Conference 2.301::cimfast

Hi,

>>  We are using CIMfast V2.2 and BASEstar Open Server For Windows NT (3.51)
Can i find out where you got the kit from (and when)??

I believe this is a known problem, and a fixed kit can probably be made
available early next week.


>>PRINT - Expression creation failed
>>PRINT - Invalid expression function name found
>>PRINT - Function was: STR_LENGTH
This error is a result of the CIMfast searching it's expression table and not
finding the STR_LENGTH expression.

Dick Long is currently supporting CIMfast on NT, and recently he found a
problem with this search on NT.  It is a binary-tree type search, but he found
the  STRCMP function used by C on NT only handles alpha-numerics correctly
(thus placing _ characters in the wrong search order), so sometimes valid
expressions were not being found in the table, and resulting in this error.

I'm not sure if he has a full kit built yet... he i currently at the BASEstar
User's group meeting in Orlando Florida, and will be back monday...
I'll mail this to him, and see if he can give us any pointers
to a new kit... otherwise you can expect one early next week, i'm sure.

thankx,
sarah

The first kit is if you are NOT using ORACLE, the second is is you are:


QUAKKS::DISK_LONG:[LONG]

CIMFAST_KIT.ZIP;1       2588  10-FEB-1997 12:02:44.12  (RWED,RWED,,RE)
CIMFAST_SQL_KIT.ZIP;1
                        2732  10-FEB-1997 12:02:36.14  (RWED,RWED,,RE)

    Sarah,
    
    Please, make sure the files are not protected. I'm not being able to
    access them.
    
    Cyro

>>    Please, make sure the files are not protected. I'm not being able to
>>    access them.
sorry..

i have moved them to cam::sys$public:

NEP2UN::SJ$DISK:[JOHNSON]> di cam::sys$public:*cim*.zip/sec

Directory CAM::PUBLIC:[PUBLIC]

CIMFAST_KIT.ZIP;1       2588  10-FEB-1997 12:02:44.00  11-FEB-1997 07:40:01.00
     [300,0]          (RWED,RWED,RWED,R)
CIMFAST_SQL_KIT.ZIP;1
                        2732  10-FEB-1997 12:02:36.00  11-FEB-1997 07:40:46.00
     [300,0]          (RWED,RWED,RWED,R)

Sarah,

I have copied the CIMFAST_SQL_KIT.ZIP file you made available to us.

We are trying to use STR_LENGTH function, but got some non-expected results, 
as described herein.

We have created a B* data point using a datatype defined as a fixed length (80) 
array of visual chars. We also have defined a CIMfast variable of string type.
Then we have set B* data point to a known string, using B* CLI command, for 
example, "CF1000".

In the CEL program we set the CIMfast variable with the B* data point and then
we ask to print STR_LENGTH of the CIMfast variable. The result is 80. We were 
expecting 6. We also have asked to print the content of CIMfast variable and 
observed it is set to CF1000. But we have observed that blanks are added, or 
maintained, at the end of the string. We supposed this was caused because we 
have defined the B* data point as a fixed lenght array. Then we made another 
try, redefining the B* datapoint as a variable lenght array. The unexpected 
result is that we got a message saying CIMfast does not support variable 
lenght array.

Are we making any conceptual error?

Thanks,

Cyro

Hi,

>>maintained, at the end of the string. We supposed this was caused because we 
>>have defined the B* data point as a fixed lenght array. Then we made another 
>>try, redefining the B* datapoint as a variable lenght array. The unexpected 
>>result is that we got a message saying CIMfast does not support variable 
>>lenght array.
>>Are we making any conceptual error?

no error on your part... CIMfast does not currently have the variable string
datatype... 
you can create your own data type to map B* variable strings onto...
you can see the bcf_create_struct_sql.com and bcf_sql.cel as examples...
attached is a note i found in the qar conference that has a different
way of mapping the data type (dcm and B* OPEN treats variable strings the same
way... the first word is the length, followed by the data).

hope this helps,
sarah

             <<< CAM::$1$DUA5:[NOTES$LIBRARY]CIMFAST_QARS.NOTE;1 >>>
                           -< BASEstar CIMfast QARs >-
================================================================================
Note 30.1                           DCM_WRITE                             1 of 1
EKAS::EGERTON                                        17 lines  13-MAR-1992 16:40
--------------------------------------------------------------------------------

Miquel,

	This is because DCM's idea of what a string is conflicts with CIMfast's.
DCM expects a varying length string when you specify the format string.  Be
warned that the same sort of thing will happen in reverse if you try to read
a string.  The following should work okay to write to the device:


#define VARYING_STRING(str) (buffer_create (word (str_length str)) str)

on init_event;
   dcm_write device "DN_EKAS"
             address "0"
             format "string:7"
             value VARYING_STRING("CIMfast");
   exit;

Sarah,

We are using the CIMfast/SQL for WNT (intel) you made available to us on
Feb,10th (see note 205.4).

At this moment we are testing it more intensively and we are observing it is
presenting some non-expected results. Just to be sure we are not making a 
conceptual error, we decided to take the demo program BCF_SQL.CEL included in 
the CIMfast kit.

We modified the demo program in order to insert one additional record, with 
the following fields contents:

NAME: cachorro12
BREED: abcdefghijkl
AGE: 876543

We are observing that at first print, the contents of the field name is 
printed correctly, i.e. with ten characters. This printout is the result of

print (sql_fetch "names"); 

However, after that, for example,as the result from the query to find all dogs
over 4 years age, the name and breed fields are being truncated: the last 
character is lost. And the numerical field (AGE) returns a wrong value.

We have observed that problems still persists with strings. We have created a
B* Open data_point of data_type with 10 characters. Reading this point with 
CLI command (get value) the value is correct. However, reading it from a 
CIMfast program, only nine characters are being returned.

Please, could you check these points and give me a feedback.

Cyro