Permalink
Switch branches/tags
Nothing to show
Commits on Jan 17, 2017
  1. Revert to original Nuget package restore command

    jack-pappas committed Jan 17, 2017
  2. Migrate travis CI builds to container-based Linux

    jack-pappas committed Jan 17, 2017
Commits on Feb 6, 2016
  1. Implemented stepwise simplification/normalization for Regex.

    jack-pappas committed Feb 6, 2016
Commits on Feb 4, 2016
  1. Fix tool path for fsharplex tests; the tests were incorrectly trying …

    jack-pappas committed Feb 4, 2016
    …to invoke Reggie instead of fsharplex. Fixes #22.
  2. Upgrade to a newer FAKE build that understands how to find F# 4.0 (no…

    jack-pappas committed Feb 4, 2016
    …w required to build Reggie).
Commits on Feb 3, 2016
  1. Leave [<InternalsVisibleTo>] in place in non-debug builds too, to fix…

    jack-pappas committed Feb 3, 2016
    … CI builds. Partially addresses #22.
Commits on Feb 2, 2016
  1. AppVeyor builds need to use VS2015.

    jack-pappas committed Feb 2, 2016
  2. Remove pause command in build.bat, as it's holding up AppVeyor builds.

    jack-pappas committed Feb 2, 2016
  3. Fix FsCheck reference paths in test projects.

    jack-pappas committed Feb 2, 2016
    Fixed a bit of code in TestHelpers to work with the TestResult.True constructor, which had a breaking change in FsCheck 2.2.4.
    Fixes #21.
  4. Add AppVeyor badge to readme.

    jack-pappas committed Feb 2, 2016
  5. Remove possibly-interfering import in Facio.Utilities.Testing.

    jack-pappas committed Feb 2, 2016
  6. Fix AppVeyor build config to use the correct filename for the Windows…

    jack-pappas committed Feb 2, 2016
    … build script.
  7. Add AppVeyor build configuration.

    jack-pappas committed Feb 2, 2016
  8. Update travis-ci build to use latest Mac OS X build of Mono (currentl…

    jack-pappas committed Feb 2, 2016
    …y 4.2.2).
  9. Fixed Reggie.Tests to target FsCheck 2.2.4, which was overlooked in m…

    jack-pappas committed Feb 2, 2016
    …y previous commit.
  10. Uniformly target FsCheck 2.2.4 and FSharp.Core 4.3.1.0 to fix build w…

    jack-pappas committed Feb 2, 2016
    …arnings (and errors on some systems).
Commits on Dec 13, 2015
  1. Fixed F# code printing for Regex.CharacterSet case, so the intervals …

    jack-pappas committed Dec 13, 2015
    …in the CharSet are printed correctly.
  2. Fixed cases that caused Regex.Simplify not to be normalizing (i.e., w…

    jack-pappas committed Dec 13, 2015
    …hen it returns, it should always have reached a fixpoint).
    
    Fixing the issue with Regex.Simplify required changing the Regex.And and Or cases to be skewed right (instead of left) by simplifyRec; this seems to have greatly impacted performance for the worse, an issue which needs to be addressed next.
    Added some tracing to Regex.simplifyRec for use while debugging/testing.
Commits on Dec 6, 2015
  1. Clean up test cases for Regex.Simplify() in Reggie.Tests by combining…

    jack-pappas committed Dec 6, 2015
    … multiple tests into a single test with a TestCaseSource.
    
    Implemented a wrapper for the 'assertProp' helper function which records each test input to a file in case the property assertion crashes (with a StackOverflowException) for that input; this makes it easier to know which input caused the crash so it can be addressed.
  2. Implemented rewrite rules in Regex.Simplify to remove repetitive occu…

    jack-pappas committed Dec 6, 2015
    …rrences of Epsilon in a chain of Or cases.
  3. Upgrade the other testing projects using FsCheck to version 2.2.2 for…

    jack-pappas committed Dec 6, 2015
    … the performance improvements available in that version.
  4. Implemented additional rewrite rules for simplifying Regex.And patterns;

    jack-pappas committed Dec 6, 2015
    these rules fix some cases that caused Regex.Simplify not to be normalizing
    (and could even cause it to fail to terminate).
Commits on Nov 8, 2015
  1. Decrease timeout for fsharplex test cases.

    jack-pappas committed Nov 8, 2015
    The timeout was originally set to two minutes because certain input specifications (e.g., the fsharp lexer) hit a performance edge case that greatly slowed things down. That was fixed a while back; now, even for large input specifications fsharplex should finish running in a few seconds or less (on a fast machine).
  2. Fix one more error related to Reggie re-namespacing.

    jack-pappas committed Nov 8, 2015
  3. Fix the remaining error from the previous commit.

    jack-pappas committed Nov 8, 2015
  4. Fix namespacing in Reggie (it should use it's own namespace, not FSha…

    jack-pappas committed Nov 8, 2015
    …rpLex).
    
    Note, this commit breaks the build -- there is a compilation error in FSharpLex/Program.fs that's easy to fix but I'd like to investigate _why_ it's an error.
Commits on Nov 7, 2015
  1. Implemented rewrite rules to better (more efficiently) handle doubly-…

    jack-pappas committed Nov 7, 2015
    …nested And and Or patterns.
  2. Use FsCheck to generate a char[] instead of a seq<char>. Fixes #20.

    jack-pappas committed Nov 7, 2015
Commits on Nov 1, 2015
  1. Upgrade Reggie.Tests to use FsCheck 2.2.0.

    jack-pappas committed Nov 1, 2015
Commits on Aug 30, 2015
  1. Modified the Ast.RuleClause record to preserve location information f…

    jack-pappas committed Aug 30, 2015
    …or the clause pattern.
    
    Implemented most of the necessary functionality to emit a warning message with location info when the pattern in an Ast.RuleClause won't ever be matched.
    Small modification to the fsharplex parser corresponding to the Ast.RuleClause change; the parser doesn't currently have a way to get the position information for the pattern, it just sets the position to None.
    Simplifications in the Compiler.compileRule function related to detection of never-matched rule clauses.
Commits on Aug 16, 2015
  1. Fix #1. fsharplex now saves the position information for user-defined…

    jack-pappas committed Aug 16, 2015
    … code (for actions executed when rules are matched) and the fslex-compatible backend emits F# #line directives to map the generated code back to the lexer specification file.
  2. [skip ci] Committing a half-completed experiment to replace the tuple…

    jack-pappas committed Aug 16, 2015
    … type used as the element type of the AVL tree used to implement CharDiet with a struct to see if there's any noticeable performance improvement. I'll finish the experiment later, but I don't want to forget about it and the changes are both minimal and commented out so there's no difference in functionality.