[flang-commits] [flang] d40e600 - [flang] Fix crash from empty -DMACRO= (bug #64837)

[Peter Klausler via flang-commits]

Author: Peter Klausler
Date: 2023-08-23T10:20:45-07:00
New Revision: d40e6005db992007d0d856f07eb6c5770a45389b

URL: https://github.com/llvm/llvm-project/commit/d40e6005db992007d0d856f07eb6c5770a45389b
DIFF: https://github.com/llvm/llvm-project/commit/d40e6005db992007d0d856f07eb6c5770a45389b.diff

LOG: [flang] Fix crash from empty -DMACRO= (bug #64837)

Some vector indexing code in the preprocessor fails with empty
tokens or token sequences in predefined macros.

Fixes https://github.com/llvm/llvm-project/issues/64837.

Differential Revision: https://reviews.llvm.org/D158451

Added: 
    flang/test/Preprocessing/bug64837.F90

Modified: 
    flang/lib/Parser/preprocessor.cpp
    flang/lib/Parser/token-sequence.h

Removed: 
    


################################################################################
diff  --git a/flang/lib/Parser/preprocessor.cpp b/flang/lib/Parser/preprocessor.cpp
index c78a9e1f4d9f48..5af16323d0129b 100644
--- a/flang/lib/Parser/preprocessor.cpp
+++ b/flang/lib/Parser/preprocessor.cpp
@@ -296,16 +296,18 @@ std::optional<TokenSequence> Preprocessor::MacroReplacement(
     if (!def->isFunctionLike()) {
       bool isRenaming{false};
       if (def->isPredefined()) {
-        std::string name{def->replacement().TokenAt(0).ToString()};
         std::string repl;
-        if (name == "__FILE__") {
-          repl = "\""s +
-              allSources_.GetPath(prescanner.GetCurrentProvenance()) + '"';
-        } else if (name == "__LINE__") {
-          std::string buf;
-          llvm::raw_string_ostream ss{buf};
-          ss << allSources_.GetLineNumber(prescanner.GetCurrentProvenance());
-          repl = ss.str();
+        if (!def->replacement().empty()) {
+          std::string name{def->replacement().TokenAt(0).ToString()};
+          if (name == "__FILE__") {
+            repl = "\""s +
+                allSources_.GetPath(prescanner.GetCurrentProvenance()) + '"';
+          } else if (name == "__LINE__") {
+            std::string buf;
+            llvm::raw_string_ostream ss{buf};
+            ss << allSources_.GetLineNumber(prescanner.GetCurrentProvenance());
+            repl = ss.str();
+          }
         }
         if (!repl.empty()) {
           ProvenanceRange insert{allSources_.AddCompilerInsertion(repl)};

diff  --git a/flang/lib/Parser/token-sequence.h b/flang/lib/Parser/token-sequence.h
index ab506da82dde6f..6b9e1f87ee0160 100644
--- a/flang/lib/Parser/token-sequence.h
+++ b/flang/lib/Parser/token-sequence.h
@@ -61,11 +61,17 @@ class TokenSequence {
   std::size_t SizeInTokens() const { return start_.size(); }
   std::size_t SizeInChars() const { return char_.size(); }
 
-  CharBlock ToCharBlock() const { return {&char_[0], char_.size()}; }
+  CharBlock ToCharBlock() const {
+    return char_.empty() ? CharBlock{} : CharBlock{&char_[0], char_.size()};
+  }
   std::string ToString() const { return ToCharBlock().ToString(); }
 
   CharBlock TokenAt(std::size_t token) const {
-    return {&char_[start_.at(token)], TokenBytes(token)};
+    if (auto bytes{TokenBytes(token)}) {
+      return {&char_[start_.at(token)], bytes};
+    } else { // char_ could be empty
+      return {};
+    }
   }
   char CharAt(std::size_t j) const { return char_.at(j); }
   CharBlock CurrentOpenToken() const {

diff  --git a/flang/test/Preprocessing/bug64837.F90 b/flang/test/Preprocessing/bug64837.F90
new file mode 100644
index 00000000000000..dadc830f5590fb
--- /dev/null
+++ b/flang/test/Preprocessing/bug64837.F90
@@ -0,0 +1,4 @@
+! RUN: %flang -E -DMACRO= %s 2>&1 | FileCheck %s
+!CHECK: integer, parameter :: p = +1
+integer, parameter :: p = MACRO +1
+end