attached io in HasIO must not be mandatory - either uri or io
has to be given
Change-Id: Id39e40f98020d4051c1ad8105f6af6018aafaea8
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/29349
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
When the io of one SECoP module fails, it takes ages to startup
because each parameter poll takes the time to wait for a timeout.
After the first communication error on an io, no more startup polls
are tried on the modules using this io.
Change-Id: I0d250953dfe91a7d68d2d2b108395cc25d471afe
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/28588
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
- repeated errors on poller are only once logged (per poll
function / read_* method)
- during exception handling, silent=True on a SECoP error indicates
that the error is already logged
+ fix the name of HardwareError
+ add test for consistency of SECoPErrors
+ catch socket.timeout in AsynTcp
Change-Id: I9df6c775cc19553b22a4d6e39591092adf7ff9a1
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/28139
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
When several poller threads are using the same io, the resposivity
of client requests is reduced, as every thread first finishes
its pending communication requests, before it is the turn of the
request thread. This is solved by using one common poller thread
for all modules sharing the same communicator.
+ fix an issue with overriding a property with a parameter, as
this is the case for pollperiod (cfg was applied to property
instead of overriding parameter)
+ separate setFastPoll arguments into flag and fast interval
+ fix missing announceUpdate call when read function fails
+ fix mechanism for triggering polls after an io connection
reconnected again.
Change-Id: I1115a61fae3de80d18416e61f40b52a0eebb637c
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/28021
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
revert basically the former change
"init generalConfig.defaults only in secop-server"
The problem of import order when setting generalConfig.defaults
has to be solved by not overriding already existing keys when
setting the default.
Change-Id: I82121e346607dd74146279c4241e13ab63c14096
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/28011
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
generalConfig.defaults must not be set on import, as this
depends on import order
Change-Id: I00395b40b4281ddc044c196713f6512068011380
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27985
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
- fix some earlyInit and initModules methods
- remove some comments
- change name of Done unique value to 'Done', this seems more
useful for __repr__ and debug logging
Change-Id: I73f0e09bfef858ddca11bba0e92e941ebc151160
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27907
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
- remove secop.poller and basic poller
- regular polls for 'important' parameters done by method doPoll
- all other parameters are polled slower (slowInterval) and
with lower priority (only one at a time when main poll is due)
- nopoll decorator for read_* to disable poll
- enablePoll attribute (default True) for disabling polling a module
- fast polls may be implemented by means of a statemachine
- configurable slow poll interval
+ allow a Parameter to override a Property (parameter
Readable.pollinterval overrides Module.pollinterval)
Change-Id: Ib1b3453041a233678b7c4b4add22ac399670e447
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27832
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
- setting the attribute using the name of an attached module
- getting the attribute results in the module object
+ change names iodev to io, iodevClass to ioClass,
sendRecv to communicate, HasIodev to HasIO
Change-Id: I200b63a5a7dc1453bf6ac998782b065645201900
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27575
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
A common problematic practice is, to declare the value parameter
with the same FloatRange than the target. Because of measurement
errors, a value might be near, but outside the limit.
In order to avoid this, we force the programmer to declare a
bigger range for the value than for the target, or to
explicitly disable this check on a module property.
It is also fine to declare the value without limits.
This behavior may be disabled via command line option or in the
general config file. For simplicity, FloatRanges inside data
structures are not considered.
+ above command line option is also used to disable the error
handling on a string to float conversion
+ log appropriate error message for string to float conversion
Change-Id: Ib78ea1fb7c821442bf5847030573c8c27822dea5
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27574
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
modules with a couple of parameters with similar read_* or
write_* methods may handle them by generic methods wrapped
with decorators ReadHandler / WriteHandler
The trinamic driver is included in this change for demonstrating
how it works.
In a further step, the special handling for the iohandler stuff can
be moved away from secop.server and secop.params, using this feature.
+ fix problem on startup of trinamic driver (needs MultiEvent.queue)
+ some other small fixes
+ apply recommended functools.wraps for wrapping
Change-Id: Ibfeff9209f53c47194628463466cee28366e17ac
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27460
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
- bin/secop-server options -v and -q applied to console logger only
- level for logfile taken from general config
- option for automatic deletion of old logfiles
- added 'comlog' level (between debug and info)
This allows to run the servers by default with 'comlog' level on
the logfiles, which helps a lot for analyzing very rare communication
errors in retrospect.
to avoid spamming of the normal log files, comlog data is stored
separately, one file per communicator
+ redesign of remote logging (no more need of LoggerAdapter)
Change-Id: Ie156a202b1e7304e50bbe830901bc75872f6ffe2
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/27427
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
rename secop.stringio to secop.io, which includes now
also BytesIO and the common base class IOBase
+ a small fix in error handling
Change-Id: I8e305e2c164f4ed131f4b36ef45edd8bd222336d
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/26393
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
