Installing ---------- Prerequisites: GHC >= 6.6, Cabal >= 1.1.6, filepath (it's possible to use Takusen with GHC-6.4 and Cabal-1.1.4; see notes in separate section below) To run or build Setup.hs you will need filepath installed. This is only needed by Setup.hs, and is not required by Takusen itself. Ensure that the database libraries for any particular database you plan to use are in your path e.g. ensure that %ORA_HOME%\bin is in your path. If the library is not in your path, then it won't be detected in the configure step, and Takusen won't be able to use it. You can fix this by adding it to your path and going through the configure/build/install cycle again. Typical build, after unzipping the distribution archive (Takusen-?.?.gz): $ ghc --make Setup $ Setup configure $ Setup build $ Setup install Typical build, using darcs to get latest code: $ mkdir takusen $ cd takusen $ darcs get http://darcs.haskell.org/takusen $ ghc --make Setup $ Setup configure $ Setup build $ Setup install Using i.e. Writing data access code ----------------------------------- There are extensive instructions and examples in the Haddock docs for module Database.Enumerator: http://darcs.haskell.org/takusen/doc/html/Database-Enumerator.html This should give you most, if not all, of the information you need to create a program that uses Takusen. Here's a little hello-world test case that uses Sqlite: {-# OPTIONS -fglasgow-exts #-} {-# OPTIONS -fallow-overlapping-instances #-} module Main where import Database.Sqlite.Enumerator import Control.Monad.Trans (liftIO) main = flip catchDB reportRethrow $ withSession (connect "sqlite_db") (do let iter (s::String) (_::String) = result s result <- doQuery (sql "select 'Hello world.'") iter "" liftIO (putStrLn result) ) If this is Main.hs, then $ ghc --make Main -o hello should build a "hello" executable. Paths, GHCi & runhaskell ------------------------ Just as with ensuring that your path is correctly set when building Takusen, you must also ensure it is correctly set when building your programs. If it is not correct, then you are likely to see linker errors. Note that the Cabal build script detects which back-ends are installed by looking for certain executables in your path. The script modifies the package description in the build stage so that it only builds libraries for the back-ends it detects. If you install a new back-end, you will need to re-run the installation process in order for Takusen to use it. PostgreSQL gotchas on Windows ----------------------------- The PostgreSQL client library is called libpq.dll on Windows, rather than the more typical pq.dll. This is fine when using ghc to compile, as gnu ld is able to figure out that this is the library to use when you pass it -lpq, but ghci is not quite so slick, and it fails to load the library. There is any easy workaround, which is to copy libpq.dll and rename it to pq.dll. If you do this, then you should be able to use ghci with PostgreSQL without problems. Don't forget to ensure that PostgreSQL's bin is in your path. In the past I've had problems with older versions of PostgreSQL and ghc-6.4.1. Specifically, the call to PQprepare would segfault. This occured with C programs (i.e. no Haskell) compiled with the gcc and ld that came with ghc-6.4.1. If ld (version 2.15.91) was replaced with an older one (2.13.91) from MSYS then the test program worked. With PostgreSQL 8.1.5.1 and ghc-6.6 (gcc 3.4.5 (mingw special) and ld 2.15.94) this problem seems to have vanished. Oracle gotchas on Windows ------------------------- Some users have reported linker errors because their Oracle bin contains hsbase.dll, which is an Oracle library related to Heterogenous Services. This DLL overrides GHC's HSbase.dll, and therefore causes linker errors. If you can control your Oracle client installation then either - don't choose Heterogenous Services when you install, or re-run the installer and remove it, or - rename hsbase.dll in Oracle bin. GHC-6.4 and Takusen ------------------- It is possible to use Takusen with GHC-6.4. We have tested with Cabal-1.1.4, so you aren't required to upgrade to 1.1.6 (but it won't hurt, either). If you use Cabal-1.1.4, then you should use the Setup-114.hs script. If you use Cabal-1.1.6, then you should use the normal Setup.hs script. You will need to install Data.Time, which is quite a chore on Windows because of the dependencies. You will also need MSYS installed in order to be able to run autoreconf, so get that out of the way first. The summary for Windows is: darcs get --partial http://www-users.cs.york.ac.uk/~ndm/filepath/ (or download and unzip the prepared distribution) normal cabal configure, build, install darcs get --partial http://www.cse.unsw.edu.au/~dons/code/fps (or download and unzip the prepared distribution) normal cabal configure, build, install darcs get http://darcs.haskell.org/packages/Win32 Edit Win32.cabal, add fps to build-depends. normal cabal configure, build, install darcs get http://semantic.org/TimeLib/TimeLib There are two packages here: time and fixed. Ignore time. Go into fixed and do cabal configure, build, install darcs get http://darcs.haskell.org/packages/time Edit time.cabal, add fixed and Win32 to build-depends. create Setup.hs: import Distribution.Simple main = defaultMainWithHooks defaultUserHooks From MSYS shell, not Windows cmd.exe: $ autoreconf $ ghc --make Setup $ Setup configure $ Setup build $ Setup install And finally, build Takusen with the normal cabal configure, build, install.