Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* add age counter * working on it * it builds lol * trying out new terms proposal * not working due to cycles lol * fix up terms * implement function extraction * oops * significant proof checker progress * some progress * manual congruence * Revert "manual congruence" This reverts commit c578a77. * proof caching * fix global variable desugaring * fixed up primitive computation * small cleanup * working on resugar * resugar in files * resugar in main * weird perf behavior * add resugaring * fix perf problems with proofs * fix more prim stuff * fix proof bug, big refactor * fix another bug * magic iteration action * rebuilding fixes * simplify terms.rkt * oops * working on terms * builds * fix desugaring bug * terms example much simpler * saturate parent again * progress on terms desugaring * more progress * running into terrible assert failure * fix another desugar bug * better term desugaring * revert rebuilding * remove unecessary thing * Fix subtleties in rebuilding * fix another bug * bug in function extract * desugar simplify * remove old print * add compiler passes * a bunch of cleanup * we need to fix extraction * tweaks for tutorial example * oh no bad bug * fix up fibonacci-demand for terms * another bug * let bindings correctly * add optimization but without canonicalization" * actually i already did that * get rid of define, fix bug in terms * revert let binding encoding * Revert "revert let binding encoding" This reverts commit b56cc58. * filter equality with global variables * fix saturation bug * extract based on parent relationship * don't find value in debug * extract as an action or query * desugar set when it should be union * desugar set to union * this doesn't work anymore * some desugaring fixes * remove a print * use print-table * fix another benchmark * more bug fixes * clean up * fix up primitive queries * another normal form bug * small test * fix desugar global variables * fix bug in vec * vec tests * fix bug in filtering for primitives * rebuild non-eq tables too * add cache to fact desugaring * working on rebuild for custom containers" * fix vec rebuild! * rebuild for set * more cleanup * better error for unbound func * fix bug in query-extract * working on fixing extract variants * fix extract variants actually * fix bug in term encoding with initialization * remove some prints * fix defaults * better resugaring * remove prints again * horrible bug with fail * fix another global var bug ugh * another desugar bug * eq-able containers work better * working on vec again * fix find * unecessary dot * more desugar fixes for unbound vars * new desugar tests * proofs tests no longer relevant * no longer need presort table * canonical vars don't need to be updated * gj intersection sizes from Eli * resugar checks * trying to fix parent table * fix another parent table bug * working on computable functions in queries * fix bug in computed funcs * remove prints * some cleanup * revert all the computable stuff * delete slow test for now * all tests pass * remove rebuild and cleanup * delete random file * another random file * remove iteration * no longer using racket script * fix minimize script XD * add yihong's microbenchmark * Add some query compilation logging Co-authored-by: Oliver Flatt <oflatt@gmail.com> * Fix naive saturation Co-authored-by: Oliver Flatt <oflatt@gmail.com> * Convert parent tuples to filters Also push them up in queries. Co-authored-by: Oliver Flatt <oflatt@gmail.com> * Fix bug in check GJ instructions Co-authored-by: Oliver Flatt <oflatt@gmail.com> * Further refine the query compiler Co-authored-by: Oliver Flatt <oflatt@gmail.com> * disable query compiler changes * more cleanup * tell query compiler to do top-down * try more heuristics * less crazy heuristic * math micro change to not using match limit for consistency * simplify * generate queries as if parent table wasn't there * remove panic * print variable costs * clean up a bit * cleanup * re-add rebuilding * fix a bug with globals * globals broke seminaive * timestamps for globals * fix seminaive for globals * oops * some fixes from @ezrosent * backing out query compiler changes * more gj back out * refactor fact desugaring * remove unneeded arg to primitive apply --------- Co-authored-by: Eli Rosenthal <ezr@google.com> Co-authored-by: Max Willsey <me@mwillsey.com>
- Loading branch information