-- This is the `blogme2-inner.lua' file of blogme2.
-- It contains the "inner layer" of the kernel of blogme2 - just the parsers.
-- Author and version: Eduardo Ochs <edrx@mat.puc-rio.br>, 2005aug24
-- License: GPL.
-- (find-fline "INTERNALS")
-- (find-fline "blogme2-inner.lua")
-- (find-fline "blogme2-middle.lua")
-- (find-fline "blogme2-outer.lua")

-- The lowest-level low-level parsers
-- Remember (see the file INTERNALS):
--    BP    Long name/meaning      Corresponding Lua pattern
-- +------+----------------------+------------------------------------------+
-- | "%s" | space char           | "[ \t\n]"                                |
-- | "%w" | word char            | "[^%[%]]"                                |
-- | "%c" | normal char          | "[^ \t\n%[%]]"                           |
-- | "%B" | bracketed expression | "%b[]"                                   |
-- | "%W" | bigword              | "(%w*%b[]*)*" (but not the empty string) |
-- +------+----------------------+------------------------------------------+
--
subj = ""		-- what we are parsing ("" is a dummy value, of course)
pos = 0			-- current position; positions are 0-based
val = nil               -- the semantic value of parsing something
b,e = 0, 0		-- beginning and end of the text inside []s

_P = {}			-- table of low-level parsers

substring = function (b, e) return strsub(subj, b+1, e) end
parse     = function (blogmepat) return _P[blogmepat]() end
parsepat  = function (patstr)
    local _, __, newpos_ = string.find(subj, patstr, pos+1)
    if _ then pos = newpos_-1; return true end
  end
_P["%s*"] = function () return parsepat("^[ \t\n]*()") end
_P["%w+"] = function () return parsepat("^[^ \t\n%[%]]+()") end
_P["%c+"] = function () return parsepat("^[^%[%]]+()") end
_P["%B"]  = function () return parsepat("^%b[]()") end

_P["%w+:string"] = function ()
    local origpos = pos
    if parse("%w+") then val = substring(origpos, pos); return true end
  end
_P["%c+:string"] = function ()
    local origpos = pos
    if parse("%c+") then val = substring(origpos, pos); return true end
  end

--          Blogme parsers that invoke "%B:eval":
--        Name                   Input        Semantic value
-- +----------------------+------------------+--------------+
-- | "%W:eval"            | "a[lexp 22]bar"  | "a22bar"     |
-- | "%s*;%W:eval"        | " a[lexp 22]bar" | "a22bar"     |
-- | "(%s*;%W)*:evallist" | " a [lexp 22]  " | {"a", 22}    |
-- | "(%c+|%B)*:eval"     | " a [lexp 22]  " | " a 22  "    |
-- +----------------------+------------------+--------------+
--
_P["%W:eval"] = function ()             
    local result = {}
    while parse("%w+:string") or parse("%B:eval") do
      tinsert(result, val)
    end
    if getn(result) == 1 then val = result[1];      return true end
    if getn(result) >  1 then val = concat(result); return true end
  end
_P["%s*;%W:eval"] = function ()
    parse("%s*")
    return parse("%W:eval")
  end
_P["(%s*;%W)*:evallist"] = function ()
    local result = {}
    while parse("%s*;%W:eval") do tinsert(result, val) end
    val = result
    return true
  end
_P["(%c+|%B)*:eval"] = function ()             -- implicit concat
    local result = {}
    while parse("%c+:string") or parse("%B:eval") do
      tinsert(result, val or "")
    end
    val = concat(result)
    return true
  end

-- Parsers with human-readable names
_P["bigword:eval"] = _P["%s*;%W:eval"]         -- almost always a string (!)
_P["rest:string"]  = _P["(%c+|%B)*:eval"]      -- implicit eval
_P["rest:list"]    = _P["(%s*;%W)*:evallist"]  -- implicit eval
_P["rest:quote"]   = function () val = substring(pos, e); return true end

-- "%B:eval" itself. Parsing it can be so confusing...
--  [ head arg1 arg2 ]
-- /\pos                    instant 0
--  /\b              /\pos  instant 1
--  /\pos           /\e     instant 2
--       /\pos              instant 3
--
_P["%B:eval"] = function ()
    local oldb, olde = b, e  -- instant 0 (pos = before the "[")
    b = pos+1                --   b = after the "["
    if parse("%B") then      -- instant 1 (pos = after the "]")
      pos, e = b, pos-1      -- instant 2 (b=pos= after "[", e = before "]")
      parse("%s*;%W:eval")   -- instant 3 (pos = after the head)
      val = run_head(val)
      b, e, pos = oldb, olde, e+1
      return true
    end
    b, e = oldb, olde
  end

-- run_head: parse the args and run head(arg1, arg2, ...)
_W = {}  -- blogme words
_A = {}  -- argument-parsing functions for blogme words
functionp = function (obj) return type(obj) == "function" end
run_head = function (head)
    local headfun = _W[head] or _G[head]
    local argsfun = _A[head]
    if functionp(headfun) and functionp(argsfun) then
      return headfun(argsfun())
    else print("Bad head:", head)
         printpos("pos:", pos)	-- printpos is defined below
         printpos("b:", b)
         printpos("e:", e)
	 error()
    end
  end

-- printpos (I need to rewrite this)
blogme_input_fname = "?"
printpos = function (str, pos)
    printf("%s (progn (find-fline \"%s\") (goto-char %d))\n",
           str, blogme_input_fname, pos+1)
  end


-- (find-fline "blogme2-inner.lua")
-- (find-fline "blogme2-middle.lua")
-- (find-fline "blogme2-outer.lua")