Experimental Physics and
Industrial Control System

Hi Dirk,

Thanks for the detailed information here; you were correct that this is
2 hex strings each 32-bits long. Using the %x format did the trick for me;
although I swear I tried that already but perhaps using a longin record
which as Mark pointed out does not give me an unsigned integer.

Because these numbers represent bit-wise interlocks I like Stephanie's idea
of using an aSub record ... just seems simpler to parse the bits this way
rather than trying to use database records.

Many thanks to all that replied ... the end result is that I used this
protocol:
get_ilock {
	out "din/r";

	in "%(\$1InterlockWords.A)x ",
	   "%(\$1InterlockWords.B)x";
	ExtraInput = Ignore;
}

Where the records are aSub records and was able to print the values to the
console from my subroutine as expected.

Thanks,
Marty

Dirk Zimoch wrote:

Hi Marty,

%R is the floating point raw format. The integer raw format is %r.

(Here I assume that you really get binary data, not the hex string "ffffffff
ffffffff", which would be a case for %x.)

I am a bit confused, because ffffffff ffffffff is 2 32 bit values or 1 64 bit
value.

Stream device can read a 32 bit big endian signed integer with %4r, a 32 bit big
endian unsigned integer with %04r, a 32 bit little endian signed integer with
%#4r and so on.

On 64 bit machines, StreamDevice can even read 64 bit integers with %8r, but no
EPICS record can store them. So you are limited to 32 bit signed or unsigned at
the moment.

In case you do not have binary data but hex strings, (and the space in your
format suggests that) you should use %x.

Dirk


On 12.11.2012 21:21, Martin L. Smith wrote:

I'm trying to read a 32-bit hex value like ffffffff ffffffff using
EPICS R3.14.12.2, streamDevice 2-5, and asyn 4-19 on 64-bit Linux OS.

When I read these values using the following protocol:
get_ilock {
out "din";
in "%(\$1FirstInterlockWord.VAL)8R ",
"%(\$1SecondInterlockWord.VAL)0R";

ExtraInput = Ignore;
}

Both records are longin records and no matter if I use the 8R or 0R I
always get the same result:

ID06ds:FirstInterlockWord 11/12/12 14:03:02.743812335 -2147483648
ID06ds:SecondInterlockWord 11/12/12 14:03:02.743823268 -2147483648

Turns out that this is hex value 80000000

If I use a strngin record I can see that I'm getting the expected data.

Is this the largest value that I can read using streamDevice or is there
some other way that I should try to read these values?

Thanks,
Marty

References:

streamDevice 32-bit hex value read problem Martin L. Smith

Re: streamDevice 32-bit hex value read problem Dirk Zimoch

Navigate by Date:

Prev: Re: question of Hytec 8516 added in synApps-5.3 jim.chen

Next: archive data server James F Ross

Index: 1994  1995  1996  1997  1998  1999  2000  2001  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: Re: streamDevice 32-bit hex value read problem Dirk Zimoch

Next: Re: strange EPICS glitch Eric Norum

Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <2012>  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024