[Mono-bugs] [Bug 81663][Wis] Changed - [PATCH] Performance: Delegate optimization, DLR and IronPython

[bugzilla-daemon at bugzilla.ximian.com]

Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by lupus at ximian.com.

http://bugzilla.ximian.com/show_bug.cgi?id=81663

--- shadow/81663	2007-07-20 12:54:08.000000000 -0400
+++ shadow/81663.tmp.29065	2007-08-01 08:11:58.000000000 -0400
@@ -269,6 +269,16 @@
 
 ------- Additional Comments From vargaz at gmail.com  2007-07-20 12:54 -------
 This look ok. You can avoid guessing how much space to reserve by
 allocating code using g_malloc, emitting the code, then allocating from
 the code manager when the length of the emitted code is known.
 
+
+------- Additional Comments From lupus at ximian.com  2007-08-01 08:11 -------
+For this cases we should delegate caching to the arch-specific code.
+In most cases, for eligible signatures, the exact type doesn't matter,
+only the number of params does. With this in place we could allocate
+a lot less memory and we could do it in a not appdomain specific way.
+For x86 not taking into account struct returns we could have a simple
+array as the cache lookup, with the number of arguments as the index
+into the cache. The code memory would be allocated with
+mono_global_codeman_reserve().