Re: Base R7.0.2.2 VME_IO Parsing

[Michael Davidsaver via Tech-talk <tech-talk@aps.anl.gov>]

On 9/23/19 7:54 AM, Graham Cox - UKRI STFC wrote:
> Thanks for the clarification Michael.
> 
> We are using the PSI branch of mrfioc2 (due to our use of the MRF VME-300 series boards), so I guess this is why we don't see the change to INST_IO.

No.  This change was a side-effect of my later adding support for the cPCI-EVG-300 on Linux/x86.

> This brings up the point discussed at the timing workshop of ICALEPCS 2017 in Barcelona about efforts to merge the PSI and 'vanilla' branches of mrfioc2.  Has any progress been made on this, or is this unlikely to happen in the (nearish) future?

While interest was expressed, it has not been followed up with effort.

At this point I think it would be easier to extend my support for the cPCI-EVG-300
and mTCA-EVM-300 to include the VME variant.


> Thanks,
> 
> Graham
> 
> -----Original Message-----
> From: Michael Davidsaver [mailto:mdavidsaver@gmail.com] 
> Sent: 23 September 2019 15:12
> To: Cox, Graham (STFC,DL,TECH)
> Cc: tech-talk@aps.anl.gov
> Subject: Re: Base R7.0.2.2 VME_IO Parsing
> 
> On 9/23/19 2:48 AM, Graham Cox - UKRI STFC via Tech-talk wrote:
>> Is this behaviour an intentional change from Base 3.14->7 ?  Or am I missing something here?
> 
> Yes, back in 2015 as part of my work on linking and locksets the parser was made stricter.
> The numbers became non-optional, were required to be decimal.  This was released in
> 3.16.0.1 (in 2016).
> 
> Also on the subject of updates.  Around the same time I switched mrfioc2 from VME_IO
> links to INST_IO.
> 
>> record(bo, "$(P)ResetMxc-Cmd") {
>>     field( DTYP, "Obj Prop bool")
>>     field( OUT , "@OBJ=$(OBJ), PROP=Reset MXC")
>>     field( DESC, "EVG Reset MXC")
>>     field( PINI, "YES")
>>     field( VAL, "1")
>>     field( UDF, "0")
>> }
> 
> https://github.com/epics-modules/mrfioc2/blob/1e703ba49944cd2137cd04641b80195347d1a404/evgMrmApp/Db/evgMrm.db#L35-L42
>