X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=TODO;h=7bc7cb1aa33523785d011d714e16bcb1438dbc7e;hb=88c736485b8d1e4416c0401dc635c3300b54ee2a;hp=673fa67c571bd58d4bc846831c7938fcded3a97e;hpb=c31622d78fe27e0c27a556df766c5a6b57612caa;p=dbsrgits%2FSQL-Translator.git

diff --git a/TODO b/TODO
index 673fa67..7bc7cb1 100644
--- a/TODO
+++ b/TODO
@@ -1,16 +1,39 @@
-Define structure of the data returned by parsers:
-  o Producers need to know what to expect
-  o For testing purposes, we need to know if parsers are returning
-    something reasonable
+*   The regular Sybase parser is only just functional.  If you are
+    interested in using Sybase, I would suggest serializing the schema
+    (via YAML or Storable) using the DBI-Sybase parser and then
+    manipulating that as you see fit. 
 
-Modules to be written/finished
-    SQL::Translator::Validator
-    SQL::Translator::Parser::xSV
-    SQL::Translator::Producer::MySQL
-    SQL::Translator::Parser::Excel
+*   Add more DBI parsers!  These have the potential to be very
+    thorough and far faster than parsing text files with
+    Parse::RecDescent.
 
-Should the parsers return an instance instead of a data structure?  It
-would make traversing the data structure easier.
+*   At least allow more pass-through of INSERT, DELETE, and UPDATE
+    statements
 
+*   Add INSERT statements for xSV, Excel parsers to automatically
+    create INSERTs for each row of data in the source file
 
-# vim: set sw=2 ts=2 tw=70 fo=trcqo:
+*   Somehow merge ClassDBI producer with CGI::FormBuilder or Template
+    Toolkit and some sort of automated CGI builder to create
+    view/create/edit/delete forms for objects based on schema defs
+
+*   Embetter the Diagram producer to use some real graphing algorithms
+    to distribute the tables so that the lines don't overlap so badly
+
+*   Integrate more with some standard XML schema representations,
+    maybe like Torque DB (http://db.apache.org/torque/).  We've
+    started messing around with XMI, too, but it isn't quite usable.
+
+*   Possibly write a basic ANSI-92 SQL parser which could be extended
+    when writing other new parsers.
+
+*   Make as many "required" modules as possible optional.  This will
+    require support in the Makefile, the tests, and the modules
+    themselves (they'll need to die gracefully if prerequisites are
+    not installed).
+
+*   Support for precompiled Parse::RecDescent grammars.
+
+*   More code generation producers, such as Java, PHP, and Python.
+
+*   Integrate Module::Pluggable as a replacement for the _list method.