Zeoslib Portal

Hello Mark, as you already know: The upper compoent layer is not related to the dbc layer. From my POV i'ts better if such conversion (user format) happen on the compoent-layer only. Reason is quite simple: Each DataSet, the more each field, can have it's own format. It simply seems to be impossible...

Hello Louis,

i can't see a reason for an ShortIntField for firebird. FB has no byte range types. Can you give more details or an example again?

The component and dbc layer are not related. That's why the dbc layer raises EZSQLException descendants and the component layer uses EDatabaseError descendants. It's an alltime design i love to keep. Inbetween loads of people are using ZDBC only because of the advertising from the mORMot project or ...

Hi,

nope i'm fine with FieldType Bit on ODBC and OleDB. It should be mapped to a boolean field.
Note for MSSQL Server use ODBC_W this server isn't ansi based anymore since ages..

So can you explain the problem a bit more precise? Compiler, Server, Driver Versions should be a good start..

Hello Luis,

welcome on V8. Be patient we'll get you in. According your proplem can you create a min example and open a ticket on https://sourceforge.net/p/zeoslib/tickets/ A bit more code would be nice to resolve the problem.

Hi, @aehimself what was the reason for your pull request? I don't get it. Usually users are setting the param-Type to lock this type. Your patch did change the log vice versa. Can you explain it? @Mark i did change that again: https://sourceforge.net/p/zeoslib/code-0/7518/ . It should also handle ht...

@Michael, I understand that the LOB "descriptors" are required and therefore normal to exist, but wouldn't it make sense to free them up if we navigate away while not caching? I do not see a regression here. All escriptors should be freed is the descriptors have been copied. See TZAbstrac...

Hi, yet i have not enough time for the forum guys, my appologies for that. So just some quick answers: Glad to hear Mark's implementation did not cause any unwanted side effects. The more NICE job, Mark! Thank you. For the record: A "Firehose" cursor can be acquired by using TZReadOnlyData...

Mark, had some test regression (if the OCILobs are NULL). Fixed by https://sourceforge.net/p/zeoslib/code-0/7419/ . According the compatibility question: 1. I have no objections if the enumerator will be added to the TZAbstractRODateSet descendants (except TZMemTable -> late fetched lobs are not sup...

Hi, usually you should see all error messages from IB/FB except the AddLogMsgToExceptionOrWarningMsg is turned off. see: TZInterbaseFirebirdConnection.HandleErrorOrWarning for i := Low(InterbaseStatusVector) to High(InterbaseStatusVector) do begin AppendSepString(ErrorString, InterbaseStatusVector[i...

Hi,

the buffer/rowsize calculation happens in ZDbcOracleResultSet.pas method: procedure TZOracleResultSet_A.Open.

It would be nice if you could resolve that. And if you could resolve the UTF16 missue i would be happy too. Looking forward i can't reproduce both of them.

Hi naumov,

yes please open a new ticket. And please provide an example. Than it's quickly fixed.

Patch applied, thank you.

Jan, there is no emulation about "autocommit" for Oracle. It's simply not supported see: https://docs.oracle.com/cd/E11882_01/ap ... LNOCI17163 patameter "mode". we just send OCI_COMMIT_ON_SUCCESS if the connection is in AutoCommit mode.

Yet i had no time to continue Mark. Thanks for the ticket approach. Anyway it should be acceptable to update the documentations according this "unexpected" behavior.
I keep an eye on that from Zeos side...