RE: Long out record PV Monitoring error in client using CaChannel-1.5 python interface

["Wang Xiaoqiang" <Xiaoqiang.Wang@psi.ch>]

 

So I have just installed SL5.4 x86_64 in a virtual machine and compiled EPICS 3.14.12 and CaChannel-1.5.

Running the script gives the same results: signed number becomes unsigned.

 

This problem is that dbr_long_t is actually epicsUInt32. While in Py_BuildValue "l" formater converts it to 64bit long.

So in the attachment, "i" is used for dbr_long_t. That solves the problem.

 

I will put up a new release soon as possible.

 

Best,

Xiaoqiang

---

From: Ritesh Sugandhi [mailto:ritesh.sugandhi@gmail.com]
Sent: Wednesday, December 01, 2010 2:00 PM
To: Wang Xiaoqiang
Cc: Di Maio Franck
Subject: Re: Long out record PV Monitoring error in client using CaChannel-1.5 python interface

Dear sir ,

thanks for your reply,

The EPICS is build with x86-64 target and the python is 64 bit. 

Now, I will rebuild the CaChannel from source and check the build process and make carefully and return back to you.

Best regards,

Ritesh sugandhi

On Wed, Dec 1, 2010 at 12:47 PM, Wang Xiaoqiang <Xiaoqiang.Wang@psi.ch> wrote:

 

Your EPICS is built with x86_64 taget? Did you build Cachannel from source? Your python is 64 bit?

Your output seems that the value is interpreted as 64 bit instead of 32bit.

 

Best,

Xiaoqiang

---

From: Ritesh Sugandhi [mailto:ritesh.sugandhi@gmail.com]

Sent: Wednesday, December 01, 2010 11:23 AM
To: Wang Xiaoqiang
Cc: Di Maio Franck; pamgurd@shaw.ca
Subject: Re: Long out record PV Monitoring error in client using CaChannel-1.5 python interface

Dear Sir,

thanks for your quick reply. I am using CaChannel-1.5 from http://sourceforge.net/projects/igor2epics/files/CaChannel/ ? and My test system configuration is as below

 

RHEL 5.5. client edition x86 -64

EPICS 3.14.11

Python 2.4.3 

I run your scripts and output is as below:

==========================================================

EPICS output is 

epics> dbpf long 1234

DBR_LONG:           1234      0x4d2               

epics> dbpf long 2147483647

DBR_LONG:           2147483647 0x7fffffff         

epics>  dbpf long 2147483648

DBR_LONG:           -2147483648 0x80000000        

epics>  dbpf long -1

DBR_LONG:           -1        0xffffffff          

epics>  dbpf long -2

DBR_LONG:           -2        0xfffffffe          

a.py script output is 

================================

[codac-dev@trunk ~]$ python a.py 

        2010-12-01 11:09:19

long    2010-12-01 11:09:19     0

long    2010-12-01 11:13:30     1234

long    2010-12-01 11:13:44     2147483647

long    2010-12-01 11:14:07     2147483648

long    2010-12-01 11:14:21     4294967295

long    2010-12-01 11:14:45     4294967294

=================================

Any suggestion will be highly appreciated. do you think that is is 32 bit vs. 64 bit os issue?

Regards,

Ritesh Sugandhi

On Wed, Dec 1, 2010 at 9:56 AM, Wang Xiaoqiang <Xiaoqiang.Wang@psi.ch> wrote:

 

What you mean by "When ever the record attends the –ive values"?

 

Are you using CaChannel 1.5 from http://sourceforge.net/projects/igor2epics/files/CaChannel/ ?

I did a test with your snippet, into which I have added the trivial missing imports. I have attached the script and database

 

My test system is:

  Scientific Linux 5.1 x86

  EPICS 3.14.8

  Python 2.4.3

 

Here is the output. Could you run my attached script and post the output?

 

### IOC Log ####

slslc06> dbpf long 1234
DBR_LONG:           1234      0x4d2
slslc06> dbpf long 2147483647
DBR_LONG:           2147483647 0x7fffffff
slslc06> dbpf long 2147483648
DBR_LONG:           2147483647 0x7fffffff
slslc06>                                  

#### Python Monitor ###

        2010-12-01 09:40:59
long    2010-12-01 09:40:59     0
long    2010-12-01 09:41:16     1234
long    2010-12-01 09:41:33     2147483647

######################

 

 

Best,

Xiaoqiang

 

---

From: Ritesh Sugandhi [mailto:ritesh.sugandhi@gmail.com]

Sent: Tuesday, November 30, 2010 4:53 PM

To: tech-talk@aps.anl.gov

Cc: pamgurd@shaw.ca; Wang Xiaoqiang; Di Maio Franck

Subject: Long out record PV Monitoring error in client using CaChannel-1.5 python interface

Dear Sir,

We are using a client application develop using CaChannel-1.5 epics python interface library. The CA server is exposing a longout record which we are interested in monitoring.  When ever the record attends the –ive values. The client shows  2**32-val instead of value. while the same code works well with ao record type .

The Code is as below:

###################################################

def eventMon(epics_args, user_args):

  global monPV, now

  print monPV, now, epics_args['pv_value']

def main():

  global stop, now

  try:

    monChan = CaChannel()

    monChan.search(monPV)

    monChan.pend_io()

    print now

  except CaChannelException, status:

    print ca.message(status)

  try:

    monChan.add_masked_array_event(ca.dbf_type_to_DBR_STS(monChan.field_type()), \

          None, ca.DBE_VALUE | ca.DBE_ALARM, eventMon)

  except CaChannelException, status:

    print ca.message(status)

  while not stop:

    now = strftime('\t%Y-%m-%d %X\t')

    ca.pend_event(0.5)

###################################################

Best Regards,

Ritesh Sugandhi 

Attachment: CaChannel_src.tar.gz
Description: CaChannel_src.tar.gz