{-# LANGUAGE Arrows #-}
{-
ParseSyntaxFiles.hs processes a directory containing Kate
XML syntax highlighting definitions. For each xml file in
the directory, it creates a syntax highlighting parser
Text/Highlighting/Kate/Syntax/(name).hs.
Finally, it creates a module
Text/Highlighting/Kate/Syntax.hs with a wrapper around
all these modules.
Usage: runghc ParseSyntaxFiles.hs xml
Requires HXT.
-}
module Main where
import Text.XML.HXT.Core
import Control.Arrow
import Control.Arrow.ArrowList
import Control.Monad (liftM)
import Data.List
import Data.Maybe
import Data.Char (toUpper, toLower, isAlphaNum)
import qualified Data.Map as Map
import System.Directory
import System.Environment
import System.Exit
import System.FilePath
import Text.PrettyPrint
import Text.Printf (printf)
import Data.Char (ord)
import Text.Highlighting.Kate.Definitions
import qualified Data.ByteString as B
import Data.ByteString.UTF8 (fromString, toString)
data SyntaxDefinition =
SyntaxDefinition { synLanguage :: String
, synAuthor :: String
, synVersion :: String
, synLicense :: String
, synExtensions :: String
, synCaseSensitive :: Bool
, synLists :: [(String, [String])]
, synContexts :: [SyntaxContext]
, synItemDatas :: [(String, String)]
, synKeywordAttr :: SyntaxKeywordAttr
} deriving (Read, Show)
data SyntaxKeywordAttr =
SyntaxKeywordAttr { keywordCaseSensitive :: Bool
, keywordDelims :: [Char]
} deriving (Read, Show)
data SyntaxContext =
SyntaxContext { contName :: String
, contAttribute :: String
, contLineEndContext :: String
, contLineBeginContext :: String
, contFallthrough :: Bool
, contFallthroughContext :: String
, contDynamic :: Bool
, contParsers :: [SyntaxParser]
} deriving (Read, Show)
data SyntaxParser =
SyntaxParser { parserType :: String
, parserAttribute :: String
, parserContext :: String
, parserLookAhead :: Bool
, parserIncludeAttrib :: Bool
, parserFirstNonSpace :: Bool
, parserColumn :: Maybe Int
, parserDynamic :: Bool
, parserString :: String -- could be a regex
, parserChar :: Char
, parserChar1 :: Char
, parserChildren :: [SyntaxParser]
} deriving (Read, Show)
-- | Converts a list of files (ending in .xml) and directories containing .xml files
-- into a list of .xml files.
argFiles :: [String] -> IO [String]
argFiles [] = error "Specify paths of xml files and/or directories containing xml syntax files."
argFiles args = do
let isXmlFile x = isSuffixOf ".xml" x
let (files, dirs) = partition isXmlFile args
dirContents <- mapM (\dir -> getDirectoryContents dir >>= return . map (combine dir) . filter isXmlFile) dirs
return $ nub (files ++ concat dirContents)
libraryPath = joinPath ["Text", "Highlighting", "Kate"]
destDir = joinPath [libraryPath, "Syntax"]
main :: IO ()
main = do
files <- getArgs >>= argFiles
destDirExists <- doesDirectoryExist destDir
if destDirExists
then return ()
else createDirectory destDir
mapM_ processOneFile files
let syntaxFile = combine libraryPath (addExtension "Syntax" "hs")
putStrLn $ "Writing " ++ syntaxFile
-- Get all syntax files, not only the newly generated ones.
names <- getDirectoryContents destDir >>= return . sort . map dropExtension . filter (isSuffixOf ".hs")
let imports = unlines $ map (\name -> "import qualified Text.Highlighting.Kate.Syntax." ++ name ++ " as " ++ name) names
let cases = unlines $ map (\name -> show (map toLower name) ++ " -> " ++ name ++ ".highlight") names
let languageExtensions = '[' :
(intercalate ", " $ map (\name -> "(" ++ show name ++ ", " ++ name ++ ".syntaxExtensions)") names) ++ "]"
syntaxFileTemplate <- liftM toString $ B.readFile (syntaxFile <.> "in")
let filledTemplate = fillTemplate 0 [("imports",imports),
("languages",show names),
("languageExtensions",languageExtensions),
("cases",cases)] syntaxFileTemplate
B.writeFile syntaxFile $ fromString filledTemplate
processOneFile :: FilePath -> IO ()
processOneFile src = do
[syntax] <- runX $ application src
let name = nameFromPath src
let outFile = joinPath [libraryPath, "Syntax", addExtension name "hs"]
let includeLangs = nub $ filter (/= name) $ map (drop 2 . parserContext) $
filter (\p -> (parserType p) == "IncludeRules" && "##" `isPrefixOf` (parserContext p)) $
concatMap contParsers $ synContexts syntax
let includeImports = map (("import qualified " ++) . langNameToModule) includeLangs
putStrLn $ "Writing " ++ outFile
B.writeFile outFile $ fromString $
"{- This module was generated from data in the Kate syntax highlighting file " ++ (takeFileName src) ++ ", version " ++
synVersion syntax ++ ",\n" ++
" by " ++ synAuthor syntax ++ " -}\n\n" ++
"module Text.Highlighting.Kate.Syntax." ++ name ++ " ( highlight, parseExpression, syntaxName, syntaxExtensions ) where\n\
\import Text.Highlighting.Kate.Definitions\n\
\import Text.Highlighting.Kate.Common\n" ++
unlines includeImports ++
"import Text.ParserCombinators.Parsec\n\
\import Control.Monad (when)\n\
\import Data.Map (fromList)\n\
\import Data.Maybe (fromMaybe, maybeToList)\n\n" ++
(if null (synLists syntax)
then ""
else "import qualified Data.Set as Set\n") ++
render (mkParser syntax) ++ "\n"
mkParser :: SyntaxDefinition -> Doc
mkParser syntax =
let name = text "-- | Full name of language." $$
text "syntaxName :: String" $$
text ("syntaxName = " ++ show (synLanguage syntax))
exts = text "-- | Filename extensions for this language." $$
text "syntaxExtensions :: String" $$
text ("syntaxExtensions = " ++ show (synExtensions syntax))
shortFormOf "dsKeyword" = ["kw"]
shortFormOf "dsDataType" = ["dt"]
shortFormOf "dsDecVal" = ["dv"]
shortFormOf "dsBaseN" = ["bn"]
shortFormOf "dsFloat" = ["fl"]
shortFormOf "dsChar" = ["ch"]
shortFormOf "dsString" = ["st"]
shortFormOf "dsComment" = ["co"]
shortFormOf "dsOthers" = ["ot"]
shortFormOf "dsAlert" = ["al"]
shortFormOf "dsFunction" = ["fu"]
shortFormOf "dsRegionMarker" = ["re"]
shortFormOf "dsError" = ["er"]
shortFormOf _ = []
styles = text ("styles = " ++ (show [(typ, shortsty) | (typ, sty) <- synItemDatas syntax, shortsty <- shortFormOf sty]))
withAttr = text "withAttribute attr txt = do" $$ (nest 2 $
text "when (null txt) $ fail \"Parser matched no text\"" $$
text "let labs = attr : maybeToList (lookup attr styles)" $$
text "st <- getState" $$
text "let oldCharsParsed = synStCharsParsedInLine st" $$
text "let prevchar = if null txt then '\\n' else last txt" $$
text "updateState $ \\st -> st { synStCharsParsedInLine = oldCharsParsed + length txt, synStPrevChar = prevchar } " $$
text "return (labs, txt)")
parseExpressionInternal = text "parseExpressionInternal = do" $$ (nest 2 $
text "context <- currentContext" $$
text "parseRules context <|> (pDefault >>= withAttribute (fromMaybe \"\" $ lookup context defaultAttributes))")
parseExpression = text "-- | Parse an expression using appropriate local context." $$
text "parseExpression :: GenParser Char SyntaxState LabeledSource" $$
text "parseExpression = do" $$ (nest 2 $
text "st <- getState" $$
text "let oldLang = synStLanguage st" $$
text ("setState $ st { synStLanguage = " ++ show (synLanguage syntax) ++ " }") $$
text ("context <- currentContext <|> (pushContext " ++ show (contName $ head $ synContexts syntax) ++
" >> currentContext)") $$
text "result <- parseRules context" $$
text "updateState $ \\st -> st { synStLanguage = oldLang }" $$
text "return result")
defaultAttributes = text $ "defaultAttributes = " ++ (show $ map (\cont -> (contName cont, contAttribute cont)) $ synContexts syntax)
-- Note: lineBeginContexts seems not to be used in any of the xml files
-- lineBeginContexts =
-- text $ "lineBeginContexts = " ++ (show $ map (\cont -> (contName cont, contLineBeginContext cont)) $ synContexts syntax)
startingContext = head (synContexts syntax)
contextCatchAll = text $ "parseRules x = fail $ \"Unknown context\" ++ x"
contexts = map (mkRules syntax) $ synContexts syntax
initialContextStack = Map.fromList [(synLanguage syntax, [contName startingContext])]
startingState = SyntaxState { synStContexts = initialContextStack
, synStLanguage = synLanguage syntax
, synStCurrentLine = ""
, synStCharsParsedInLine = 0
, synStPrevChar = '\n'
, synStCaseSensitive = synCaseSensitive syntax
, synStKeywordCaseSensitive = keywordCaseSensitive (synKeywordAttr syntax)
, synStCaptures = [] }
initState = text $ "startingState = " ++ show startingState
sourceLineParser = text "parseSourceLine = manyTill parseExpressionInternal pEndLine"
mainParser = text "parseSource = do " $$
(nest 2 $ text "lineContents <- lookAhead wholeLine" $$
text "updateState $ \\st -> st { synStCurrentLine = lineContents }" $$
-- text "context <- currentContext" $$
-- text "pushContext (fromMaybe \"#stay\" $ lookup context lineBeginContexts)" $$
text "result <- manyTill parseSourceLine eof" $$
text "return $ map normalizeHighlighting result")
mainFunction = text $ "-- | Highlight source code using this syntax definition.\n\
\highlight :: String -> Either String [SourceLine]\n\
\highlight input =\n\
\ case runParser parseSource startingState \"source\" input of\n\
\ Left err -> Left $ show err\n\
\ Right result -> Right result"
endLineParser = text "pEndLine = do" $$
(nest 2 $ text "lookAhead $ newline <|> (eof >> return '\\n')" $$
text "context <- currentContext" $$
text "case context of" $$
(nest 2 $ (vcat $ map (\cont -> text (show $ contName cont) <> text " -> " <>
switchContext (contLineEndContext cont) <>
if "#pop" `isPrefixOf` (contLineEndContext cont)
then text " >> pEndLine"
else text " >> pHandleEndLine") $ synContexts syntax) $$
(text $ "_ -> pHandleEndLine")))
{- text "pushContext (fromMaybe \"#stay\" $ lookup context lineBeginContexts)" $$ -}
-- we use 'words "blah blah2 blah3"' to keep ghc from inlining the list, which makes compiling take a long time
listDef (n, list) = text $ listName n ++ " = Set.fromList $ words $ " ++
show (if keywordCaseSensitive (synKeywordAttr syntax)
then unwords list
else map toLower (unwords list))
lists = vcat $ map listDef $ synLists syntax
regexDef re = text $ compiledRegexName re ++ " = compileRegex " ++ show re
regexes = vcat $ map regexDef $ nub $ [parserString x | x <- concatMap contParsers (synContexts syntax),
parserType x == "RegExpr", parserDynamic x == False]
in vcat $ intersperse (text "") $ [name, exts, mainFunction, parseExpression, mainParser, initState, sourceLineParser,
endLineParser, withAttr, styles, parseExpressionInternal, lists, regexes,
defaultAttributes {- , lineBeginContexts -}] ++ contexts ++ [contextCatchAll]
mkAlternatives :: [Doc] -> Doc
mkAlternatives docs =
let contents = vcat $ intersperse (text "<|>") docs
in if length docs > 1
then char '(' <> contents <> char ')'
else contents
mkRules :: SyntaxDefinition -> SyntaxContext -> Doc
mkRules syntax context =
let fallthroughParser = if contFallthrough context
then [parens (switchContext (contFallthroughContext context) <>
text " >> return ([], \"\")")]
else []
in text ("parseRules " ++ show (contName context) ++ " = ") $$
if null (contParsers context) && null fallthroughParser
then nest 2 (text "pzero")
else nest 2 $ (text "do (attr, result) <- " <>
(mkAlternatives $ (map (mkSyntaxParser syntax context) $ contParsers context) ++ fallthroughParser)) $$
text (" return (attr, result)")
mkSyntaxParser :: SyntaxDefinition -> SyntaxContext -> SyntaxParser -> Doc
mkSyntaxParser syntax context parser =
let mainParser = text $ case parserType parser of
"DetectChar" -> "pDetectChar " ++ show (parserDynamic parser) ++ " " ++ show (parserChar parser)
"Detect2Chars" -> "pDetect2Chars " ++ show (parserDynamic parser) ++ " " ++
show (parserChar parser) ++ " " ++ show (parserChar1 parser)
"AnyChar" -> "pAnyChar " ++ show (parserString parser)
"StringDetect" -> "pString " ++ show (parserDynamic parser) ++ " " ++ show (parserString parser)
"RegExpr" -> if parserDynamic parser
then "pRegExprDynamic " ++ show (parserString parser)
else "pRegExpr " ++ compiledRegexName (parserString parser)
"keyword" -> "pKeyword " ++ show (keywordDelims $ synKeywordAttr syntax) ++ " " ++ list
where list = case lookup string (synLists syntax) of
Just _ -> listName string
Nothing -> "Set.empty"
string = parserString parser
"Int" -> "pInt"
"Float" -> "pFloat"
"HlCOct" -> "pHlCOct"
"HlCHex" -> "pHlCHex"
"HlCStringChar" -> "pHlCStringChar"
"HlCChar" -> "pHlCChar"
"RangeDetect" -> "pRangeDetect " ++ show (parserChar parser) ++ " " ++ show (parserChar1 parser)
"LineContinue" -> "pLineContinue"
"IncludeRules" -> case parserContext parser of
('#':'#':xs) -> langNameToModule xs ++ ".parseExpression" ++
if parserIncludeAttrib parser || null (parserAttribute parser)
then ""
else " >>= ((withAttribute " ++ show (parserAttribute parser) ++ ") . snd)"
xs -> "parseRules " ++ show xs
"DetectSpaces" -> "pDetectSpaces"
"DetectIdentifier" -> "pDetectIdentifier"
_ -> "pUnimplemented"
parserDoc = char '(' <>
(case (parserColumn parser) of
Just c -> text $ "pColumn " ++ show c ++ " >> "
_ -> empty) <>
(if parserFirstNonSpace parser
then text "pFirstNonSpace >> "
else empty) <>
if parserType parser == "IncludeRules"
then mainParser <> char ')'
else (if parserLookAhead parser
then text "lookAhead (" <> mainParser <> text ") >> return ([],\"\") "
else mainParser <> text " >>= withAttribute " <>
text (if null (parserAttribute parser)
then show (contAttribute context)
else show (parserAttribute parser))) <>
char ')' <>
(if parserContext parser `elem` ["", "#stay"]
then empty
else text " >>~ " <> switchContext (parserContext parser))
childParsers = parserChildren parser
in char '(' <>
(if null childParsers
then parserDoc
else text "withChildren " <> parserDoc <> char ' ' <> (mkAlternatives $ map (mkSyntaxParser syntax context) childParsers)) <>
char ')'
switchContext next =
case next of
x | "#pop" `isPrefixOf` x -> char '(' <>
text (concat $ intersperse " >> " $ replicate (length (filter (=='#') x)) "popContext") <> char ')'
"#stay" -> text "return ()"
x -> text ("pushContext " ++ show x)
langNameToModule str = "Text.Highlighting.Kate.Syntax." ++
case str of
"Alerts" -> "Alert"
"Alerts_indent" -> "Alert_indent"
"C++" -> "Cpp"
"CSS" -> "Css"
"Doxygen" -> "Doxygen"
"HTML" -> "Html"
"Javadoc" -> "Javadoc"
"JavaScript" -> "Javascript"
"SQL (MySQL)" -> "SqlMysql"
"DoxygenLua" -> "Doxygenlua"
x -> x
listName :: String -> String
listName n = "list_" ++ normalize n
compiledRegexName :: String -> String
compiledRegexName n = "regex_" ++ normalize n
normalize :: String -> String
normalize "" = ""
normalize (x:xs) | isAlphaNum x = x : normalize xs
normalize (' ':xs) = '_':normalize xs
normalize (x:xs) = printf "'%2x" (ord x) ++ normalize xs
capitalize :: String -> String
capitalize (x:xs) = toUpper x : xs
capitalize [] = []
nameFromPath :: FilePath -> String
nameFromPath = concat . map capitalize . words .
(map (\c -> if c == '-' then ' ' else c)) . takeFileName .
dropExtension
application :: String -> IOSArrow b SyntaxDefinition
application src
= readDocument [withValidate no, withInputEncoding utf8] src
>>>
multi (hasName "language")
>>>
extractSyntaxDefinition
extractSyntaxDefinition :: IOSArrow XmlTree SyntaxDefinition
extractSyntaxDefinition = proc x -> do
lang <- getAttrValue "name" -< x
author <- getAttrValue "author" -< x
version <- getAttrValue "version" -< x
license <- getAttrValue "license" -< x
sources <- getAttrValue "extensions" -< x
caseSensitive <- getAttrValue "casesensitive" -< x
itemdatas <- getItemDatas -< x
lists <- getLists -< x
contexts <- getContexts -< x
keywordAttr <- getKeywordAttrs -< x
returnA -< SyntaxDefinition { synLanguage = lang
, synAuthor = author
, synVersion = version
, synLicense = license
, synExtensions = sources
, synCaseSensitive = vBool True caseSensitive
, synLists = lists
, synContexts = contexts
, synItemDatas = itemdatas
, synKeywordAttr = if null keywordAttr
then defaultKeywordAttr
else head keywordAttr }
getItemDatas :: IOSArrow XmlTree [(String,String)]
getItemDatas = multi (hasName "itemDatas")
>>>
(listA $ getChildren
>>>
hasName "itemData"
>>>
getAttrValue "name" &&& getAttrValue "defStyleNum")
getLists :: IOSArrow XmlTree [(String, [String])]
getLists = listA $ multi (hasName "list")
>>>
getAttrValue "name" &&& getListContents
getListContents :: IOSArrow XmlTree [String]
getListContents = listA $ getChildren
>>>
hasName "item"
>>>
getChildren
>>>
getText
>>>
arr stripWhitespace
getContexts :: IOSArrow XmlTree [SyntaxContext]
getContexts = listA $ multi (hasName "context")
>>>
proc x -> do
name <- getAttrValue "name" -< x
attribute <- getAttrValue "attribute" -< x
lineEndContext <- getAttrValue "lineEndContext" -< x
lineBeginContext <- getAttrValue "lineBeginContext" -< x
fallthrough <- getAttrValue "fallthrough" -< x
fallthroughContext <- getAttrValue "fallthroughContext" -< x
dynamic <- getAttrValue "dynamic" -< x
parsers <- getParsers -< x
returnA -< SyntaxContext
{ contName = name
, contAttribute = attribute
, contLineEndContext = if null lineEndContext then "#stay" else lineEndContext
, contLineBeginContext = if null lineBeginContext then "#stay" else lineBeginContext
, contFallthrough = vBool False fallthrough
, contFallthroughContext = if null fallthroughContext then "#stay" else fallthroughContext
, contDynamic = vBool False dynamic
, contParsers = parsers }
getParsers :: IOSArrow XmlTree [SyntaxParser]
getParsers = listA $ getChildren
>>>
proc x -> do
name <- getName -< x
attribute <- getAttrValue "attribute" -< x
context <- getAttrValue "context" -< x
char0 <- getAttrValue "char" -< x
char1 <- getAttrValue "char1" -< x
str <- getAttrValue "String" -< x
includeAttrib <- getAttrValue "includeAttrib" -< x
lookahead <- getAttrValue "lookAhead" -< x
firstNonSpace <- getAttrValue "firstNonSpace" -< x
column <- getAttrValue "column" -< x
dynamic <- getAttrValue "dynamic" -< x
children <- getParsers -< x
let tildeRegex = name == "RegExpr" && length str > 0 && head str == '^'
returnA -< SyntaxParser
{ parserType = name
, parserAttribute = attribute
, parserContext = context
, parserLookAhead = vBool False lookahead
, parserIncludeAttrib = vBool False includeAttrib
, parserFirstNonSpace = vBool False firstNonSpace
, parserColumn = if tildeRegex
then Just 0
else if null column
then Nothing
else Just (read column)
, parserDynamic = vBool False dynamic
, parserString = if tildeRegex then drop 1 str else str
, parserChar = if length char0 == 1 then head char0 else '*'
, parserChar1 = if length char1 == 1 then head char1 else '*'
, parserChildren = children }
getKeywordAttrs :: IOSArrow XmlTree [SyntaxKeywordAttr]
getKeywordAttrs = listA $ multi $ hasName "keywords"
>>>
proc x -> do
caseSensitive <- getAttrValue "casesensitive" -< x
weakDelim <- getAttrValue "weakDeliminator" -< x
additionalDelim <- getAttrValue "additionalDeliminator" -< x
returnA -< SyntaxKeywordAttr
{ keywordCaseSensitive = vBool True caseSensitive
, keywordDelims = (standardDelims ++ additionalDelim) \\ weakDelim }
standardDelims = " \n\t.():!+,-<=>%&*/;?[]^{|}~\\"
defaultKeywordAttr = SyntaxKeywordAttr { keywordCaseSensitive = True
, keywordDelims = standardDelims }
stripWhitespaceLeft = dropWhile isWhitespace
isWhitespace x = elem x [' ', '\t', '\n']
stripWhitespace = reverse . stripWhitespaceLeft . reverse . stripWhitespaceLeft
vBool :: Bool -> String -> Bool
vBool defaultVal value = case value of
z | z `elem` ["true","yes","1"] -> True
z | z `elem` ["false","no","0"] -> False
_ -> defaultVal
-- | Fill template. The template variables in the source text are
-- surrounded by @'s: e.g., @myvar@.
fillTemplate :: Int -> [(String,String)] -> String -> String
fillTemplate _ _ [] = []
fillTemplate _ [] lst = lst
fillTemplate n subs ('\n':xs) = '\n' : fillTemplate 0 subs xs
fillTemplate n subs ('@':xs) =
let (pref, suff) = break (=='@') xs
in if length pref > 0 && all isAlphaNum pref && length suff > 0
then case lookup pref subs of
Just v -> intercalate ('\n':replicate n ' ') (lines v) ++
fillTemplate (n + length v) subs (tail suff)
Nothing -> '@' : fillTemplate (n+1) subs xs
else '@' : fillTemplate (n+1) subs xs
fillTemplate n subs (x:xs) = x : fillTemplate (n+1) subs xs