Ticket #2339 (closed defect: fixed)

Opened 4 years ago

Last modified 3 years ago

ODBC WVARCHAR Access Lacking

Reported by: warmerdam Owned by: warmerdam
Priority: normal Milestone: 1.5.3
Component: OGR_SF Version: unspecified
Severity: normal Keywords: odbc WCHAR NVARCHAR UTF-8
Cc:

Description

Currently SQL Server NVARCHAR and similar native language fields are requested as SQL_C_CHAR by OGR which appears to trigger a translation to the current locale's encoding. If the field contains something like Arabic text this results in the text being corrupted (mostly turned to '?' characters).

The desirable solution would be to fetch the text as UCS-2 wchar_t and translate to UTF-8 for returning from OGR.

This bug is broadly address by rfc23_ogr_unicode but a more narrow fix will be required for OGR 1.5.2.

Change History

Changed 4 years ago by warmerdam

  • status changed from new to assigned

I have prepared a preliminary update to cpl_odbc.cpp that is MSVC only, and fetches WCHAR type fields as wchar_t and then converts to UTF-8 using the win32 WideCharToMultiByte?() function. This is committed in 1.5 branch (r14361).

This is intended to be a narrow (non-disruptive) change for the stable branch while trunk will get a broader treatment based on rfc23_ogr_unicode but using a similar approach at the ODBC level.

Changed 3 years ago by warmerdam

  • status changed from assigned to closed
  • resolution set to fixed

RFC 23 has been implemented in trunk for the ODBC driver. No further action required I think.

Note: See TracTickets for help on using tickets.