Merge pull request #8921 from haskell/mergify/bp/3.10/pr-8896

Regenerate Lexer.hs with latest Alex (fix #8892) (backport #8896)

Author: mergify[bot]

.github/workflows/quick-jobs.yml

@@ -47,7 +47,7 @@ jobs:
       - name: Update Hackage index
         run: cabal v2-update
       - name: Install alex
-        run: cabal v2-install alex --constraint='alex ==3.2.6'
+        run: cabal v2-install alex --constraint='alex ==3.2.7.3'
       - uses: actions/checkout@v3
       - name: Regenerate files
         run: |

Cabal-syntax/src/Distribution/Fields/Lexer.hs

@@ -27,6 +27,7 @@ module Distribution.Fields.LexerMonad (
     LexWarning(..),
     LexWarningType(..),
     addWarning,
+    addWarningAt,
     toPWarnings,
 
   ) where
@@ -153,3 +154,8 @@ setStartCode c = Lex $ \s -> LexResult s{ curCode = c } ()
 addWarning :: LexWarningType -> Lex ()
 addWarning wt = Lex $ \s@LexState{ curPos = pos, warnings = ws  } ->
     LexResult s{ warnings = LexWarning wt pos : ws } ()
+
+-- | Add warning at specific position
+addWarningAt :: Position -> LexWarningType -> Lex ()
+addWarningAt pos wt = Lex $ \s@LexState{ warnings = ws  } ->
+    LexResult s{ warnings = LexWarning wt pos : ws } ()

Cabal-syntax/src/Distribution/Fields/LexerMonad.hs

@@ -1,4 +1,4 @@
-Octree-0.5.cabal:39:3: Non breaking spaces at 39:3, 41:3, 43:3
+Octree-0.5.cabal:39:1: Non breaking spaces at 39:1, 41:1, 43:1
 cabal-version: >=1.8
 name:          Octree
 version:       0.5

Cabal-tests/tests/ParserTests/regressions/Octree-0.5.format

@@ -1,4 +1,4 @@
-monad-param.cabal:19:3: Tabs used as indentation at 19:3, 20:3
+monad-param.cabal:19:1: Tabs used as indentation at 19:1, 20:1
 name:         monad-param
 version:      0.0.1
 license:      BSD3

Cabal-tests/tests/ParserTests/regressions/monad-param.format

@@ -1,4 +1,4 @@
-th-lift-instances.cabal:15:9: Tabs used as indentation at 15:9
+th-lift-instances.cabal:15:1: Tabs used as indentation at 15:1
 cabal-version:      >=1.10
 name:               th-lift-instances
 version:            0.1.4

Cabal-tests/tests/ParserTests/regressions/th-lift-instances.format

@@ -1,6 +1,6 @@
 # cabal check
 Warning: These warnings may cause trouble when distributing the package:
-Warning: pkg.cabal:1:2: Byte-order mark found at the beginning of the file
+Warning: pkg.cabal:1:1: Byte-order mark found at the beginning of the file
 Warning: The following errors will cause portability problems on other environments:
 Warning: ./pkg.cabal starts with an Unicode byte order mark (BOM). This may cause problems with older cabal versions.
 Warning: Hackage would reject this package.

cabal-testsuite/PackageTests/Check/PackageFiles/BOM/cabal.out

@@ -0,0 +1,11 @@
+synopsis: Regenerate Lexer.hs to avoid out-of-bound array access due to a bug in Alex
+packages: Cabal-syntax
+prs: #8896
+issues: #8892
+
+description: {
+
+- Regenerate Cabal-syntax's Lexer.hs with Alex 3.2.7.3 which includes a fix for
+  an out-of-bound array access (only noticeable with GHC's JavaScript backend).
+
+}

changelog.d/issue-8892

@@ -83,22 +83,22 @@ $instresc        = $printable
 tokens :-
 
 <0> {
-  @bom?  { \_ len _ -> do
-              when (len /= 0) $ addWarning LexWarningBOM
+  @bom?  { \pos len _ -> do
+              when (len /= 0) $ addWarningAt pos LexWarningBOM
               setStartCode bol_section
               lexToken
          }
 }
 
 <bol_section, bol_field_layout, bol_field_braces> {
-  @nbspspacetab* @nl         { \_pos len inp -> checkWhitespace len inp >> adjustPos retPos >> lexToken }
+  @nbspspacetab* @nl         { \pos len inp -> checkWhitespace pos len inp >> adjustPos retPos >> lexToken }
   -- no @nl here to allow for comments on last line of the file with no trailing \n
   $spacetab* "--" $comment*  ;  -- TODO: check the lack of @nl works here
                                 -- including counting line numbers
 }
 
 <bol_section> {
-  @nbspspacetab*   { \pos len inp -> checkLeadingWhitespace len inp >>
+  @nbspspacetab*   { \pos len inp -> checkLeadingWhitespace pos len inp >>
                                      if B.length inp == len
                                        then return (L pos EOF)
                                        else setStartCode in_section
@@ -123,7 +123,7 @@ tokens :-
 }
 
 <bol_field_layout> {
-  @nbspspacetab* { \pos len inp -> checkLeadingWhitespace len inp >>= \len' ->
+  @nbspspacetab* { \pos len inp -> checkLeadingWhitespace pos len inp >>= \len' ->
                                   if B.length inp == len
                                     then return (L pos EOF)
                                     else setStartCode in_field_layout
@@ -172,17 +172,17 @@ toki t pos  len  input = return $! L pos (t (B.take len input))
 tok :: Token -> Position -> Int -> ByteString -> Lex LToken
 tok  t pos _len _input = return $! L pos t
 
-checkLeadingWhitespace :: Int -> ByteString -> Lex Int
-checkLeadingWhitespace len bs
+checkLeadingWhitespace :: Position -> Int -> ByteString -> Lex Int
+checkLeadingWhitespace pos len bs
     | B.any (== 9) (B.take len bs) = do
-        addWarning LexWarningTab
-        checkWhitespace len bs
-    | otherwise = checkWhitespace len bs
+        addWarningAt pos LexWarningTab
+        checkWhitespace pos len bs
+    | otherwise = checkWhitespace pos len bs
 
-checkWhitespace :: Int -> ByteString -> Lex Int
-checkWhitespace len bs
+checkWhitespace :: Position -> Int -> ByteString -> Lex Int
+checkWhitespace pos len bs
     | B.any (== 194) (B.take len bs) = do
-        addWarning LexWarningNBSP
+        addWarningAt pos LexWarningNBSP
         return $ len - B.count 194 (B.take len bs)
     | otherwise = return len