diff options
author | Kartik K. Agaram <vc@akkartik.com> | 2017-10-30 13:45:17 -0700 |
---|---|---|
committer | Kartik K. Agaram <vc@akkartik.com> | 2017-10-30 13:45:17 -0700 |
commit | aa68aeb34e2476af8bfea8ba11e08e3d5f1d19f1 (patch) | |
tree | 6d2e7b113e9862033ccbd08f297d383f62afb382 /054static_dispatch.cc | |
parent | 7530c98b44c3d2228c9ec64b1839a1da6e1a9aaa (diff) | |
download | mu-aa68aeb34e2476af8bfea8ba11e08e3d5f1d19f1.tar.gz |
4098
Finally, make the seemingly-trivial change to buffer methods that I was envisioning 2 days ago. I still have zero confidence in our heuristic for picking the generic method to specialize for a call-site. Waiting for issues to reveal themselves.
Diffstat (limited to '054static_dispatch.cc')
-rw-r--r-- | 054static_dispatch.cc | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/054static_dispatch.cc b/054static_dispatch.cc index 702e5331..0af3ddeb 100644 --- a/054static_dispatch.cc +++ b/054static_dispatch.cc @@ -184,18 +184,22 @@ string best_variant(instruction& inst, const recipe& caller_recipe) { vector<recipe_ordinal> candidates; // Static Dispatch Phase 1 +//? cerr << inst.name << " phase 1\n"; candidates = strictly_matching_variants(inst, variants); if (!candidates.empty()) return best_variant(inst, candidates).name; // Static Dispatch Phase 2 +//? cerr << inst.name << " phase 2\n"; candidates = strictly_matching_variants_except_literal_against_address_or_boolean(inst, variants); if (!candidates.empty()) return best_variant(inst, candidates).name; +//? cerr << inst.name << " phase 3\n"; // Static Dispatch Phase 3 //: (shape-shifting recipes in a later layer) // End Static Dispatch Phase 3 // Static Dispatch Phase 4 +//? cerr << inst.name << " phase 4\n"; candidates = matching_variants(inst, variants); if (!candidates.empty()) return best_variant(inst, candidates).name; |