Parse indices of associative arrays properly

ShellCheck/AST.hs

@@ -21,6 +21,7 @@ module ShellCheck.AST where
 
 import Control.Monad
 import Control.Monad.Identity
+import Text.Parsec
 import qualified ShellCheck.Regex as Re
 
 data Id = Id Int deriving (Show, Eq, Ord)
@@ -51,6 +52,8 @@ data Token =
     | T_Arithmetic Id Token
     | T_Array Id [Token]
     | T_IndexedElement Id Token Token
+    -- Store the index as string, and parse as arithmetic or string later
+    | T_UnparsedIndex Id SourcePos String
     | T_Assignment Id AssignmentMode String (Maybe Token) Token
     | T_Backgrounded Id Token
     | T_Backticked Id [Token]
@@ -145,7 +148,7 @@ tokenEquals a b = kludge a == kludge b
 instance Eq Token where
     (==) = tokenEquals
 
-analyze :: Monad m => (Token -> m ()) -> (Token -> m ()) -> (Token -> Token) -> Token -> m Token
+analyze :: Monad m => (Token -> m ()) -> (Token -> m ()) -> (Token -> m Token) -> Token -> m Token
 analyze f g i =
     round
   where
@@ -153,7 +156,7 @@ analyze f g i =
         f t
         newT <- delve t
         g t
-        return . i $ newT
+        i newT
     roundAll = mapM round
 
     roundMaybe Nothing = return Nothing
@@ -363,10 +366,11 @@ getId t = case t of
         T_CoProc id _ _ -> id
         T_CoProcBody id _ -> id
         T_Include id _ _ -> id
+        T_UnparsedIndex id _ _ -> id
 
 blank :: Monad m => Token -> m ()
 blank = const $ return ()
-doAnalysis f = analyze f blank id
-doStackAnalysis startToken endToken = analyze startToken endToken id
-doTransform i = runIdentity . analyze blank blank i
+doAnalysis f = analyze f blank (return . id)
+doStackAnalysis startToken endToken = analyze startToken endToken (return . id)
+doTransform i = runIdentity . analyze blank blank (return . i)
 

ShellCheck/ASTLib.hs

@@ -21,6 +21,7 @@ module ShellCheck.ASTLib where
 
 import ShellCheck.AST
 
+import Control.Monad.Writer
 import Control.Monad
 import Data.List
 import Data.Maybe
@@ -251,3 +252,16 @@ getCommandSequences t =
         T_IfExpression _ thens elses -> map snd thens ++ [elses]
         otherwise -> []
 
+getAssociativeArrays t =
+    nub . execWriter $ doAnalysis f t
+  where
+    f :: Token -> Writer [String] ()
+    f t@(T_SimpleCommand {}) = fromMaybe (return ()) $ do
+        name <- getCommandName t
+        guard $ name == "declare"
+        let flags = getAllFlags t
+        guard $ elem "A" $ map snd flags
+        let args = map fst . filter ((==) "" . snd) $ flags
+        let names = mapMaybe getLiteralString args
+        return $ tell names
+    f _ = return ()

ShellCheck/Analytics.hs

@@ -1322,7 +1322,9 @@ checkBraceExpansionVars params t@(T_BraceExpansion id list) = mapM_ check list
         (`isUnqualifiedCommand` "eval") <$> getClosestCommand (parentMap params) t
 checkBraceExpansionVars _ _ = return ()
 
-prop_checkForDecimals = verify checkForDecimals "((3.14*c))"
+prop_checkForDecimals1 = verify checkForDecimals "((3.14*c))"
+prop_checkForDecimals2 = verify checkForDecimals "foo[1.2]=bar"
+prop_checkForDecimals3 = verifyNot checkForDecimals "declare -A foo; foo[1.2]=bar"
 checkForDecimals params t@(TA_Expansion id _) = potentially $ do
     guard $ not (hasFloatingPoint params)
     str <- getLiteralString t
@@ -2116,6 +2118,8 @@ prop_checkUnassignedReferences19= verifyNotTree checkUnassignedReferences "reado
 prop_checkUnassignedReferences20= verifyNotTree checkUnassignedReferences "printf -v foo bar; echo $foo"
 prop_checkUnassignedReferences21= verifyTree checkUnassignedReferences "echo ${#foo}"
 prop_checkUnassignedReferences22= verifyNotTree checkUnassignedReferences "echo ${!os*}"
+prop_checkUnassignedReferences23= verifyTree checkUnassignedReferences "declare -a foo; foo[bar]=42;"
+prop_checkUnassignedReferences24= verifyNotTree checkUnassignedReferences "declare -A foo; foo[bar]=42;"
 checkUnassignedReferences params t = warnings
   where
     (readMap, writeMap) = execState (mapM tally $ variableFlow params) (Map.empty, Map.empty)

ShellCheck/Parser.hs

@@ -897,6 +897,20 @@ readNormalishWord end = do
     checkPossibleTermination pos x
     return $ T_NormalWord id x
 
+readIndexSpan = do
+    id <- getNextId
+    x <- many (readNormalWordPart "]" <|> someSpace <|> otherLiteral)
+    return $ T_NormalWord id x
+  where
+    someSpace = do
+        id <- getNextId
+        str <- spacing1
+        return $ T_Literal id str
+    otherLiteral = do
+        id <- getNextId
+        str <- many1 $ oneOf quotableChars
+        return $ T_Literal id str
+
 checkPossibleTermination pos [T_Literal _ x] =
     when (x `elem` ["do", "done", "then", "fi", "esac"]) $
         parseProblemAt pos WarningC 1010 $ "Use semicolon or linefeed before '" ++ x ++ "' (or quote to make it literal)."
@@ -2251,7 +2265,9 @@ readEvalSuffix = many1 (readIoRedirect <|> readCmdWord <|> evalFallback)
 
 -- Get whatever a parser would parse as a string
 readStringForParser parser = do
+    state <- Ms.get
     pos <- lookAhead (parser >> getPosition)
+    Ms.put state
     readUntil pos
   where
     readUntil endPos = anyChar `reluctantlyTill` (getPosition >>= guard . (== endPos))
@@ -2316,11 +2332,12 @@ readAssignmentWord = try $ do
         return $ T_Literal id ""
 
 readArrayIndex = do
+    id <- getNextId
     char '['
-    optional space
-    x <- readArithmeticContents
+    pos <- getPosition
+    str <- readStringForParser readIndexSpan
     char ']'
-    return x
+    return $ T_UnparsedIndex id pos str
 
 readArray = called "array assignment" $ do
     id <- getNextId
@@ -2477,12 +2494,12 @@ verifyEof = eof <|> choice [
         try (lookAhead p)
         action
 
-prop_readScript1 = isOk readScript "#!/bin/bash\necho hello world\n"
-prop_readScript2 = isWarning readScript "#!/bin/bash\r\necho hello world\n"
-prop_readScript3 = isWarning readScript "#!/bin/bash\necho hello\xA0world"
-prop_readScript4 = isWarning readScript "#!/usr/bin/perl\nfoo=("
-prop_readScript5 = isOk readScript "#!/bin/bash\n#This is an empty script\n\n"
-readScript = do
+prop_readScript1 = isOk readScriptFile "#!/bin/bash\necho hello world\n"
+prop_readScript2 = isWarning readScriptFile "#!/bin/bash\r\necho hello world\n"
+prop_readScript3 = isWarning readScriptFile "#!/bin/bash\necho hello\xA0world"
+prop_readScript4 = isWarning readScriptFile "#!/usr/bin/perl\nfoo=("
+prop_readScript5 = isOk readScriptFile "#!/bin/bash\n#This is an empty script\n\n"
+readScriptFile = do
     id <- getNextId
     pos <- getPosition
     optional $ do
@@ -2497,7 +2514,8 @@ readScript = do
             annotations <- readAnnotations
             commands <- withAnnotations annotations readCompoundListOrEmpty
             verifyEof
-            return $ T_Annotation annotationId annotations $  T_Script id sb commands
+            let script = T_Annotation annotationId annotations $  T_Script id sb commands
+            reparseIndices script
         else do
             many anyChar
             return $ T_Script id sb []
@@ -2549,6 +2567,9 @@ readScript = do
 
     readUtf8Bom = called "Byte Order Mark" $ string "\xFEFF"
 
+readScript = do
+    script <- readScriptFile
+    reparseIndices script
 
 isWarning p s = parsesCleanly p s == Just False
 isOk p s =      parsesCleanly p s == Just True
@@ -2635,6 +2656,33 @@ parseShell sys name contents = do
     second (ContextName pos str) = ParseNote pos InfoC 1009 $
         "The mentioned parser error was in this " ++ str ++ "."
 
+-- Go over all T_UnparsedIndex and reparse them as either arithmetic or text
+-- depending on declare -A statements.
+reparseIndices root =
+   analyze blank blank f root
+  where
+    associative = getAssociativeArrays root
+    isAssociative s = s `elem` associative
+    f (T_Assignment id mode name (Just (T_UnparsedIndex _ pos src)) value) = do
+        new <- parsed name pos src
+        return $ T_Assignment id mode name (Just new) value
+    f (T_Assignment id mode name Nothing (T_Array id2 words)) = do
+        newwords <- mapM (fix name) words
+        return $ T_Assignment id mode name Nothing (T_Array id2 newwords)
+    f t = return t
+
+    fix name word =
+        case word of
+            T_IndexedElement id (T_UnparsedIndex _ pos src) value -> do
+                new <- parsed name pos src
+                return $ T_IndexedElement id new value
+            otherwise -> return word
+
+    parsed name pos src =
+        if isAssociative name
+        then subParse pos readIndexSpan src
+        else subParse pos (optional space >> readArithmeticContents) src
+
 reattachHereDocs root map =
     doTransform f root
   where