Re: gateway problem

["Noboru Yamamoto" <noboru.yamamoto@kek.jp>]

Hi Ralph, 

I'm not sure if it is same phenomena Yuichi has, but here is 
I confirmed tonight.

My Environment:
EPICS 3.14.9 
gateway 2_0_2_1 ( and 2_0_0_0)
MacOSX 10.4.10 running on IntelMac.

Symptom:

for a waveform record with FTVL=DOUBLE in iocsh.

a) "caput -a" through DIRECT CA access to IOCSH works OK.
b) Python CA( which uses native DBR type ) works OK for both direct
and through-a-gateway access to this record.
c) "caput -a" through a gateway fail.
d) "caput -a -c" through a gateway also fail with "write callback timeout".

IF I changed FTVL of this record to STRING.
All methods, a)-d), works OK.

So, this is consistent with the report from Gasper.

I also read the source code of caput, and found it send data as STRING
for "caput -a". 

This is almost all I know about this issue.

best regards,

Noboru


--- "Ralph Lange" <Ralph.Lange@bessy.de> ---

>Hello Yuichi,
>
>could you be a bit more specific about the nature of the problem you are 
>experiencing?
>
>Regards,
>Ralph
>
>
>Yuichi ITOH wrote:
>> Hi
>>
>> I am trying to use gateway, and I have encountered with a problem.
>> The problem appeares when writing array data through gateway using "caput -a" command.
>> (Gateway2.0.2.1 built with Base3.14.9 on Scientific Linux 4.2)
>>
>> Writing a scalar data through gateway, it will successfully completed.
>> Writing array data without gateway, it will successfully completed.
>> Environment variable EPICS_CA_MAX_ARRAY_BYTES is correctly set, of course.
>>
>> Has anyone else encountered with this problem?
>>
>> Thanks,
>> Yuichi ITOH
>>
>>