Re: Epics 7.x.x.x and AD R3.x

[Mark Rivers <rivers@cars.uchicago.edu>]

Hi Mark,


Michael is right, for Windows not only do you need the master branch of pvaDriver, you also need some fixes to EPICS 7.


We have found a number of problems, some of which are specific to the version of Visual Studio, some of which depend on static/dynamic, and some of which depend on debug/no debug.  One of these problems, with the windows-x64-debug architecture with Visual Studio 2010 I found (and Michael fixed) just today.


If you want to use EPICS 7 on Windows right now I suggest using the "master" branch, which is actually called core/master or master, depending on the specific module.  Presumably there will be a new release of EPICS 7.x soon that fixes these problems.


Mark




________________________________
From: Michael Davidsaver <mdavidsaver@gmail.com>
Sent: Tuesday, February 6, 2018 7:33 PM
To: engbretson@anl.gov; Mark Rivers; 'EPICS Tech Talk'
Subject: Re: Epics 7.x.x.x and AD R3.x

On 02/06/2018 03:03 PM, Mark Engbretson wrote:
>
>
> Mark,
>
>
>
> At some point I thought that you had said that you had built epics 7 and a recent version of Area Detector and it all worked.

He did, after some windows specific fixes.  In this case for https://github.com/epics-base/pvDataCPP/issues/49


> Can you correct me if this is an incorrect memory?
>
>
>
> Asking since I had been trying to compile Epics base 7.0.1.1 and AD R3-1 and R3-2 and while everything compiles, all my AD softioc's crash and burn if I build them against the pvadriver-1-3 with a null this-> pointer when the pvadriver is instantiated attempting to create a mutex lock.    Happen with MSVC 2017 and 2010. And happens if I used EPICS-CPP-4.7.0. and not the versions in epics 7.
>
>
>
> Note sure if I should be looking in the epics base code for an issue of if this has to be in the pvadriver since everything else seems to work if I don't build against it.
>
>
>
>
>