| Wed, 11 Jan 2012 13:53:35 -0500 |
Adam Chlipala |
New, more principled heuristic for Especialize: only specialize uniform function arguments; that is, arguments that don't change across recursive calls |
| Wed, 11 Jan 2012 11:08:48 -0500 |
Adam Chlipala |
Tweak Especialize heuristic |
| Mon, 09 Jan 2012 09:51:39 -0500 |
Adam Chlipala |
Tweak Especialize heuristic to prevent non-termination |
| Sat, 07 Jan 2012 15:56:22 -0500 |
Adam Chlipala |
First part of changes to avoid depending on C function call argument order of evaluation (omitting normal Ur function calls, so far) |
| Thu, 06 Jan 2011 12:31:13 -0500 |
Adam Chlipala |
Expand valueish again |
| Thu, 06 Jan 2011 09:25:15 -0500 |
Adam Chlipala |
Broaden definition of valueish |
| Fri, 24 Dec 2010 12:51:46 -0500 |
Adam Chlipala |
Add an extra Especialize pass before Rpcify |
| Tue, 21 Dec 2010 13:57:12 -0500 |
Adam Chlipala |
Hopeful fix to stop Especialize infinite looping |
| Tue, 19 Oct 2010 17:54:49 -0400 |
Adam Chlipala |
Fixes for nasty bugs in Reduce and Especialize |
| Sun, 22 Aug 2010 13:43:46 -0400 |
Adam Chlipala |
Reduce: Inline let-bound variables whose types involve functions |
| Thu, 03 Jun 2010 13:04:37 -0400 |
Adam Chlipala |
Some serious bug-fix work to get HTML example to compile; this includes fixing a bug with 'val' patterns in Unnest and the need for more local reduction in Especialize |
| Tue, 16 Mar 2010 10:09:01 -0400 |
Adam Chlipala |
Undo an Especialize change that turned out to be unecessary |
| Sat, 06 Mar 2010 19:14:48 -0500 |
Adam Chlipala |
Got split1 working, but noticed a nasty type inference bug with transplanted unification variables |
| Sat, 06 Mar 2010 16:15:26 -0500 |
Adam Chlipala |
Standard library moduls Incl and Mem; tweaks to Especialize and Unpoly |
| Thu, 07 Jan 2010 14:02:58 -0500 |
Adam Chlipala |
Basis.debug; more restrictive type for Basis.form; weaken definition of polymorphic-ness for especialization |
| Wed, 23 Dec 2009 12:25:34 -0500 |
Adam Chlipala |
Stop skipping Especialization of generated functions; fix Compiler.parseUrp; expose uw_really_write(); allow more NULL arguments to uw_register_transactional() |
| Sun, 20 Dec 2009 15:17:43 -0500 |
Adam Chlipala |
Another try at reasonable Especialize, this time with a custom traversal |
| Tue, 15 Dec 2009 19:26:52 -0500 |
Adam Chlipala |
Especialize uses a termination measure based on number of arguments introduced |
| Tue, 15 Dec 2009 13:20:13 -0500 |
Adam Chlipala |
Fix new Especialize security bug: do not duplicate free variables as specialized arguments |
| Tue, 15 Dec 2009 12:26:00 -0500 |
Adam Chlipala |
Initial generalization of Especialize, with security bug known |
| Thu, 21 May 2009 11:45:04 -0400 |
Adam Chlipala |
Change List.mapM' to avoid leaving functions around |
| Sat, 16 May 2009 12:41:33 -0400 |
Adam Chlipala |
Fix a Core_untangle bug that missed closure variable references; XHTMLize |
| Thu, 14 May 2009 13:18:31 -0400 |
Adam Chlipala |
Proper lifting of MonoEnv stored expressions; avoidance of onchange clobbering |
| Thu, 14 May 2009 08:13:54 -0400 |
Adam Chlipala |
Improving/reordering Unpoly and Especialize; pathmaps |
| Sun, 12 Apr 2009 14:19:15 -0400 |
Adam Chlipala |
Classes as optional arguments to Basis.tag |
| Sun, 22 Feb 2009 17:17:01 -0500 |
Adam Chlipala |
demo/hello compiles with kind polymorphism |
| Thu, 27 Nov 2008 12:43:28 -0500 |
Adam Chlipala |
Avoid Especializing polymorphic code |
| Thu, 27 Nov 2008 12:34:44 -0500 |
Adam Chlipala |
Ditch use of ElabEnv.env in Especialize, to realize big speed-up |
