| Sun, 01 Nov 2015 16:33:14 -0500 |
Adam Chlipala |
After a tricky debugging session, limit visibility of type-class instances from anonymous modules |
| Sat, 24 Oct 2015 17:23:07 -0400 |
Adam Chlipala |
More Elaborate work to avoid instantiating unification variables with out-of-scope regular variables |
| Thu, 27 Aug 2015 16:28:45 -0400 |
Adam Chlipala |
Proper error message for excessively fancy FFI signatures |
| Wed, 28 Jan 2015 08:47:04 -0500 |
Adam Chlipala |
Improve wildification for records of type-class witnesses |
| Sat, 03 Jan 2015 10:21:26 -0500 |
Adam Chlipala |
Fixing one small issue with changes to signature matching (grandfathered into release) |
| Tue, 30 Dec 2014 17:02:31 -0500 |
Adam Chlipala |
Fix a bug in subsignature checking for submodules |
| Tue, 23 Dec 2014 13:42:20 -0500 |
Adam Chlipala |
Improve wildify heuristic for finding record type-class witnesses |
| Tue, 23 Dec 2014 11:23:27 -0500 |
Adam Chlipala |
Another try at a proper fix for constraint matching in subsignature checking |
| Fri, 05 Dec 2014 19:41:27 -0500 |
Adam Chlipala |
Move code from last changeset, to improve performance |
| Fri, 01 Aug 2014 16:11:36 -0400 |
Adam Chlipala |
When unifying constructor-level unification variables, also unify their kinds |
| Tue, 29 Jul 2014 14:46:06 -0400 |
Adam Chlipala |
Retweak the last tweak to allow type inference to succeed in a strict superset of the places where it used to succeed |
| Tue, 29 Jul 2014 14:38:50 -0400 |
Adam Chlipala |
Small tweak to ordering of unification rules, to support better record matching |
| Fri, 02 May 2014 19:19:09 -0400 |
Adam Chlipala |
New lessSafeFfi |
| Fri, 02 May 2014 17:16:02 -0400 |
Adam Chlipala |
Monadic bind supports patterns |
| Thu, 20 Feb 2014 15:50:33 -0500 |
Adam Chlipala |
Tweaked parameter renaming for functors, so now demos and the original bug-triggering application work |
| Thu, 20 Feb 2014 10:27:15 -0500 |
Adam Chlipala |
Some more nested functor bug-fixing, including generating fresh internal names at applications; still need to debug issues with datatype constructors |
| Fri, 14 Feb 2014 15:11:22 -0500 |
Adam Chlipala |
Fix a subtle renaming problem in elaborating functor applications |
| Wed, 29 Jan 2014 18:19:25 -0500 |
Adam Chlipala |
Add another simple case for record unification between alpha-equivalent constructors |
| Sun, 24 Nov 2013 09:56:50 -0500 |
Adam Chlipala |
Sneaky still-in-today's-release update of 'table' signature item handling |
| Mon, 04 Nov 2013 15:14:23 -0500 |
Adam Chlipala |
Interpret 'table' signature items more flexibly, automatically adding (Ur) constraints to support a kind of subtyping over (SQL) constraint sets |
| Thu, 17 Oct 2013 09:22:38 -0400 |
Adam Chlipala |
Save disjointness environments across top-level declarations |
| Thu, 26 Sep 2013 16:22:06 -0400 |
Adam Chlipala |
Get -root working properly again |
| Mon, 19 Aug 2013 12:38:43 -0400 |
Adam Chlipala |
Potentially exponential search through where to head-normalize in [decompileCon] |
| Mon, 19 Aug 2013 12:25:32 -0400 |
Adam Chlipala |
Allow [where con] to descend within submodule structure; open submodule constraints while checking later signature items |
| Sun, 29 Jul 2012 12:27:13 -0400 |
Adam Chlipala |
Remove 'class' declaration; now use 'con' instead |
| Thu, 17 May 2012 16:22:05 -0400 |
Adam Chlipala |
Implicit records of folders |
| Sat, 05 May 2012 12:58:13 -0400 |
Adam Chlipala |
-dumpTypesOnError |
| Sat, 05 May 2012 12:45:35 -0400 |
Adam Chlipala |
Send daemon output to calling process |
| Fri, 04 May 2012 10:33:04 -0400 |
Adam Chlipala |
More diagnostic information about some type class resolution failures |
| Thu, 03 May 2012 09:56:41 -0400 |
Adam Chlipala |
Catch duplicate top-level module names; fix dropping of constraints during incremental elaboration; document treatment of record types as type class instance types |
| Wed, 02 May 2012 08:56:19 -0400 |
Adam Chlipala |
Don't modify the module cache after elaboration failures |
| Wed, 02 May 2012 08:34:50 -0400 |
Adam Chlipala |
Better discovery of type class instances from recursive definitions |
| Sun, 29 Apr 2012 16:23:03 -0400 |
Adam Chlipala |
'urweb daemon start' and 'urweb daemon stop' |
| Sun, 29 Apr 2012 13:17:31 -0400 |
Adam Chlipala |
Initial support for reusing elaboration results |
| Sun, 22 Apr 2012 09:08:45 -0400 |
Adam Chlipala |
-unifyMore |
| Sat, 21 Apr 2012 14:57:00 -0400 |
Adam Chlipala |
Saving proper environments, to use in displaying nested error messages |
| Sat, 21 Apr 2012 13:57:10 -0400 |
Adam Chlipala |
Avoid kindof failures in elaboration |
| Mon, 16 Apr 2012 09:53:48 -0400 |
Adam Chlipala |
Simplify types for printing with '-dumpTypes' |
| Mon, 09 Jan 2012 16:20:10 -0500 |
Adam Chlipala |
Prevent horrifying loops in unification variable graph |
| Sat, 07 Jan 2012 11:01:21 -0500 |
Adam Chlipala |
Make wildification a bit smarter about ordering of new wildcard declarations |
| Fri, 06 Jan 2012 09:40:20 -0500 |
Adam Chlipala |
Revert last unification tweak |
| Thu, 05 Jan 2012 19:13:31 -0500 |
Adam Chlipala |
Tweak new unification heuristic |
| Thu, 05 Jan 2012 17:10:43 -0500 |
Adam Chlipala |
Prevent unifications of 'others' pieces in record summaries, when both pieces contain unification variables (to prevent undesired unifications) |
| Sun, 18 Dec 2011 12:00:36 -0500 |
Adam Chlipala |
Order constructors properly in wildification, to avoid spuriously displeasing the new scoping check |
| Sun, 18 Dec 2011 11:29:13 -0500 |
Adam Chlipala |
Add a new scoping check for unification variables, to fix a type inference bug |
| Sat, 03 Dec 2011 16:39:45 -0500 |
Adam Chlipala |
Nicer record summary error messages |
| Sat, 05 Nov 2011 13:12:07 -0400 |
Adam Chlipala |
Tweaks to choices of source positions to use in error messages, including for subSgn |
| Sat, 05 Nov 2011 11:33:59 -0400 |
Adam Chlipala |
Strengthen consNeq, to avoid unwise unifications |
| Sat, 05 Nov 2011 09:56:52 -0400 |
Adam Chlipala |
Fix confusing error message when record unification fails because enclosed types are incompatible |
| Wed, 02 Nov 2011 09:37:20 -0400 |
Adam Chlipala |
Remove debug print |
| Wed, 02 Nov 2011 08:39:01 -0400 |
Adam Chlipala |
Handle application in decompileCon |
| Sat, 29 Oct 2011 17:30:34 -0400 |
Adam Chlipala |
Shorter, more focused error messages about undetermined unification variables |
| Fri, 28 Oct 2011 08:54:44 -0400 |
Adam Chlipala |
Clear reducedSummaries |
| Thu, 27 Oct 2011 08:36:31 -0400 |
Adam Chlipala |
Harmonize have/need terminology in error messages; display canceled record summaries on errors |
| Fri, 21 Oct 2011 08:51:30 -0400 |
Adam Chlipala |
Catch UnboundNamed exceptions that arise from missing signature items |
| Sat, 15 Oct 2011 10:31:30 -0400 |
Adam Chlipala |
Change error message display order: only show disjointness/type class failures if all record unifications succeeded |
| Sat, 08 Oct 2011 14:16:13 -0400 |
Adam Chlipala |
Fix wildification for signatures with synonyms |
| Sat, 03 Sep 2011 08:57:13 -0400 |
Adam Chlipala |
Better detection of type class witnesses in local bindings |
| Sun, 07 Aug 2011 16:53:06 -0400 |
Adam Chlipala |
'-dumpTypes' |
| Sat, 06 Aug 2011 13:07:50 -0400 |
Adam Chlipala |
Better wildification, avoiding some unintentional variable capture |
