charset conversion error for sql (VAR)CHAR to SQL_C_WCHAR conversion [ODBC181]

Question

charset conversion error for sql (VAR)CHAR to SQL_C_WCHAR conversion [ODBC181]

firebird-automations opened this issue 10 years ago · 10 comments

firebird-automations commented 10 years ago

Submitted by: Jojakim Stahl (jstahl)

Votes: 1

The charset conversion from db cp encoded string to SQL_C_WCHAR (wide char string) mixes up length Information etc.
It's based on the assumption 1 Byte in db cp = 2 Byte in wide notation. This relationship is not valid. For UTF8 1-4 bytes may result in 1 or 2 wide chars = 2 or 4 bytes.

Commits: 9fca7df d94a113

Answer 1 · 2014-11-03T15:16:25.000Z

Commented by: Jojakim Stahl (jstahl)

uploaded patch

Answer 2 · 2014-11-03T15:16:25.000Z

Modified by: Jojakim Stahl (jstahl)

Attachment: bug-181.patch [ 12613 ]

Answer 3 · 2015-06-17T23:37:39.000Z

Commented by: Guy Dion (gdfbu)

I had the following problem:

SELECT LOWER('ÉÈËïÎÀÖÙÛÇ') FROM RDB$DATABASE

correctly returned 'çöùûàéèëïî' with JayBird JDBC driver but returned a truncated string 'çöùûà' using ODBC driver 2.0.3 or 2.0.4 (see issue ODBC184).

I have applied this patch (to the trunk version obtained here http://firebird.cvs.sourceforge.net/viewvc/firebird/OdbcJdbc/) and recompiled using Ms Visual Studio 2013. It has corrected the problem.

Tested with both 32 bits and 64 bits ODBC connection under Windows 7 64 bits.

Answer 4 · 2016-07-05T15:16:31.000Z

Commented by: @alexpotapchenko

Jojakim, can you send final version of OdbcConvert.cpp (OdbcDesc.cpp, IscStatementMetaData.cpp etc.) with all applied patches (160, 161, 162, 180 etc.)?
Some your patches does not apply by turn.

Answer 5 · 2016-07-06T10:45:49.000Z

Commented by: Jojakim Stahl (jstahl)

Hello, I attached the sources which I currently use to build my "own/private" version of the driver which I am using with our application. The sources are base on cvs head from 2015-02-25.
Nice that you are going to integrate the fixes, perhaps I can abandon my fork then.
Greetings