Vincent Rogier

Hi, Can you provide DDL/DML and sample code for it ? I cannot reproduce it on Windows (19c client) nor Linux (19c client) with a Oracle 19c server (on Windows)....

Hi, @AidyCC have you been able to test with @cjbj suggestion ? In case it did not change the outcome, can you provide a sample code ? Thanks, Vincent

Hi, thus, on your environment, the following crashes whatever the variable ORA_OCI_NO_OPTIMIZED_FETCH is set (I was not aware of this varialbe until Chris mentionned it) ? OCIRowidToChar(rowid, NULL, &rowid_length, def->rs->stmt->con->err);...

Ok, at least the symbol OCIRowidToChar is present. I do not have access to a MacOS... In ocilib, in resultset.c line 1801: OCIRowidToChar(rowid, NULL, &rowid_length, def->rs->stmt->con->err); Can you add a...

Thanks. There is something I don't understand. In previous test you made in ANSI, you had a wierd output "E...". This time, with same code base (except the printf), you...

Ok. If I summarize: - in ansi mode, it works ok - in unicode mode, ocirawidtochar() crashes @cjbj it seems to be an issue with this method when oci environment...

One last test you can do, it is to provide a valid local buffer, let's say a unsigned char buffer[100], go the first call to the function instead of passing...

Hi, Thanks. It is normal that it works as you're not initializing OCI in UTF16 mode. When using OCI_CHARSET_WIDE, ocilib passes the flag OCI_UTF16 to OCIEnvCreate() 2nd parameter 'mode'. Vincent

Hi, @AidyCC there are several issues in the raw oci code you've attempted to provide. Mismatch between char/wchar_t, string length computing, row id that are always char* even in UTF16...

@AidyCC With the code I have provided, do you encounter the issue, in your environment, when the macro UTF16 is defined ?