commit 713e9113770f7c35304e030c1111b93d3e4c8cb2
parent 3b9a79f8e0368d8bb9c2c43a8ac1cb0bf78c2918
Author: Tomas Hlavaty <tom@logand.com>
Date: Wed, 7 Sep 2011 00:56:35 +0200
README updated
Diffstat:
| M | README | | | 159 | +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---- |
1 file changed, 151 insertions(+), 8 deletions(-)
diff --git a/README b/README
@@ -1,3 +1,12 @@
+http://ondoc.logand.com/d/900/5/
+http://users.encs.concordia.ca/~haarslev/publications/jvlc92/node6.html
+http://www.cliki.net/MOP%20design%20patterns
+http://www.alu.org/mop/concepts.html
+http://www.alu.org/mop/dictionary.html
+http://www.lispworks.com/documentation/lw60/LW/html/lw-227.htm
+
+
+
- meta-syntax with keywords (small, fast, not extensible)
- syntax with macros (M-. etc)
@@ -41,11 +50,145 @@
http://www.bricklin.com/firstspreadsheetquestion.htm
- But VisiCalc was not an accounting program at all, it just made it
- possible for people to do accounting. Programs that were overly
- tuned for such function (Javelin, Lotus Improv, etc.) completely
- failed...What made VisiCalc novel was the ability to not only
- interact but have it learn by example. Again, VisiCalc doesn't
- summarize or do anything, it is just a tool to allow others to work
- out their ideas and reduce the tedium of repeating the same
- calculations.
+> But VisiCalc was not an accounting program at all, it just made it
+> possible for people to do accounting. Programs that were overly
+> tuned for such function (Javelin, Lotus Improv, etc.) completely
+> failed...What made VisiCalc novel was the ability to not only
+> interact but have it learn by example. Again, VisiCalc doesn't
+> summarize or do anything, it is just a tool to allow others to work
+> out their ideas and reduce the tedium of repeating the same
+> calculations.
+
+
+
+
+
+
+
+ 2sql
+
+ Tomas Hlavaty
+ <http://logand.com/contact.html>
+
+ 2011-09-06
+
+ Updated: 2011-09-07
+
+2sql is an extensible Lisp to SQL compiler and runtime with ORM layer
+built on top of it.
+
+ Table of Contents
+ =================
+
+ - Motivation and comparison with other projects
+ - Dependencies
+ - 2sql compiler
+ - Printer
+ - Compiler
+ - Macros
+ - Backends
+ - Portability
+ - Object Relational Mapping
+ - Custom persistent slot types
+ - Inheritance
+ - Schema evolution
+ - Portability
+
+ Motivation and comparison with other projects
+ =============================================
+
+ Dependencies
+ ============
+
+ - split-sequence
+ - ironclad
+ - babel
+ - cl-postgres
+ - sqlite
+ - closer-mop
+
+ 2sql compiler
+ =============
+
+ Printer
+ -------
+
+ Compiler
+ --------
+
+ Macros
+ ------
+
+ Backends
+ --------
+
+ Portability
+ -----------
+
+ Object Relational Mapping
+ =========================
+
+Use 2SQL-ORM:DEFPCLASS to define persistent classes or use DEFCLASS
+directly and inherit from 2SQL-ORM::PERSISTENT-OBJECT and specify
+:METACLASS 2SQL-ORM::PERSISTENT-CLASS. If the persistent class
+inherits from mixins, those must have the persistent class metaclass
+too.
+
+To create a persistent instance, use MAKE-PINSTANCE.
+
+Instance cache must be manually cleared as necessary when using UPDATE
+or DELETE-FROM SQL commands except when:
+
+ - SETF on a persistent slot is used;
+ - DELETE-PINSTANCE is called.
+
+ Custom persistent slot types
+ ----------------------------
+
+Types must be specified for all persistent slots because the SQL types
+are derived from the Lisp types. Use 2SQL-ORM:DEFPTYPE to define new
+Lisp type with SQL type mapping. For example,
+
+CODE
+ (2sql-orm:defptype username () '(2sql-orm:text 32 3))
+END
+
+defines a new type username where (< 3 (length a) 32).
+
+ Inheritance
+ -----------
+
+Persistent class inheritance is supported. The concept lives only on
+the lisp side and does not map to table inheritance even if supported
+by the database backend.
+
+ Schema evolution
+ ----------------
+
+TODO Schema evolution not implemented yet
+
+ Portability
+ -----------
+
+2sql-orm is less portable then plain 2sql because of feature mismatch
+between database backends and varied MOP compatibility between
+different Lisps.
+
+ | | sbcl | ccl | clisp | ecl |
+ |------------+------+-----+-------+-----|
+ | postgresql | Y | | | |
+ | sqlite | | | | |
+ | oracle | | | | |
+ | mysql | | | | |
+
+Currently works on sbcl only.
+
+ - TODO fix ccl issue with proxy: ccl is strict with types. Slot
+ type should probably be something like (or proxy user) or so.
+
+ - TODO clisp crashes
+
+ - TODO ecl has issues with deftype?
+
+ - TODO sqlite doesn't have sequences, needed for oid (mysql have
+ similar problem?)
