ADO RDD xHarbour

[AHF]

Antonio,

It seems that in usrrdd.c are missing some functions like FieldSize FieldLen Dbrlocklist.

Shouldnt these also be there?

[Antonio Linares]

Antonio,

What this needed for:

Code: Select all

STATIC FUNCTION ADO_RAWLOCK( nWA, nAction, nRecNo )

// LOCAL oRecordSet := USRRDD_AREADATA( nWA )[ WA_RECORDSET ]

   /* TODO */

   HB_SYMBOL_UNUSED( nRecNo )
   HB_SYMBOL_UNUSED( nWA )
   HB_SYMBOL_UNUSED( nAction )

   RETURN HB_SUCCESS

It does nothing but it is needed. In case that you need it to do something, then you can implement it.

[Antonio Linares]

Antonio,

You should download the complete source code for the harbour RDDs and use it to review the RDD code:

https://github.com/harbour/core/tree/master/src/rdd

[AHF]

Antonio,

I dont have any c expert and even so it would be dificult to understand it without proper documentation.

In order to proceed I need the following functions working:


OrdFor()
FieldSize() or FieldLen()
FieldDec()
DbrlockList()

I'm in a dead end ! What do you propose?

[Antonio Linares]

Antonio,

Its not a matter of knowing C, instead it is a matter of properly searching

You asked for OrdFor().

1. Is there a UR_FOR or similar in harbour\include\hbusrrdd.ch ?

I find this:

/* DBSCOPEINFO */
#define UR_SI_BFOR 1
#define UR_SI_CFOR 2

Also:

/* DBORDERCONDINFO */
#define UR_ORC_ACTIVE 1
#define UR_ORC_CFOR 2

2. Have we implemented those functions in adordd.prg ?

3. In SQL we may need to store a string cFor value only, no need for the codeblock.

I know this is not easy, but slowly we can keep going...

[AHF]

Antonio,

1. Is there a UR_FOR or similar in harbour\include\hbusrrdd.ch ?

As far as I see no.

3. In SQL we may need to store a string cFor value only, no need for the codeblock.

Its absolutly necessary to build the selects respecting indexes order and conditions.
If the index is created (temporary) then on ADO_CREATE we have all info we just save it for later use.
If the index its usual then it must be defined previously in the index array.
So the problem its only with ordfor() to get the values saved.

2. Have we implemented those functions in adordd.prg ?

What do you mean by those functions ?

I can have such a function(s) but then I will disrupt other linked rdds, besides that since we dont have nWA we can not use it like (alias)->(ordfor(2))

Code: Select all

FUNCTION OrdFor(nOrder)

    //works only for current selected  workarea not alias->(ordfor(n))

    LOCAL cTablename := USRRDD_AREADATA( Select() )[ WA_TABLENAME ]
    LOCAL aWAData := USRRDD_AREADATA( Select() )
    LOCAL cFor := ""
    
    IF nOrder = 0
    
       nOrder := aWAData[WA_INDEXACTIVE]
       cFor := aWAData[WA_INDEXFOR][nOrder]
       
    ELSEIF nOrder <= LEN(aWAData[WA_INDEXES] )
    
       cFor := aWAData[WA_INDEXFOR][nOrder]
       
    ENDIF   
    
 RETURN cFor
 

I think all mentioned functions must be defined at rdd level to be used correctly as they probably are in dbfcdx rdd for ex.

The adordd its working already with "indexes" with zero app code change and if you get an error you will have all area info.

Code: Select all

 20:    ENCCLIST                           NomeRDD: ADORDD
     ==============================
     RecNo    RecCount    BOF   EOF
        1         34      .F.   .F.

     Índices em uso                        TagName
        => NRENCOMEND+CODIGOPROD+ARMAZEM      ENCCPRO
           CODIGOPROD+ARMAZEM+NRENCOMEND      PROENCC
           GUIA+CODCLIENTE+CODIGOPROD+ARMAZEM MECPD
           CODCLIENTE                         CDENCCL
           NRFACTUR+CODCLIENTE+CODIGOPROD+ARMA PTFACT
           NRFACTUR+CODIGOPROD+ARMAZEM+ANO +SE NRFPROD
           CODIGOPROD+NRCONTRATO              CCONTRT

     Relações em uso 
        Registos fechados 

 30: => OFERTAS                            NomeRDD: ADORDD
     ==============================
     RecNo    RecCount    BOF   EOF
        1          3      .F.   .F.

     Índices em uso                        TagName
           CODCLIENTE+CODIGOPROD+ARMAZEM+DATAO CLIPRO
           CODIGOPROD+ARMAZEM+CODCLIENTE+DATAO PROCLI
           CODCLIENTE+DATAOFERTA +NUMEROOFER+C  CLIPRO2
           CODIGOPROD+ARMAZEM+DATAOFERTA +NUME PROCLI2
        => CODCLIENTE                         TEMP2

     Relações em uso 
        Registos fechados 
 

Dont get me wrong its not a matter of difficulty but if its in fact possible to work with our approach idea and that we are not simply loosing time as we are trying to join together two completly diferent worlds xBase and SQL.
Besides that we have till June to have such a solution.

Is there some rdd expert that could help us? How can we solve this? Can you check how its working in other rdds?

[AHF]

Antonio,

I never looked at .ch.
I understand now how it works.

Ive define UR_INFO and now I get the call from dbrlocklist do get the array.

We re back again

[Antonio Linares]

Antonio,

Very good!

[Antonio Linares]

Antonio,

if you post here or provide a url to download your most recent version, I guess many users may be interested
in testing it and surely will provide a great feedback

[AHF]

Antonio,

Ill do that maybe today after finish all locking, seeks and finds.

Dbrlocklist() and Dbinfo() are working
.
Ive checked the paramaters in hbusrdd.c

Code: Select all

hb_usrInfo( AREAP pArea, HB_USHORT uiInfoType, PHB_ITEM pInfo )

Can I assume that when we have pSomeVar its always a pointer thus passed by reference?
Thus in this case ADO_ORDERINFO(nWa, nInfo, uReturn) uReturn passed reference

The Recno() only works with :Bookmarks as:

Code: Select all

         nRecno := oRecordSet:BookMark
         // iif( oRecordSet:AbsolutePosition == -3, oRecordSet:RecordCount() + 1, oRecordSet:AbsolutePosition )
 

The commented line was in previous adordd but I think its safer like this. Do you agree?

[Antonio Linares]

Antonio,

yes I think it is fine.

You are doing a great work, congratulations!

[Antonio Linares]

Once you complete it, you could provide it to Harbour dev team so it will get updated in the Harbour official repository
and you will become an official Harbour contributor

[AHF]

Antonio,

In order to finish network and concurrent activity I need to work with transactions.

For that I need to know if in ADORDD each time we open a new recordset its in the same connection.

Its crucial to know this because transactions are per connection.

Do you have any idea how to test that?

[Antonio Linares]

Antonio,

AS far as I know I would say yes, but Mr. Rao is the right one to confirm us this

[AHF]

Antonio,

Ive already ask him in my oyher post ADO exclusive open.

It seems that in the source of tdatarow we could check that.