jplexer.lua

-- Authors: Steve Donovan, David Manura.
-- Source: https://github.com/stevedonovan/Penlight/blob/master/lua/pl/lexer.lua
--[[
Copyright (C) 2009 Steve Donovan, David Manura.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
]]

--- Lexical scanner for creating a sequence of tokens from text.
-- `lexer.scan(s)` returns an iterator over all tokens found in the
-- string `s`. This iterator returns two values, a token type string
-- (such as 'string' for quoted string, 'iden' for identifier) and the value of the
-- token.
--
-- Versions specialized for Lua and C are available; these also handle block comments
-- and classify keywords as 'keyword' tokens. For example:
--
--    > s = 'for i=1,n do'
--    > for t,v in lexer.lua(s)  do print(t,v) end
--    keyword for
--    iden    i
--    =       =
--    number  1
--    ,       ,
--    iden    n
--    keyword do
--
-- See the Guide for further @{06-data.md.Lexical_Scanning|discussion}
-- @module pl.lexer

local yield,wrap = coroutine.yield,coroutine.wrap
local strfind = string.find
local strsub = string.sub
local append = table.insert

local function assert_arg(idx,val,tp)
    if type(val) ~= tp then
        error("argument "..idx.." must be "..tp, 2)
    end
end

local lexer = {}
jps.lexer = lexer

local NUMBER1 = '^[%+%-]?%d+%.?%d*[eE][%+%-]?%d+'
local NUMBER2 = '^[%+%-]?%d+%.?%d*'
local NUMBER3 = '^0x[%da-fA-F]+'
local NUMBER4 = '^%d+%.?%d*[eE][%+%-]?%d+'
local NUMBER5 = '^%d+%.?%d*'
local IDEN = '^[%a_][%w_]*'
local WSPACE = '^%s+'
local STRING0 = [[^(['\"]).-\\%1]]
local STRING1 = [[^(['\"]).-[^\]%1]]
local STRING3 = "^((['\"])%2)" -- empty string
local PREPRO = '^#.-[^\\]\n'

local plain_matches,lua_matches,cpp_matches,lua_keyword,cpp_keyword

local function tdump(tok)
    return yield(tok,tok)
end

local function ndump(tok,options)
    if options and options.number then
        tok = tonumber(tok)
    end
    return yield("number",tok)
end

-- regular strings, single or double quotes; usually we want them
-- without the quotes
local function sdump(tok,options)
    if options and options.string then
        tok = tok:sub(2,-2)
    end
    return yield("string",tok)
end

-- long Lua strings need extra work to get rid of the quotes
local function sdump_l(tok,options)
    if options and options.string then
        tok = tok:sub(3,-3)
    end
    return yield("string",tok)
end

local function chdump(tok,options)
    if options and options.string then
        tok = tok:sub(2,-2)
    end
    return yield("char",tok)
end

local function cdump(tok)
    return yield('comment',tok)
end

local function wsdump (tok)
    return yield("space",tok)
end

local function pdump (tok)
    return yield('prepro',tok)
end

local function plain_vdump(tok)
    return yield("iden",tok)
end

local function lua_vdump(tok)
    if lua_keyword[tok] then
        return yield("keyword",tok)
    else
        return yield("iden",tok)
    end
end

local function cpp_vdump(tok)
    if cpp_keyword[tok] then
        return yield("keyword",tok)
    else
        return yield("iden",tok)
    end
end

--- create a plain token iterator from a string or file-like object.
-- @param s the string
-- @param matches an optional match table (set of pattern-action pairs)
-- @param filter a table of token types to exclude, by default {space=true}
-- @param options a table of options; by default, {number=true,string=true},
-- which means convert numbers and strip string quotes.
function lexer.scan (s,matches,filter,options)
    --assert_arg(1,s,'string')
    local file = type(s) ~= 'string' and s
    filter = filter or {space=true}
    options = options or {number=true,string=true}
    if filter then
        if filter.space then filter[wsdump] = true end
        if filter.comments then
            filter[cdump] = true
        end
    end
    if not matches then
        if not plain_matches then
            plain_matches = {
                {WSPACE,wsdump},
                {NUMBER3,ndump},
                {IDEN,plain_vdump},
                {NUMBER1,ndump},
                {NUMBER2,ndump},
                {STRING3,sdump},
                {STRING0,sdump},
                {STRING1,sdump},
                {'^.',tdump}
            }
        end
        matches = plain_matches
    end
    local function lex ()
        local i1,i2,idx,res1,res2,tok,pat,fun,capt
        local line = 1
        if file then s = file:read()..'\n' end
        local sz = #s
        local idx = 1
        --print('sz',sz)
        while true do
            for _,m in ipairs(matches) do
                pat = m[1]
                fun = m[2]
                i1,i2 = strfind(s,pat,idx)
                if i1 then
                    tok = strsub(s,i1,i2)
                    idx = i2 + 1
                    if not (filter and filter[fun]) then
                        lexer.finished = idx > sz
                        res1,res2 = fun(tok,options)
                    end
                    if res1 then
                        local tp = type(res1)
                        -- insert a token list
                        if tp=='table' then
                            yield('','')
                            for _,t in ipairs(res1) do
                                yield(t[1],t[2])
                            end
                        elseif tp == 'string' then -- or search up to some special pattern
                            i1,i2 = strfind(s,res1,idx)
                            if i1 then
                                tok = strsub(s,i1,i2)
                                idx = i2 + 1
                                yield('',tok)
                            else
                                yield('','')
                                idx = sz + 1
                            end
                            --if idx > sz then return end
                        else
                            yield(line,idx)
                        end
                    end
                    if idx > sz then
                        if file then
                            --repeat -- next non-empty line
                                line = line + 1
                                s = file:read()
                                if not s then return end
                            --until not s:match '^%s*$'
                            s = s .. '\n'
                            idx ,sz = 1,#s
                            break
                        else
                            return
                        end
                    else break end
                end
            end
        end
    end
    return wrap(lex)
end

local function isstring (s)
    return type(s) == 'string'
end

--- insert tokens into a stream.
-- @param tok a token stream
-- @param a1 a string is the type, a table is a token list and
-- a function is assumed to be a token-like iterator (returns type & value)
-- @param a2 a string is the value
function lexer.insert (tok,a1,a2)
    if not a1 then return end
    local ts
    if isstring(a1) and isstring(a2) then
        ts = {{a1,a2}}
    elseif type(a1) == 'function' then
        ts = {}
        for t,v in a1() do
            append(ts,{t,v})
        end
    else
        ts = a1
    end
    tok(ts)
end

--- get everything in a stream upto a newline.
-- @param tok a token stream
-- @return a string
function lexer.getline (tok)
    local t,v = tok('.-\n')
    return v
end

--- get current line number. <br>
-- Only available if the input source is a file-like object.
-- @param tok a token stream
-- @return the line number and current column
function lexer.lineno (tok)
    return tok(0)
end

--- get the rest of the stream.
-- @param tok a token stream
-- @return a string
function lexer.getrest (tok)
    local t,v = tok('.+')
    return v
end

--- get the Lua keywords as a set-like table.
-- So <code>res["and"]</code> etc would be <code>true</code>.
-- @return a table
function lexer.get_keywords ()
    if not lua_keyword then
        lua_keyword = {
            ["and"] = true, ["break"] = true,  ["do"] = true,
            ["else"] = true, ["elseif"] = true, ["end"] = true,
            ["false"] = true, ["for"] = true, ["function"] = true,
            ["if"] = true, ["in"] = true,  ["local"] = true, ["nil"] = true,
            ["not"] = true, ["or"] = true, ["repeat"] = true,
            ["return"] = true, ["then"] = true, ["true"] = true,
            ["until"] = true,  ["while"] = true
        }
    end
    return lua_keyword
end


--- create a Lua token iterator from a string or file-like object.
-- Will return the token type and value.
-- @param s the string
-- @param filter a table of token types to exclude, by default {space=true,comments=true}
-- @param options a table of options; by default, {number=true,string=true},
-- which means convert numbers and strip string quotes.
function lexer.lua(s,filter,options)
    filter = filter or {space=true,comments=true}
    lexer.get_keywords()
    if not lua_matches then
        lua_matches = {
            {WSPACE,wsdump},
            {NUMBER3,ndump},
            {IDEN,lua_vdump},
            {NUMBER4,ndump},
            {NUMBER5,ndump},
            {STRING3,sdump},
            {STRING0,sdump},
            {STRING1,sdump},
            {'^%-%-%[%[.-%]%]',cdump},
            {'^%-%-.-\n',cdump},
            {'^%[%[.-%]%]',sdump_l},
            {'^==',tdump},
            {'^~=',tdump},
            {'^<=',tdump},
            {'^>=',tdump},
            {'^%.%.%.',tdump},
            {'^%.%.',tdump},
            {'^.',tdump}
        }
    end
    return lexer.scan(s,lua_matches,filter,options)
end

--- create a C/C++ token iterator from a string or file-like object.
-- Will return the token type type and value.
-- @param s the string
-- @param filter a table of token types to exclude, by default {space=true,comments=true}
-- @param options a table of options; by default, {number=true,string=true},
-- which means convert numbers and strip string quotes.
function lexer.cpp(s,filter,options)
    filter = filter or {comments=true}
    if not cpp_keyword then
        cpp_keyword = {
            ["class"] = true, ["break"] = true,  ["do"] = true, ["sizeof"] = true,
            ["else"] = true, ["continue"] = true, ["struct"] = true,
            ["false"] = true, ["for"] = true, ["public"] = true, ["void"] = true,
            ["private"] = true, ["protected"] = true, ["goto"] = true,
            ["if"] = true, ["static"] = true,  ["const"] = true, ["typedef"] = true,
            ["enum"] = true, ["char"] = true, ["int"] = true, ["bool"] = true,
            ["long"] = true, ["float"] = true, ["true"] = true, ["delete"] = true,
            ["double"] = true,  ["while"] = true, ["new"] = true,
            ["namespace"] = true, ["try"] = true, ["catch"] = true,
            ["switch"] = true, ["case"] = true, ["extern"] = true,
            ["return"] = true,["default"] = true,['unsigned']  = true,['signed'] = true,
            ["union"] =  true, ["volatile"] = true, ["register"] = true,["short"] = true,
        }
    end
    if not cpp_matches then
        cpp_matches = {
            {WSPACE,wsdump},
            {PREPRO,pdump},
            {NUMBER3,ndump},
            {IDEN,cpp_vdump},
            {NUMBER4,ndump},
            {NUMBER5,ndump},
            {STRING3,sdump},
            {STRING1,chdump},
            {'^//.-\n',cdump},
            {'^/%*.-%*/',cdump},
            {'^==',tdump},
            {'^!=',tdump},
            {'^<=',tdump},
            {'^>=',tdump},
            {'^->',tdump},
            {'^&&',tdump},
            {'^||',tdump},
            {'^%+%+',tdump},
            {'^%-%-',tdump},
            {'^%+=',tdump},
            {'^%-=',tdump},
            {'^%*=',tdump},
            {'^/=',tdump},
            {'^|=',tdump},
            {'^%^=',tdump},
            {'^::',tdump},
            {'^.',tdump}
        }
    end
    return lexer.scan(s,cpp_matches,filter,options)
end

--- get a list of parameters separated by a delimiter from a stream.
-- @param tok the token stream
-- @param endtoken end of list (default ')'). Can be '\n'
-- @param delim separator (default ',')
-- @return a list of token lists.
function lexer.get_separated_list(tok,endtoken,delim)
    endtoken = endtoken or ')'
    delim = delim or ','
    local parm_values = {}
    local level = 1 -- used to count ( and )
    local tl = {}
    local function tappend (tl,t,val)
        val = val or t
        append(tl,{t,val})
    end
    local is_end
    if endtoken == '\n' then
        is_end = function(t,val)
            return t == 'space' and val:find '\n'
        end
    else
        is_end = function (t)
            return t == endtoken
        end
    end
    local token,value
    while true do
        token,value=tok()
        if not token then return nil,'EOS' end -- end of stream is an error!
        if is_end(token,value) and level == 1 then
            append(parm_values,tl)
            break
        elseif token == '(' then
            level = level + 1
            tappend(tl,'(')
        elseif token == ')' then
            level = level - 1
            if level == 0 then -- finished with parm list
                append(parm_values,tl)
                break
            else
                tappend(tl,')')
            end
        elseif token == delim and level == 1 then
            append(parm_values,tl) -- a new parm
            tl = {}
        else
            tappend(tl,token,value)
        end
    end
    return parm_values,{token,value}
end

--- get the next non-space token from the stream.
-- @param tok the token stream.
function lexer.skipws (tok)
    local t,v = tok()
    while t == 'space' do
        t,v = tok()
    end
    return t,v
end

local skipws = lexer.skipws

--- get the next token, which must be of the expected type.
-- Throws an error if this type does not match!
-- @param tok the token stream
-- @param expected_type the token type
-- @param no_skip_ws whether we should skip whitespace
function lexer.expecting (tok,expected_type,no_skip_ws)
    assert_arg(1,tok,'function')
    assert_arg(2,expected_type,'string')
    local t,v
    if no_skip_ws then
        t,v = tok()
    else
        t,v = skipws(tok)
    end
    if t ~= expected_type then error ("expecting "..expected_type,2) end
    return v
end