y-shashank / ruby-odbc Goto Github PK

- Ruby 1.[6-9], MySQL 3.22/MyODBC (local), unixODBC 2.1.0
  on Linux 2.2-x86 and 2.6-x86_64

- Ruby 1.6.4, MySQL 3.22/MyODBC (local), libiodbc 2.50
  on Linux 2.2-x86

- Ruby 1.[6-8], MySQL 3.22/MyODBC (remote), MS Jet Engine, MSVC++ 6.0
  on Windows NT4SP6

- Ruby 1.6.[3-5], MySQL 3.22/MyODBC (remote), MS Jet Engine, cygwin,
  on Windows NT4SP6 and 2000

- Ruby 1.8.*, SQLite/ODBC >= 0.67, libiodbc 3.52.4 on Fedora Core 3 x86

- Ruby 2.0.0, SQLite/ODBC >= 0.93, unixODBC 2.2.14 on Ubuntu 12.04 x86

- Ruby >= 2.4

Michael Neumann <neumann @nospam@ s-direktnet.de> and
Will Merrell <wmerrell @nospam@ catalystcorp.com> reported successful
compilation with Cygwin on Win32.

- Ruby 1.6.[3-8] or Ruby >= 1.7
- unixODBC 2.x or libiodbc 3.52 on UN*X 

$ ruby -Cext extconf.rb [--enable-dlopen|--disable-dlopen]
$ make -C ext
# make -C ext install

--enable/disble-dlopen turns on/off special initialization
code to make ruby-odbc agnostic to unixODBC/iODBC driver
manager shared library names when GCC is used for compile.
In cases where unixODBC or iODBC is installed in non-standard
locations, use the option --with-odbc-dir=<non-standard-location>
when running extconf.rb

$ ruby -Cext/utf8 extconf.rb [--enable-dlopen|--disable-dlopen]
$ make -C ext/utf8
# make -C ext/utf8 install

C:..>ruby -Cext extconf.rb
C:..>cd ext
C:..>nmake
C:..>nmake install
C:..>ruby -Cutf8 extconf.rb
C:..>cd utf8
C:..>nmake
C:..>nmake install

$ ruby -Ctest test.rb DSN [uid] [pwd]

Refer to doc/odbc.html

The difference between utf8 and non-utf8 versions are:

  - non-utf8 version uses normal SQL.* ANSI functions
  - utf8 version uses SQL.*W UNICODE functions and
    requires/returns all strings in UTF8 format

Thus, depending on the -K option of ruby one could use
that code snippet:

  ...
  if $KCODE == "UTF8" then
    require 'odbc_utf8'
  else
    require 'odbc'
  fi

It is also possible to load both non-utf8 and utf8 version
into ruby:

  ...
  # non-utf8 version
  require 'odbc'
  # utf8 version
  require 'odbc_utf8'

Whichever is loaded first, gets the module name 'ODBC'.
The second loaded module will be named 'ODBC_UTF8' (for
'odbc_utf8') or 'ODBC_NONE' (for 'odbc'). That should
allow to use both versions simultaneously in special
situations.

- heavier testing
- improve documentation

Christian Werner
mailto:chw @nospam@ ch-werner.de
http://www.ch-werner.de/rubyodbc