| commit | 120e028a84056279b9b6cdb9365e92dbb88bf924 | |
| author | Ruslan Sakevych <ruslans@fb.com> | |
| Mon, 12 Feb 2018 13:23:13 +0000 (12 05:23 -0800) | ||
| committer | Hhvm Bot <hhvm-bot@users.noreply.github.com> | |
| Mon, 12 Feb 2018 13:27:09 +0000 (12 05:27 -0800) | ||
| tree | f08f002b1e77c6da5ea0e26b49f0c9ddb1b79e59 | treesnapshot (tar.gz zip) |
| parent | 3a44e3dcab6a298b2d304bda242137bde1ae17d0 | commitdiff |
Remove syntax node introspection from the parser after it was parsed
Summary:
We are using "hacky" tricks to remove inspection of the node after it was created. Some of the cases were trivial like returning additional `Token.t` along with `Syntax.t` so that we can make some decision based on the former one.
Changing `PrefixUnaryExpression` was the hardest one. I had to store(cache) necessary information in the module state each time `make_prefix_unary_expression` was called, so that functions afterwards could inspect it.
Better datastructure for caching prefix unary expressions would be map/hashtable, but in our case is simple enough to just use list and iterate.
Differential Revision: D6890521
fbshipit-source-id: 4b6a130819a141d7699765ff2eed6f615834b7c2
Summary:
We are using "hacky" tricks to remove inspection of the node after it was created. Some of the cases were trivial like returning additional `Token.t` along with `Syntax.t` so that we can make some decision based on the former one.
Changing `PrefixUnaryExpression` was the hardest one. I had to store(cache) necessary information in the module state each time `make_prefix_unary_expression` was called, so that functions afterwards could inspect it.
Better datastructure for caching prefix unary expressions would be map/hashtable, but in our case is simple enough to just use list and iterate.
Differential Revision: D6890521
fbshipit-source-id: 4b6a130819a141d7699765ff2eed6f615834b7c2