From 29b9ffb4741f1d04734b289c398008337a5f031f Mon Sep 17 00:00:00 2001
From: Andreas Abel <andreas.abel@ifi.lmu.de>
Date: Tue, 5 Oct 2021 15:59:17 +0200
Subject: [PATCH] Fix #388: check token defs for empty regular expression

---
 source/CHANGELOG.md                   |  2 ++
 source/src/BNFC/GetCF.hs              | 24 ++++++++++++++++++++++--
 testing/fail-lbnf/388-empty-regex.cf  |  3 +++
 testing/fail-lbnf/388-empty-regex.err |  4 ++++
 4 files changed, 31 insertions(+), 2 deletions(-)
 create mode 100644 testing/fail-lbnf/388-empty-regex.cf
 create mode 100644 testing/fail-lbnf/388-empty-regex.err

diff --git a/source/CHANGELOG.md b/source/CHANGELOG.md
index 56fe6dd5..ffb87113 100644
--- a/source/CHANGELOG.md
+++ b/source/CHANGELOG.md
@@ -2,6 +2,8 @@
 
 Unreleased
 
+* LBNF: empty tokens types are now forbidden [#388]
+
 # 2.9.3
 
 Andreas Abel <andreas.abel@gu.se>  September 2021
diff --git a/source/src/BNFC/GetCF.hs b/source/src/BNFC/GetCF.hs
index 919b92c0..a02852e2 100644
--- a/source/src/BNFC/GetCF.hs
+++ b/source/src/BNFC/GetCF.hs
@@ -47,6 +47,7 @@ import System.IO   (hPutStrLn, stderr)
 -- Local imports:
 
 import qualified BNFC.Abs as Abs
+import BNFC.Abs (Reg(RAlts))
 import BNFC.Par
 
 import BNFC.CF
@@ -537,9 +538,16 @@ checkComments cf = concat
   where
   prags = cfgPragmas cf
 
--- | Check if any of the user-defined terminal categories is nullable.
+-- | Check if any of the user-defined terminal categories is nullable or empty.
 checkTokens :: CFG f -> Maybe String
-checkTokens cf
+checkTokens cf =
+  case catMaybes [ checkTokensEmpty cf, checkTokensNullable cf ] of
+    [] -> Nothing
+    ss -> Just $ concat ss
+
+-- | Check if any of the user-defined terminal categories is nullable.
+checkTokensNullable :: CFG f -> Maybe String
+checkTokensNullable cf
   | null pxs  = Nothing
   | otherwise = Just $ unlines $ concat
       [ [ "ERROR: The following tokens accept the empty string:" ]
@@ -548,6 +556,18 @@ checkTokens cf
   where
     pxs = [ px | TokenReg px _ regex <- cfgPragmas cf, nullable regex ]
 
+-- | Check if any of the user-defined terminal categories is nullable.
+checkTokensEmpty :: CFG f -> Maybe String
+checkTokensEmpty cf
+  | null pxs  = Nothing
+  | otherwise = Just $ unlines $ concat
+      [ [ "ERROR: The following tokens accept nothing:" ]
+      , printNames pxs
+      ]
+  where
+    -- The regular expression is already simplified, so we match against 0 directly.
+    pxs = [ px | TokenReg px _ (RAlts "") <- cfgPragmas cf ]
+
 
 -- we should actually check that
 -- (1) coercions are always between variants
diff --git a/testing/fail-lbnf/388-empty-regex.cf b/testing/fail-lbnf/388-empty-regex.cf
new file mode 100644
index 00000000..d774c0eb
--- /dev/null
+++ b/testing/fail-lbnf/388-empty-regex.cf
@@ -0,0 +1,3 @@
+-- 2021-10-05 Issue #388, empty regular expression
+
+token Empty [""]+ | [""] | (char - char);
diff --git a/testing/fail-lbnf/388-empty-regex.err b/testing/fail-lbnf/388-empty-regex.err
new file mode 100644
index 00000000..47ec4ccd
--- /dev/null
+++ b/testing/fail-lbnf/388-empty-regex.err
@@ -0,0 +1,4 @@
+ERROR: The following tokens accept nothing:
+  fail-lbnf/388-empty-regex.cf:3:7: Empty
+
+Aborting.  (Use option --force to continue despite errors.)