Add SnoopPrecompile

Project.toml

@@ -1,22 +1,24 @@
 name = "Automa"
 uuid = "67c07d97-cdcb-5c2c-af73-a7f9c32a568b"
 authors = ["Kenta Sato <[email protected]>", "Jakob Nybo Nissen <[email protected]"]
-version = "0.8.2"
-
-[weakdeps]
-TranscodingStreams = "3bb67fe8-82b1-5028-8e26-92a6c54297fa"
+version = "1.0.0"
 
 [deps]
 ScanByte = "7b38b023-a4d7-4c5e-8d43-3f3097f304eb"
+SnoopPrecompile = "66db9d55-30c0-4569-8b51-7e840670fc0c"
+TranscodingStreams = "3bb67fe8-82b1-5028-8e26-92a6c54297fa"
+
+[weakdeps]
 TranscodingStreams = "3bb67fe8-82b1-5028-8e26-92a6c54297fa"
 
 [extensions]
 AutomaStream = "TranscodingStreams"
 
 [compat]
 ScanByte = "0.3.3"
+SnoopPrecompile = "1"
 TranscodingStreams = "0.9"
-julia = "1.5"
+julia = "1.6"
 
 [extras]
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"

src/Automa.jl

@@ -45,6 +45,8 @@ end
 
 using .RegExp: RE, @re_str, opt, rep, rep1, onenter!, onexit!, onall!, onfinal!, precond!
 
+include("workload.jl")
+
 # This list of exports lists the API
 export RE,
     @re_str,

src/tokenizer.jl

@@ -22,8 +22,8 @@ end
 
 # By default, the counter C is 1
 Tokenizer{E}(data) where E = Tokenizer{E, typeof(data), 1}(data)
-Base.IteratorSize(::Type{<:Tokenizer}) = Base.SizeUnknown()
-Base.eltype(::Type{<:Tokenizer{E}}) where E = Tuple{Int, Int32, E}
+Base.IteratorSize(::Type{Tokenizer{E, D, C}}) where {E, D, C} = Base.SizeUnknown()
+Base.eltype(::Type{Tokenizer{E, D, C}}) where {E, D, C} = Tuple{Int, Int32, E}
 
 """
     tokenize(::Type{E}, data, version=1)

src/workload.jl

@@ -0,0 +1,44 @@
+ using SnoopPrecompile: @precompile_all_calls
+
+@precompile_all_calls begin
+let
+    # Create a buffer validator
+    regex = let
+        name = onexit!(onenter!(re"[^\t\r\n]+", :mark), :name)
+        field = onexit!(onenter!(re"[^\t\r\n]+", :mark), :field)
+        nameline = name * rep('\t' * name)
+        record = onexit!(field * rep('\t' * field), :record)
+        nameline * re"\r?\n" * record * rep(re"\r?\n" * record) * rep(re"\r?\n")
+    end
+    generate_buffer_validator(:foo, regex)
+
+    # Create an ordinary parser
+    machine = compile(regex)
+    actions = Dict(
+        :mark => :(pos = p),
+        :name => :(push!(headers, String(data[pos:p-1]))),
+        :field => quote
+            n_fields += 1
+            push!(fields, String(data[pos:p-1]))
+        end,
+        :record => quote
+            n_fields == length(headers) || error("Malformed TSV")
+            n_fields = 0
+        end
+    )
+
+    generate_code(machine, actions)
+
+    # Create a tokenizer
+    tokens = [
+        re"[A-Za-z_][0-9A-Za-z_!]*!",
+        re"\(",
+        re"\)",
+        re",",
+        re"abc",
+        re"\"",
+        re"[\t\f ]+",
+    ];
+    make_tokenizer(tokens)
+end
+end