--- 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 discussion
res["and"]
etc would be true
.
-- @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},
{STRING1,sdump},
{STRING2,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},
{STRING2,sdump},
{'^//.-\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 is_delim
if type(delim) == 'function' then
is_delim = delim
else
is_delim = function(t)
return t == delim
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 level == 1 and is_delim(token) 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
return lexer