| commit | 7026f263174db7e16c981f6e2011add54a27e9a7 | |
| author | Nick Benton <pnb@fb.com> | |
| Thu, 25 May 2017 15:42:30 +0000 (25 08:42 -0700) | ||
| committer | Hhvm Bot <hhvm-bot@users.noreply.github.com> | |
| Thu, 25 May 2017 15:53:59 +0000 (25 08:53 -0700) | ||
| tree | 635d4bba3dc8ccfc0adbbc8712a444d645a0f54c | treesnapshot (tar.gz zip) |
| parent | db729e32129fb3d895d8b410ae7c80b9c98d8623 | commitdiff |
actually check adata in semdiff again
Summary:
We recently switched to having an explicit adata section in programs
and referring to entries therein by id in the instruction stream,
rather than having the literal data within the instruction
stream. This fixes comparison of ILiteral instructions to again
compare the actual referenced data values, rather than the ids.
Additionally, allowed DefFunc instructions to take an integer argument. That allows
recently added tests to go through without breaking the parser, but how this should
be dealt with may need revisiting.
Reviewed By: hubyrod
Differential Revision: D5129147
fbshipit-source-id: 1e863a5ad90cdcc1192239435ac2dfc4fbac1f8d
Summary:
We recently switched to having an explicit adata section in programs
and referring to entries therein by id in the instruction stream,
rather than having the literal data within the instruction
stream. This fixes comparison of ILiteral instructions to again
compare the actual referenced data values, rather than the ids.
Additionally, allowed DefFunc instructions to take an integer argument. That allows
recently added tests to go through without breaking the parser, but how this should
be dealt with may need revisiting.
Reviewed By: hubyrod
Differential Revision: D5129147
fbshipit-source-id: 1e863a5ad90cdcc1192239435ac2dfc4fbac1f8d
| hphp/hack/src/hhbc/Hhas_parser_actions.ml | diffblobblamehistory | |
| hphp/hack/src/hhbc/semdiff/rhl.ml | diffblobblamehistory | |
| hphp/hack/src/hhbc/semdiff/semdiff.ml | diffblobblamehistory |