From 7c1f2dfd69afe119322e19b3a4061876b3dd1033 Mon Sep 17 00:00:00 2001 From: Bruce Hill Date: Fri, 14 Sep 2018 14:00:48 -0700 Subject: [PATCH] Fixes and improvements, and ported nomsu.3.peg to new tidier syntax. --- nomsu.3.peg | 166 ++++++++++++++++++++++++++------------------ nomsu.4.peg | 25 +++---- nomsu_compiler.lua | 71 +++++++++---------- nomsu_compiler.moon | 34 ++++----- parser.lua | 119 +------------------------------ parser.moon | 68 +----------------- pretty_errors.lua | 1 + pretty_errors.moon | 1 + 8 files changed, 167 insertions(+), 318 deletions(-) diff --git a/nomsu.3.peg b/nomsu.3.peg index 244a621..21fa86c 100644 --- a/nomsu.3.peg +++ b/nomsu.3.peg @@ -3,69 +3,94 @@ file: {:curr_indent: ' '* :} (((action / expression / inline_block / indented_block) eol !.) / file_chunks / empty_block) - %ws* (!! .+ -> "Parse error" !!)? + {:curr_indent: %nil :} + !. -shebang: "#!" (!"nomsu" [^%nl])* "nomsu" %ws+ "-V" %ws* {:version: [0-9.]+ :} [^%nl]* +shebang: "#!" (!"nomsu" [^%nl])* "nomsu" ws+ "-V" ws* {:version: [0-9.]+ :} [^%nl]* file_chunks (FileChunks): {:curr_indent: ' '* :} shebang? comment? blank_lines? (top_block (nl_nodent section_division top_block)*) blank_lines? + ws* unexpected_chunk? + {:curr_indent: %nil :} top_block (Block): {:curr_indent: ' '* :} comment? blank_lines? statement (nl_nodent statement)* + {:curr_indent: %nil :} empty_block (Block): {:curr_indent: ' '* :} comment? blank_lines? + {:curr_indent: %nil :} -nodent: =curr_indent !(" ") +nodent: (unexpected_indent [^%nl]* / =curr_indent) indent: =curr_indent " " -blank_lines: %nl ((nodent comment / %ws*) %nl)* -eol: %ws* eol_comment? (!. / &%nl) +blank_lines: %nl ((nodent comment / ws*) %nl)* +eol: ws* eol_comment? (!. / &%nl) nl_nodent: blank_lines nodent -nl_indent: blank_lines {:curr_indent: indent :} (comment nl_nodent)? +nl_indent: blank_lines {:curr_indent: indent :} (comment nl_nodent)* -comment: - "#" (({} {~ [^%nl]* (%nl+ (indent -> '') [^%nl]*)* ~} %userdata) => add_comment) -eol_comment: - "#" (({} {[^%nl]*} %userdata) => add_comment) +comment (Comment): + "#" {~ [^%nl]* (%nl+ (indent -> '') [^%nl]*)* ~} +eol_comment (Comment): + "#" {[^%nl]*} + +unexpected_code: ws* _unexpected_code +_unexpected_code (Error): + {:error: {~ [^%nl]+ -> "Couldn't parse this code" ~} :} +unexpected_chunk (Error): + {:error: {~ .+ -> "Couldn't parse this code" ~} :} +unexpected_indent (Error): + {:error: {~ (=curr_indent ws+) -> "Messed up indentation" ~} :} + {:hint: {~ '' -> 'Either make sure this line is aligned with the one above it, or make sure the previous line ends with something that uses indentation, like ":" or "(..)"' ~} :} +missing_paren_err (Error): + {:error: {~ eol -> 'Line ended without finding a closing )-parenthesis' ~} :} + {:hint: {~ '' -> 'Put a ")" here' ~} :} +missing_quote_err (Error): + {:error: {~ eol -> 'Line ended before finding a closing double quotation mark' ~} :} + {:hint: {~ "" -> "Put a quotation mark here" ~} :} +missing_bracket_error (Error): + {:error: {~ eol -> "Line ended before finding a closing ]-bracket" ~} :} + {:hint: {~ '' -> 'Put a "]" here' ~} :} +missing_brace_error (Error): + {:error: {~ eol -> "Line ended before finding a closing }-brace" ~} :} + {:hint: {~ '' -> 'Put a "}" here' ~} :} section_division: ("~")^+3 eol inline_block: - "(" %ws* inline_block %ws* ")" / raw_inline_block + "(" ws* inline_block ws* ")" / raw_inline_block raw_inline_block (Block): - (!"::") ":" %ws* ((inline_statement (%ws* ";" %ws* inline_statement)*) / !(eol nl_indent)) + (!"::") ":" ws* ((inline_statement (ws* ";" ws* inline_statement)*) / !(eol nl_indent)) indented_block (Block): - ":" eol nl_indent statement (nl_nodent statement)* (%nl (%ws* %nl)* nodent comment)* + ":" eol nl_indent statement (nl_nodent statement)* + (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* + {:curr_indent: %nil :} statement: - (action / expression) (eol / (!! [^%nl]+ -> "Unexpected code while parsing line" !!)) + (action / expression) (eol / unexpected_code) inline_statement: (inline_action / inline_expression) noindex_inline_expression: number / variable / inline_text / inline_list / inline_dict / inline_nomsu / ( "(" - %ws* (inline_action / inline_expression) %ws* - (%ws* ',' %ws* (inline_action / inline_expression) %ws*)* - (")" - / (!! eol -> 'Line ended without finding a closing )-parenthesis' !!) - / (!! [^%nl]+ -> 'Unexpected code while parsing subexpression' !!) - ) + ws* (inline_action / inline_expression) ws* + (ws* ',' ws* (inline_action / inline_expression) ws*)* + (")" / missing_paren_err / unexpected_code) ) inline_expression: index_chain / noindex_inline_expression indented_expression: indented_text / indented_nomsu / indented_list / indented_dict / ({| "(..)" nl_indent - (action / expression) (nl_nodent comment)* - (eol / (!! [^%nl]+ -> "Unexpected code while parsing indented expression" !!)) + (action / expression) (eol / unexpected_code) + (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* + {:curr_indent: %nil :} |} -> unpack) - / (nl_indent (!! [^%nl]* -> "Unexpected indentation. Perhaps you meant to put a ':' or '(..)' on the previous line?" !!) (nl_nodent [^%nl]*)*) expression: inline_expression / indented_expression @@ -79,45 +104,42 @@ index_chain (IndexChain): -- Actions need either at least 1 word, or at least 2 tokens inline_action (Action): !section_division - ({:target: inline_arg :} %ws* "::" %ws*)? - ( (inline_arg (%ws* (inline_arg / word))+) - / (word (%ws* (inline_arg / word))*)) - (%ws* inline_block)? + ({:target: inline_arg :} ws* "::" ws*)? + ( (inline_arg (ws* (inline_arg / word))+) + / (word (ws* (inline_arg / word))*)) + (ws* inline_block)? inline_arg: inline_expression / inline_block action (Action): !section_division - ({:target: arg :} (nl_nodent "..")? %ws* "::" (nl_nodent "..")? %ws*)? - ( (arg ((nl_nodent "..")? %ws* (arg / word))+) - / (word ((nl_nodent "..")? %ws* (arg / word))*)) + ({:target: arg :} (nl_nodent "..")? ws* "::" (nl_nodent "..")? ws*)? + ( (arg ((nl_nodent "..")? ws* (arg / word))+) + / (word ((nl_nodent "..")? ws* (arg / word))*)) arg: expression / inline_block / indented_block -word: !number { %operator_char+ / %ident_char+ } +word: !number { operator_char+ / ident_char+ } text_word (Text): word inline_text (Text): - !('".."' eol) - '"' - ({~ (('\"' -> '"') / ('\\' -> '\') / %escaped_char / [^%nl\"])+ ~} - / inline_text_interpolation)* - ('"' - / (!! eol -> 'Line ended before finding a closing double quotation mark' !!) - / (!! [^%nl]+ -> 'Unexpected code while parsing Text' !!)) + !(indented_text) + ('"' _inline_text* ('"' / missing_quote_err / unexpected_code)) +_inline_text: + {~ (('\"' -> '"') / ('\\' -> '\') / escaped_char / [^%nl\"]+)+ ~} + / inline_text_interpolation inline_text_interpolation: "\" ( variable / inline_list / inline_dict / inline_text / ("(" - %ws* (inline_action / inline_expression) %ws* - (%ws* ',' %ws* (inline_action / inline_expression) %ws*)* - (")" - / (!! eol -> 'Line ended without finding a closing )-parenthesis' !!) - / (!! [^%nl]+ -> 'Unexpected code while parsing Text interpolation' !!))) + ws* (inline_action / inline_expression) ws* + (ws* ',' ws* (inline_action / inline_expression) ws*)* + (")" / missing_paren_err / unexpected_code)) ) indented_text (Text): - '".."' eol %nl {%nl*} {:curr_indent: indent :} + '".."' eol %nl {%nl+}? {:curr_indent: indent :} (indented_plain_text / text_interpolation / {~ %nl+ (=curr_indent -> "") ~})* - (!! [^%nl]+ -> "Unexpected code while parsing Text" !!)? + unexpected_code? + {:curr_indent: %nil :} -- Tracking text-lines-within-indented-text as separate objects allows for better debugging line info indented_plain_text (Text): {~ (("\\" -> "\") / (("\" blank_lines =curr_indent "..") -> "") / (!text_interpolation "\") / [^%nl\]+)+ @@ -129,43 +151,51 @@ number (Number): (("-"? (([0-9]+ "." [0-9]+) / ("." [0-9]+) / "0x" [0-9a-fA-F]+ -- Variables can be nameless (i.e. just %) and can only contain identifier chars. -- This ensures you don't get weird parsings of `%x+%y` or `%'s thing`. -variable (Var): "%" {%ident_char*} +variable (Var): "%" {ident_char*} inline_list (List): !('[..]') - "[" %ws* - (inline_list_item (%ws* ',' %ws* inline_list_item)* (%ws* ',')?)? %ws* - ("]" / (","? ( - (!! eol -> "Line ended before finding a closing ]-bracket" !!) - /(!! [^%nl]+ -> "Unexpected code while parsing List" !!) - ))) + "[" ws* + (inline_list_item (ws* ',' ws* inline_list_item)* (ws* ',')?)? ws* + ("]" / (","? (missing_bracket_error / unexpected_code))) indented_list (List): "[..]" eol nl_indent - list_line (nl_nodent list_line)* (nl_nodent comment)* - (","? (!! [^%nl]+ -> "Unexpected code while parsing List" !!))? + list_line (nl_nodent list_line)* + (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* + (","? unexpected_code)? list_line: - (inline_list_item %ws* "," %ws*)+ eol - / (inline_list_item %ws* "," %ws*)* (action / expression) eol + (inline_list_item ws* "," ws*)+ eol + / (inline_list_item ws* "," ws*)* (action / expression) eol inline_list_item: inline_action / inline_expression inline_dict (Dict): !('{..}') - "{" %ws* - (inline_dict_entry (%ws* ',' %ws* inline_dict_entry)*)? %ws* - ("}" / (","? ( - (!! eol -> "Line ended before finding a closing }-brace" !!) - / (!! [^%nl]* -> "Unexpected code while parsing Dictionary" !!) - ))) + "{" ws* + (inline_dict_entry (ws* ',' ws* inline_dict_entry)*)? ws* + ("}" / (","? (missing_brace_error / unexpected_code))) indented_dict (Dict): "{..}" eol nl_indent - dict_line (nl_nodent dict_line)* (nl_nodent comment)* - (","? (!! [^%nl]+ -> "Unexpected code while parsing Dictionary" !!))? + dict_line (nl_nodent dict_line)* + (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* + (","? unexpected_code)? dict_line: - (inline_dict_entry %ws* "," %ws*)+ eol - / (inline_dict_entry %ws* "," %ws*)* dict_entry eol + (inline_dict_entry ws* "," ws*)+ eol + / (inline_dict_entry ws* "," ws*)* dict_entry eol dict_entry(DictEntry): - dict_key (%ws* ":" %ws* (action / expression))? + dict_key (ws* ":" ws* (action / expression))? inline_dict_entry(DictEntry): - dict_key (%ws* ":" %ws* (inline_action / inline_expression)?)? + dict_key (ws* ":" ws* (inline_action / inline_expression)?)? dict_key: text_word / inline_expression + +operator_char: ['`~!@$^&*+=|<>?/-] +ident_char: [a-zA-Z0-9_] / %utf8_char +ws: [ %tab] + +escaped_char: + ("\"->'') ( + (([xX]->'') ((({[0-9a-fA-F]^2} %number_16) -> tonumber) -> tochar)) + / ((([0-9] [0-9]^-2) -> tonumber) -> tochar) + / ("a"->ascii_7) / ("b"->ascii_8) / ("t"->ascii_9) / ("n"->ascii_10) + / ("v"->ascii_11) / ("f"->ascii_12) / ("r"->ascii_13) + ) diff --git a/nomsu.4.peg b/nomsu.4.peg index eb0c878..8d8296e 100644 --- a/nomsu.4.peg +++ b/nomsu.4.peg @@ -11,14 +11,14 @@ shebang: "#!" (!"nomsu" [^%nl])* "nomsu" ws+ "-V" ws* {:version: [0-9.]+ :} [^%n file_chunks (FileChunks): {:curr_indent: ' '* :} shebang? comment? blank_lines? - (top_block (section_division nl_nodent top_block)*) + (top_block (nl_nodent section_division top_block)*) blank_lines? ws* unexpected_chunk? {:curr_indent: %nil :} top_block (Block): {:curr_indent: ' '* :} - comment? blank_lines? statement (nl_nodent statement?)* + comment? blank_lines? statement (nl_nodent statement)* {:curr_indent: %nil :} empty_block (Block): @@ -71,7 +71,7 @@ inline_block: raw_inline_block (Block): (!"::") ":" ws* ((inline_statement (ws* ";" ws* inline_statement)*) / !(eol nl_indent)) indented_block (Block): - ":" eol nl_indent statement (nl_nodent statement?)* + ":" eol nl_indent statement (nl_nodent statement)* (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* {:curr_indent: %nil :} @@ -149,19 +149,16 @@ indented_text (Text): cool_quote: '"' !(%nl+ !(=curr_indent)) cool_indented_text (Text): - ({| - '"' - _inline_text* - (('\' %nl+ {:curr_indent: indent :} '..') - / disallowed_interpolation? {%nl+} {:curr_indent: indent :}) - (indented_cool_plain_text / text_interpolation / {~ %nl+ (=curr_indent -> "") ~})* - ('"' eol / missing_quote_err) - |} -> unpack) + '"' + _inline_text* + (('\' %nl+ {:curr_indent: indent :} '..') + / disallowed_interpolation? {%nl+} {:curr_indent: indent :}) + (indented_cool_plain_text / text_interpolation / {~ %nl+ (=curr_indent -> "") ~})* + ('"' eol / missing_quote_err) indented_cool_plain_text (Text): {~ ((("\" blank_lines =curr_indent "..") -> "") / (!text_interpolation ((!("\n") escaped_char) / ('\\' -> '\') / '\')) / (cool_quote / [^%nl"\])+)+ (%nl+ (=curr_indent -> ""))* ~} - -- Tracking text-lines-within-indented-text as separate objects allows for better debugging line info indented_plain_text (Text): {~ (("\\" -> "\") / (("\" blank_lines =curr_indent "..") -> "") / (!text_interpolation "\") / [^%nl\]+)+ @@ -182,7 +179,7 @@ inline_list (List): ("]" / (","? (missing_bracket_error / unexpected_code))) indented_list (List): "[..]" eol nl_indent - list_line (nl_nodent list_line?)* + list_line (nl_nodent list_line)* (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* (","? unexpected_code)? list_line: @@ -197,7 +194,7 @@ inline_dict (Dict): ("}" / (","? (missing_brace_error / unexpected_code))) indented_dict (Dict): "{..}" eol nl_indent - dict_line (nl_nodent dict_line?)* + dict_line (nl_nodent dict_line)* (%nl (ws* %nl)* nodent (comment / eol / unexpected_code))* (","? unexpected_code)? dict_line: diff --git a/nomsu_compiler.lua b/nomsu_compiler.lua index 510de16..5578e0e 100644 --- a/nomsu_compiler.lua +++ b/nomsu_compiler.lua @@ -62,12 +62,45 @@ table.copy = function(t) return _tbl_0 end)(), getmetatable(t)) end +local make_tree +make_tree = function(tree, userdata) + local cls = AST[tree.type] + tree.source = Source(userdata.filename, tree.start, tree.stop) + tree.start, tree.stop = nil, nil + tree.type = nil + do + local _accum_0 = { } + local _len_0 = 1 + for _index_0 = 1, #tree do + local t = tree[_index_0] + if AST.is_syntax_tree(t, "Comment") then + _accum_0[_len_0] = t + _len_0 = _len_0 + 1 + end + end + tree.comments = _accum_0 + end + if #tree.comments == 0 then + tree.comments = nil + end + for i = #tree, 1, -1 do + if AST.is_syntax_tree(tree[i], "Comment") then + table.remove(tree, i) + end + end + tree = setmetatable(tree, cls) + cls.source_code_for_tree[tree] = userdata.source + if tree.__init then + tree:__init() + end + return tree +end local Parsers = { } local max_parser_version = 0 for version = 1, 999 do local _continue_0 = false repeat - if not (version == 4) then + if not (version == 4 or version == 3) then _continue_0 = true break end @@ -84,40 +117,8 @@ for version = 1, 999 do break end max_parser_version = version - local make_tree - make_tree = function(tree, userdata) - local cls = AST[tree.type] - tree.source = Source(userdata.filename, tree.start, tree.stop) - tree.start, tree.stop = nil, nil - tree.type = nil - do - local _accum_0 = { } - local _len_0 = 1 - for _index_0 = 1, #tree do - local t = tree[_index_0] - if AST.is_syntax_tree(t, "Comment") then - _accum_0[_len_0] = t - _len_0 = _len_0 + 1 - end - end - tree.comments = _accum_0 - end - if #tree.comments == 0 then - tree.comments = nil - end - for i = #tree, 1, -1 do - if AST.is_syntax_tree(tree[i], "Comment") then - table.remove(tree, i) - end - end - tree = setmetatable(tree, cls) - cls.source_code_for_tree[tree] = userdata.source - if tree.__init then - tree:__init() - end - return tree - end - Parsers[version] = make_parser(peg_file:read("*a"), make_tree) + local peg_contents = peg_file:read("*a") + Parsers[version] = make_parser(peg_contents, make_tree) peg_file:close() _continue_0 = true until true diff --git a/nomsu_compiler.moon b/nomsu_compiler.moon index 4300327..606bf45 100644 --- a/nomsu_compiler.moon +++ b/nomsu_compiler.moon @@ -38,10 +38,25 @@ table.copy = (t)-> setmetatable({k,v for k,v in pairs(t)}, getmetatable(t)) -- consider non-linear codegen, rather than doing thunks for things like comprehensions -- Re-implement nomsu-to-lua comment translation? +make_tree = (tree, userdata)-> + cls = AST[tree.type] + tree.source = Source(userdata.filename, tree.start, tree.stop) + tree.start, tree.stop = nil, nil + tree.type = nil + tree.comments = [t for t in *tree when AST.is_syntax_tree(t, "Comment")] + if #tree.comments == 0 then tree.comments = nil + for i=#tree,1,-1 + if AST.is_syntax_tree(tree[i], "Comment") + table.remove(tree, i) + tree = setmetatable(tree, cls) + cls.source_code_for_tree[tree] = userdata.source + if tree.__init then tree\__init! + return tree + Parsers = {} max_parser_version = 0 for version=1,999 - continue unless version == 4 -- TODO: remove + continue unless version == 4 or version == 3 -- TODO: remove peg_file = io.open("nomsu.#{version}.peg") if not peg_file and package.nomsupath for path in package.nomsupath\gmatch("[^;]+") @@ -49,21 +64,8 @@ for version=1,999 break if peg_file break unless peg_file max_parser_version = version - make_tree = (tree, userdata)-> - cls = AST[tree.type] - tree.source = Source(userdata.filename, tree.start, tree.stop) - tree.start, tree.stop = nil, nil - tree.type = nil - tree.comments = [t for t in *tree when AST.is_syntax_tree(t, "Comment")] - if #tree.comments == 0 then tree.comments = nil - for i=#tree,1,-1 - if AST.is_syntax_tree(tree[i], "Comment") - table.remove(tree, i) - tree = setmetatable(tree, cls) - cls.source_code_for_tree[tree] = userdata.source - if tree.__init then tree\__init! - return tree - Parsers[version] = make_parser(peg_file\read("*a"), make_tree) + peg_contents = peg_file\read("*a") + Parsers[version] = make_parser(peg_contents, make_tree) peg_file\close! MAX_LINE = 80 -- For beautification purposes, try not to make lines much longer than this value diff --git a/parser.lua b/parser.lua index 8a94314..5eed3be 100644 --- a/parser.lua +++ b/parser.lua @@ -113,126 +113,9 @@ setmetatable(NOMSU_DEFS, { end }) local Parser = { - version = 3, + version = 4, patterns = { } } -do - local peg_tidier = re.compile([[ file <- %nl* {~ (def/comment) (%nl+ (def/comment))* %nl* ~} - def <- anon_def / captured_def - anon_def <- ({ident} (" "*) ":" - {~ ((%nl " "+ def_line?)+) / def_line ~}) -> "%1 <- %2" - captured_def <- ({ident} (" "*) "(" {ident} ")" (" "*) ":" - {~ ((%nl " "+ def_line?)+) / def_line ~}) -> "%1 <- (({} {| %3 |} {} %%userdata) -> %2)" - def_line <- (err / [^%nl])+ - err <- ("(!!" { (!("!!)") .)* } "!!)") -> "(({} (%1) %%userdata) => error)" - ident <- [a-zA-Z_][a-zA-Z0-9_]* - comment <- "--" [^%nl]* - ]]) - for version = 1, Parser.version do - local peg_file = io.open("nomsu." .. tostring(version) .. ".peg") - if not peg_file and package.nomsupath then - for path in package.nomsupath:gmatch("[^;]+") do - peg_file = io.open(path .. "/nomsu." .. tostring(version) .. ".peg") - if peg_file then - break - end - end - end - assert(peg_file, "could not find nomsu .peg file") - local nomsu_peg = peg_tidier:match(peg_file:read('*a')) - peg_file:close() - Parser.patterns[version] = re.compile(nomsu_peg, NOMSU_DEFS) - end -end -local _anon_chunk = 0 -Parser.parse = function(nomsu_code, source, version) - if source == nil then - source = nil - end - if version == nil then - version = nil - end - source = source or nomsu_code.source - nomsu_code = tostring(nomsu_code) - if not (source) then - source = Source("anonymous chunk #" .. tostring(_anon_chunk), 1, #nomsu_code) - _anon_chunk = _anon_chunk + 1 - end - version = version or nomsu_code:match("^#![^\n]*nomsu[ ]+-V[ ]*([0-9.]+)") - local syntax_version = version and tonumber(version:match("^[0-9]+")) or Parser.version - local userdata = { - errors = { }, - source = source, - comments = { } - } - local tree = Parser.patterns[syntax_version]:match(nomsu_code, nil, userdata) - if not tree or type(tree) == 'number' then - error("In file " .. tostring(colored.blue(tostring(source or ""))) .. " failed to parse:\n" .. tostring(colored.onyellow(colored.black(nomsu_code)))) - end - if next(userdata.errors) then - local keys - do - local _accum_0 = { } - local _len_0 = 1 - for k, v in pairs(userdata.errors) do - _accum_0[_len_0] = k - _len_0 = _len_0 + 1 - end - keys = _accum_0 - end - table.sort(keys) - local errors - do - local _accum_0 = { } - local _len_0 = 1 - for _index_0 = 1, #keys do - local k = keys[_index_0] - _accum_0[_len_0] = userdata.errors[k] - _len_0 = _len_0 + 1 - end - errors = _accum_0 - end - error("Errors occurred while parsing (v" .. tostring(syntax_version) .. "):\n\n" .. table.concat(errors, "\n\n"), 0) - end - local comments - do - local _accum_0 = { } - local _len_0 = 1 - for p, c in pairs(userdata.comments) do - _accum_0[_len_0] = { - comment = c, - pos = p - } - _len_0 = _len_0 + 1 - end - comments = _accum_0 - end - table.sort(comments, function(a, b) - return a.pos > b.pos - end) - local comment_i = 1 - local walk_tree - walk_tree = function(t) - local comment_buff = { } - while comments[#comments] and comments[#comments].pos <= t.source.start do - table.insert(comment_buff, table.remove(comments)) - end - for _index_0 = 1, #t do - local x = t[_index_0] - if AST.is_syntax_tree(x) then - walk_tree(x) - end - end - while comments[#comments] and comments[#comments].pos <= t.source.stop do - table.insert(comment_buff, table.remove(comments)) - end - if #comment_buff > 0 then - t.comments = comment_buff - end - end - walk_tree(tree) - return tree -end Parser.is_operator = function(s) return not not (NOMSU_DEFS.operator_char ^ 1 * -1):match(s) end diff --git a/parser.moon b/parser.moon index f2e5b17..2adfe1b 100644 --- a/parser.moon +++ b/parser.moon @@ -74,73 +74,7 @@ setmetatable(NOMSU_DEFS, {__index:(key)=> return make_node }) -Parser = {version:3, patterns:{}} -do - -- Just for cleanliness, I put the language spec in its own file using a slightly modified - -- version of the lpeg.re syntax. - peg_tidier = re.compile [[ - file <- %nl* {~ (def/comment) (%nl+ (def/comment))* %nl* ~} - def <- anon_def / captured_def - anon_def <- ({ident} (" "*) ":" - {~ ((%nl " "+ def_line?)+) / def_line ~}) -> "%1 <- %2" - captured_def <- ({ident} (" "*) "(" {ident} ")" (" "*) ":" - {~ ((%nl " "+ def_line?)+) / def_line ~}) -> "%1 <- (({} {| %3 |} {} %%userdata) -> %2)" - def_line <- (err / [^%nl])+ - err <- ("(!!" { (!("!!)") .)* } "!!)") -> "(({} (%1) %%userdata) => error)" - ident <- [a-zA-Z_][a-zA-Z0-9_]* - comment <- "--" [^%nl]* - ]] - for version=1,Parser.version - peg_file = io.open("nomsu.#{version}.peg") - if not peg_file and package.nomsupath - for path in package.nomsupath\gmatch("[^;]+") - peg_file = io.open(path.."/nomsu.#{version}.peg") - break if peg_file - assert(peg_file, "could not find nomsu .peg file") - nomsu_peg = peg_tidier\match(peg_file\read('*a')) - peg_file\close! - Parser.patterns[version] = re.compile(nomsu_peg, NOMSU_DEFS) - -_anon_chunk = 0 -Parser.parse = (nomsu_code, source=nil, version=nil)-> - source or= nomsu_code.source - nomsu_code = tostring(nomsu_code) - unless source - source = Source("anonymous chunk ##{_anon_chunk}", 1, #nomsu_code) - _anon_chunk += 1 - version or= nomsu_code\match("^#![^\n]*nomsu[ ]+-V[ ]*([0-9.]+)") - syntax_version = version and tonumber(version\match("^[0-9]+")) or Parser.version - userdata = { - errors: {}, :source, comments: {} - } - tree = Parser.patterns[syntax_version]\match(nomsu_code, nil, userdata) - if not tree or type(tree) == 'number' - error "In file #{colored.blue tostring(source or "")} failed to parse:\n#{colored.onyellow colored.black nomsu_code}" - - if next(userdata.errors) - keys = [k for k,v in pairs(userdata.errors)] - table.sort(keys) - errors = [userdata.errors[k] for k in *keys] - error("Errors occurred while parsing (v#{syntax_version}):\n\n"..table.concat(errors, "\n\n"), 0) - - comments = [{comment:c, pos:p} for p,c in pairs(userdata.comments)] - -- Sort in descending order so we can pop the first comments off the end one at a time - table.sort comments, (a,b)-> a.pos > b.pos - comment_i = 1 - walk_tree = (t)-> - export comment_i - comment_buff = {} - while comments[#comments] and comments[#comments].pos <= t.source.start - table.insert(comment_buff, table.remove(comments)) - for x in *t - if AST.is_syntax_tree x - walk_tree x - while comments[#comments] and comments[#comments].pos <= t.source.stop - table.insert(comment_buff, table.remove(comments)) - t.comments = comment_buff if #comment_buff > 0 - walk_tree tree - - return tree +Parser = {version:4, patterns:{}} Parser.is_operator = (s)-> return not not (NOMSU_DEFS.operator_char^1 * -1)\match(s) diff --git a/pretty_errors.lua b/pretty_errors.lua index 7402703..ae6b48f 100644 --- a/pretty_errors.lua +++ b/pretty_errors.lua @@ -42,6 +42,7 @@ format_error = function(err) err_msg = err_msg .. "\n\027[2m" .. tostring(fmt_str:format(err_linenum)) .. tostring(err_line) .. "\027[0m" end local _, err_linenum_end, err_linepos_end = string2.line_at(err.source, err.stop) + err_linenum_end = err_linenum_end or err_linenum if err_linenum_end == err_linenum then err_msg = err_msg .. "\n" .. tostring(pointer) else diff --git a/pretty_errors.moon b/pretty_errors.moon index 3b583b7..b0d81f9 100644 --- a/pretty_errors.moon +++ b/pretty_errors.moon @@ -34,6 +34,7 @@ format_error = (err)-> err_line = "\027[0m#{before}\027[41;30m#{during}#{nl_indicator}\027[0m#{after}" err_msg ..= "\n\027[2m#{fmt_str\format(err_linenum)}#{err_line}\027[0m" _, err_linenum_end, err_linepos_end = string2.line_at(err.source, err.stop) + err_linenum_end or= err_linenum if err_linenum_end == err_linenum err_msg ..= "\n#{pointer}" else