and "spam2" to test concurrent
streaming and load balancing.
2018-02-27 17:35:16 -08:00
74fbd22d40
pvput: fix enum_t
...
Need to get value.choices
2018-02-27 14:29:50 -08:00
d605eaca11
pvput avoid unnecessary iteration
2018-02-27 13:27:38 -08:00
ef74bdeccd
Merge remote-tracking branch 'origin/master'
...
* origin/master:
implement suggestions from Andrew
termination and error handling changes.
2018-02-21 11:20:20 -08:00
909cef9200
minor
2018-02-21 11:19:58 -08:00
b5bf6a4ccd
client: (shallow) copy into Monitor::root
...
Avoid exposing refs. from monitor queue
which might then be used after release().
Also allows users the possibility of caching
getSubField() until root.get() actually changes.
2018-02-21 11:19:21 -08:00
79ada524fa
getRequester() may return NULL
2018-02-20 11:28:32 -08:00
50cee1d161
minor
2018-02-20 10:18:39 -08:00
fe52a1673e
Merge pull request #88 from mrkraimer/master
...
termination and error handling changes.
2018-02-16 05:29:43 -05:00
5b32b4e072
implement suggestions from Andrew
2018-02-15 15:57:00 -05:00
685490fc1a
no NDEBUG
...
Apparently the Base makefiles never have defined NDEBUG.
So no point for this conditional.
2018-02-14 08:55:01 -08:00
18758fa497
throw new doesn't work
...
javaism...
2018-02-13 18:07:58 -08:00
c6b1d3b738
fairQueue: logic error allows perpetual clear()
...
oops, the argument isn't cleared in all
cases, and can be swapped in.
As in the case of clear().
2018-02-13 17:58:56 -08:00
a266777d82
destroyAllChannels() swap out channels list
...
avoid possible modifications while iterating.
2018-02-13 17:58:56 -08:00
8e88f534bb
spamme advertise NTScalar
2018-02-13 17:49:02 -08:00
d12a6cad9b
Move PVACCESS_DEFAULT_PROVIDER def out of header
2018-02-08 11:32:18 -08:00
b397e3e928
Revert "convert global string "constants" to macros"
...
This reverts commit 49b2507985
2018-02-08 11:00:40 -08:00
952c482d68
merge with latest from epics-base
2018-02-08 07:58:52 -05:00
5c1b638db5
avoid spurious Leak warning
...
copying an array just to see if it is empty
isn't so efficient.
2018-02-07 20:17:40 -08:00
e8347f17b4
minor
2018-02-07 20:16:38 -08:00
24ffccf90d
better TransportRegistry::remove()
...
now with the same complexity as the original...
2018-02-06 18:20:33 -08:00
40868b7b8b
cleanup
2018-02-06 18:15:13 -08:00
c92045620a
oops
2018-02-06 16:54:54 -08:00
fcaeb414a0
termination and error handling changes.
...
epicsAtExit is no longer used to call ca_context_destroy.
It is now called from the destructor for CAChannelProvider.
Several changes were made for handling errors.
2018-02-06 10:20:29 -05:00
5f9508a3b1
demote getTransport() fialure message to debug
...
the connect() will be retried, so this isn't fatal.
2018-02-01 20:04:03 -08:00
d41d15d59e
drop some dead code
2018-02-01 19:51:32 -08:00
606669c62b
less verbose error message when ::connect() fails
2018-02-01 19:51:15 -08:00
9b828852c0
Transport::getRemoteAddress() return reference
...
make it clear that result is always !NULL
2018-02-01 19:09:44 -08:00
f1defe4e9f
mark NamedLockPattern as deprecated
...
just too hard to use (exception) safely
2018-02-01 19:00:38 -08:00
58f996fd98
update transportRegistry
...
avoid leaks of SOCKET and leaving mutex locked
when exceptions are thrown.
2018-02-01 18:54:25 -08:00
5a86835a6f
add ref-counter for Transport ABC
2018-02-01 11:17:56 -08:00
a039b48c1f
caProvider: remove deprecation warnings
...
make no-op destroy() private to prevent
accidental usage.
2018-01-31 12:49:10 -08:00
Michael Davidsaver