EPICS "changes" to ChannelProvider::createChannel()

Experimental Physics and Industrial Control System

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 <2018> 2019 2020 2021 2022 2023 2024	Index	2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 <2018> 2019 2020 2021 2022 2023 2024
<== Date ==>		<== Thread ==>

Subject:	"changes" to ChannelProvider::createChannel()
From:	Michael Davidsaver <[email protected]>
To:	EPICS core-talk <[email protected]>
Cc:	PVData Developers <[email protected]>
Date:	Fri, 18 May 2018 11:16:44 -0700

Unless you, or your site, has written a sub-class of epics::pvAccess::ChannelProvider you
can safely ignore the following.


First a short PSA.  I'm in the process of adding a server API to P4P (PVAccess 4 Python).
This is causing me to investigate the PVA server code with an eye towards clarifying
the semantics of how a ChannelProvider must behave.

For example, I previously mandated that ChannelProvider::createChannel() must be
synchronous[1].  That is, it can only return NULL on error, and must call
ChannelRequester::channelCreated() before returning.

So far the PVA server hasn't actually required this, and has ignored the return of
createChannel() and allowed delayed creation (and search via channelFind()).
However, there is an issue[2] with adjudication when a server has multiple
ChannelProviders, and/or when a process has multiple servers.



To address this I'll probably have to *start enforcing* that channelFind() and createChannel()
be synchronous for server providers as well.



This means that any ChannelProvider which attempts to be asynchronous, or simply forgets
to call channelCreated(), will see the request fail with an error.

I don't think this will break existing code as all the provider implementations which I
checked are already synchronous (PVA client, caProvider, pva2pva, and pvDatabaseCPP).

If there are any implementations I can't see which depend on this behavior, please reply,
or otherwise contact me.

I'll follow up this mail when I have something which can be tested against.


[1] docs for epics::pvAccess::ChannelProvider::createChannel()
http://epics-base.github.io/pvAccessCPP/classepics_1_1pv_access_1_1_channel_provider.html#a52fcce45922422d812ba9f4bcef7612c

[2] https://github.com/epics-base/pvAccessCPP/issues/100

Navigate by Date:: Prev: Jenkins build is back to normal : epics-base-3.14-cyg64 #235 APS Jenkins; Next: Re: Size encoding Michael Davidsaver; Index: 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 <2018> 2019 2020 2021 2022 2023 2024
Navigate by Thread:: Prev: Jenkins build is back to normal : epics-base-3.14-win64s #245 APS Jenkins; Next: Jenkins build is back to normal : epics-base-3.14-win32 #279 APS Jenkins; Index: 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 <2018> 2019 2020 2021 2022 2023 2024