[Python-checkins] r52366 - sandbox/trunk/import_in_py/importer.py

[brett.cannon]

Author: brett.cannon
Date: Tue Oct 17 21:38:39 2006
New Revision: 52366

Modified:
   sandbox/trunk/import_in_py/importer.py
Log:
Note how one could implement a sqlite3 back-end.


Modified: sandbox/trunk/import_in_py/importer.py
==============================================================================
--- sandbox/trunk/import_in_py/importer.py	(original)
+++ sandbox/trunk/import_in_py/importer.py	Tue Oct 17 21:38:39 2006
@@ -60,10 +60,54 @@
     + XXX
       
 sqlite3 importer use-case:
-* Simple way
-    + XXX
+* Simple way (only support just bytecode or source)
+    + DB
+        - For source just a column for module name and another containing the
+          source.
+        - For bytecode, a column for module name, another containing .pyc
+          output.
+    + Importer
+        - Query DB to see if module name is there.
+    + Loader
+        - Read entry in DB for module into a StringIO instance.
+        - Set 'name' attribute on StringIO instance to be DB path plus module name.
+    + Use the py/pyc file handler set for just source or bytecode.
 * Feature-rich way
-    + XXX
+    + DB
+        - Module name.
+        - Source code.
+        - Marshalled code of the source.
+            * Set to NULL when source code is modified, which removes need of
+              storing the bytecode's timestamp.
+        - Magic number for bytecode.
+    + Importer
+        - Query DB to see if module is listed.
+        - Whether it is in bytecode or source form is irrelevant.
+    + Loader
+        - See if bytecode is in DB; if so then have that used, otherwise source.
+        - Pass in tuple of DB instance and requested module.
+    + Subclass PyPycBaseHandler
+        - Have both source and bytecode opaque objects be a tuple of the DB
+          instance and the module name that is being handled.
+        - get_bytecode()
+            * Reads DB and returns magic number, 0 for timestamp, and bytecode.
+            * Timestamp is unneeded as existence of bytecode implicitly means
+              that it is not outdated.
+        - find_source()
+            * Can return same tuple as passed into handle_code().
+        - verify_timestamp()
+            * ``return True``.
+        - get_code_from_source()
+            * Get source from DB.
+            * compile.
+            * return code object and 0 for timestamp.
+        - write_bytecode()
+            * Get marshalled string from code object.
+            * Write marshalled string to DB.
+            * Write current magic number to DB.
+        - get_location()
+            * Return the path to the DB and the module name.
+            * Could even return SQL statement to get module if one cared.
 
 """
 from __future__ import with_statement