From 0018bf63fcce35f6704cff5c34d398b32706fd75 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Valim?= Date: Mon, 19 Feb 2018 00:44:52 +0100 Subject: Compute MD5s in dialyzer using the .beam file The previous mechanism was based on the core file which meant that for every module in the PLT, we had to fetch its .beam file, retrieve Erlang AST, compile that down to core, serialize it into a binary and then get its MD5. In a project with stdlib, kernel, elixir and a small application in the PLT, relying only on the .beam sped up --check_plt from 10s to 0.8s. --- lib/dialyzer/src/dialyzer_plt.erl | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'lib/dialyzer/src') diff --git a/lib/dialyzer/src/dialyzer_plt.erl b/lib/dialyzer/src/dialyzer_plt.erl index 95c8b5ebce..2af4534396 100644 --- a/lib/dialyzer/src/dialyzer_plt.erl +++ b/lib/dialyzer/src/dialyzer_plt.erl @@ -531,17 +531,19 @@ compute_md5_from_files(Files) -> lists:keysort(1, [{F, compute_md5_from_file(F)} || F <- Files]). compute_md5_from_file(File) -> - case filelib:is_regular(File) of - false -> + case beam_lib:all_chunks(File) of + {ok, _, Chunks} -> + %% We cannot use beam_lib:md5 because it does not consider + %% the debug_info chunk, where typespecs are likely stored. + %% So we consider almost all chunks except the useless ones. + Filtered = [[ID, Chunk] || {ID, Chunk} <- Chunks, ID =/= "CInf", ID =/= "Docs"], + erlang:md5(lists:sort(Filtered)); + {error, beam_lib, {file_error, _, enoent}} -> Msg = io_lib:format("Not a regular file: ~ts\n", [File]), throw({dialyzer_error, Msg}); - true -> - case dialyzer_utils:get_core_from_beam(File) of - {error, Error} -> - throw({dialyzer_error, Error}); - {ok, Core} -> - erlang:md5(term_to_binary(Core)) - end + {error, beam_lib, _} -> + Msg = io_lib:format("Could not compute MD5 for .beam: ~ts\n", [File]), + throw({dialyzer_error, Msg}) end. init_diff_list(RemoveFiles, AddFiles) -> -- cgit v1.2.3