Skip to content
Commits on May 24, 2016
  1. @jhellerstein

    Update operational.md

    Put the costs back in to motivate the stratification discussion.
    jhellerstein committed May 24, 2016
  2. @jhellerstein

    Update operational.md

    Remove costs from transitive closure on paths, to avoid concerns about infinite loops on cyclic data.
    jhellerstein committed May 24, 2016
Commits on May 19, 2016
  1. @jhellerstein

    Merge pull request #326 from sciolizer/master

    Syntax highlighting in getstarted.md
    jhellerstein committed May 18, 2016
  2. @sciolizer
Commits on Feb 28, 2014
  1. @neilconway

    Code cleanup for metrics.

    Note that this code is actually broken :-\ I didn't take the time to either fix
    it or rip it out, but this at least improves matters slightly.
    neilconway committed Feb 27, 2014
Commits on Feb 19, 2014
  1. @neilconway
Commits on Feb 16, 2014
  1. @neilconway

    Update changelog.

    neilconway committed Feb 15, 2014
  2. @neilconway

    Merge pull request #324 from vjoel/rakefile

    add rakefile with test and release tasks; add version file
    neilconway committed Feb 15, 2014
Commits on Feb 15, 2014
  1. @vjoel
Commits on Feb 14, 2014
  1. @neilconway

    Minor refactor.

    neilconway committed Feb 13, 2014
Commits on Jan 31, 2014
  1. @neilconway

    Minor code cleanup.

    neilconway committed Jan 31, 2014
Commits on Jan 16, 2014
  1. @neilconway
  2. @neilconway
  3. @neilconway

    Tweak stratification algorithm.

    Previously, we computed the minimum legal strata for each relation, and then
    placed every rule with that relation in its LHS in that strata. That is not
    incorrect, but it is unnecessarily conservative. Consider a program like:
    
    (r1) a <= b
    (r2) a <= c.group(...)
    (r3) c <= d
    
    Clearly, r2 can only be evaluated after c has been completely determined, but we
    needn't wait for c to be determined before evaluating r1. Hence, we can place r1
    and r3 in stratum 0 and r2 in stratum 1.
    
    Along the way, add a new builtin relation, t_rule_stratum, that contains the
    stratum associated with each rule.
    neilconway committed Jan 15, 2014
Commits on Jan 15, 2014
  1. @neilconway

    Minor code cleanup for strat.

    neilconway committed Jan 15, 2014
  2. @neilconway
Commits on Jan 4, 2014
  1. @neilconway

    Merge pull request #323 from freenerd/master

    Fixing docs/getstarted.md example code
    neilconway committed Jan 4, 2014
Commits on Jan 3, 2014
  1. @freenerd

    Fixing docs/getstarted.md example code

    Code snippets in the guide are fixed to be same as in /examples/chat
    freenerd committed Jan 3, 2014
Commits on Dec 11, 2013
  1. @neilconway

    Tweak new notin test case.

    neilconway committed Dec 10, 2013
  2. @neilconway

    Fix a bug in chaining multiple notins after a join.

    Somewhat obscure, but arises in RSE/GC-generated code. Basic problem is that we
    only want to call setup_accessors once; for lefts/rights, we were calling it
    multiple times. Unclear why that mucked things up, but it was certainly wrong
    anyway.
    
    On examination, the code for assigning a schema to a pairs join is completely
    bogus anyway (see comment) -- but I won't fix that for now.
    neilconway committed Dec 10, 2013
Commits on Dec 4, 2013
  1. @neilconway

    Tiny perf optimization.

    neilconway committed Dec 4, 2013
Commits on Nov 24, 2013
  1. @neilconway

    Minor code cleanup.

    Don't invoke #to_s on something that is always a String.
    neilconway committed Nov 23, 2013
Commits on Nov 12, 2013
  1. @neilconway
  2. @neilconway

    Tweak comment.

    neilconway committed Nov 12, 2013
Commits on Nov 11, 2013
  1. @neilconway

    Fix typo.

    neilconway committed Nov 11, 2013
Commits on Oct 15, 2013
  1. @neilconway

    Minor refactor for BudCollection.

    Try to avoid directly accessing BudCollection fields like @delta, @tick_delta
    from outside BudCollection.
    neilconway committed Oct 14, 2013
  2. @neilconway

    Improve tc_metrics test.

    neilconway committed Oct 14, 2013
  3. @neilconway

    Minor perf tweak.

    Avoid a needless conjunction in a conditional.
    neilconway committed Oct 14, 2013
  4. @neilconway
Commits on Oct 14, 2013
  1. @neilconway

    Fix typo in comment.

    neilconway committed Oct 14, 2013
  2. @neilconway

    Fix corner-case bug in BudCollection#include?

    If the collection has an empty set of key columns and a single tuple, we
    *cannot* assume that include?(t) is true for any t.
    neilconway committed Oct 13, 2013
Commits on Oct 12, 2013
  1. @palvaro

    Ensure that programs are only rejected as unstratifiable when negatio…

    …n occurs within a dependency cycle. Fixes #321.
    palvaro committed Oct 11, 2013
Commits on Oct 3, 2013
  1. @neilconway
  2. @neilconway

    Fix a bug in the stratification algorithm.

    The previous coding was wrong -- in some cases, a cycle through negation was not
    detected correctly. This meant that some unstratifiable programs were not
    rejected (and instead we might produce incorrect results for them).
    
    Along the way, do some minor code cleanup.
    neilconway committed Oct 3, 2013
  3. @neilconway

    Update changelog.

    neilconway committed Oct 2, 2013
Something went wrong with that request. Please try again.