expat: Use upstream fuzzer build (retain dictionaries) (#3041)

4 files changed, 21 insertions, 77 deletions

diff --git a/projects/expat/Dockerfile b/projects/expat/Dockerfile
index d3499e7e..64efafff 100644
--- a/projects/expat/Dockerfile
+++ b/projects/expat/Dockerfile
@@ -16,8 +16,8 @@
 
 FROM gcr.io/oss-fuzz-base/base-builder
 MAINTAINER mike.aizatsky@gmail.com
-RUN apt-get update && apt-get install -y make autoconf automake libtool docbook2x
+RUN apt-get update && apt-get install -y make autoconf automake libtool docbook2x cmake
 
 RUN git clone --depth 1 https://github.com/libexpat/libexpat expat
 WORKDIR expat
-COPY build.sh parse_fuzzer.* *.dict $SRC/
+COPY build.sh *.dict $SRC/
diff --git a/projects/expat/build.sh b/projects/expat/build.sh
index b185038f..21c4f511 100755
--- a/projects/expat/build.sh
+++ b/projects/expat/build.sh
@@ -15,35 +15,25 @@
 #
 ################################################################################
 
-cd expat
+mkdir -p build
+cd build
+cmake ../expat -DEXPAT_BUILD_FUZZERS=ON -DEXPAT_OSSFUZZ_BUILD=ON -DEXPAT_SHARED_LIBS=OFF
+make -j$(nproc)
 
-./buildconf.sh
-./configure
-make clean
-make -j$(nproc) all
-
-ENCODING_TYPES="UTF_16 \
-  UTF_8 \
-  ISO_8859_1 \
-  US_ASCII \
-  UTF_16BE \
-  UTF_16LE"
-
-for encoding in $ENCODING_TYPES; do
-  fuzz_target_name=parse_${encoding}_fuzzer
-
-  $CXX $CXXFLAGS -std=c++11 -Ilib/ -DENCODING_${encoding} \
-      $SRC/parse_fuzzer.cc -o $OUT/${fuzz_target_name} \
-      $LIB_FUZZING_ENGINE lib/.libs/libexpat.a
-
-  # Use dictionaries in proper encoding for 16-bit encoding types.
-  if [[ $encoding == *"UTF_16"* ]]; then
-    cp $SRC/xml_${encoding}.dict  $OUT/${fuzz_target_name}.dict
+for fuzzer in fuzz/*;
+do
+  cp $fuzzer $OUT
+  fuzzer_name=$(basename $fuzzer)
+  if [[ ${fuzzer_name} =~ ^.*UTF-16$ ]];
+  then
+    cp $SRC/xml_UTF_16.dict $OUT/${fuzzer_name}.dict
+  elif [[ ${fuzzer_name} =~ ^.*UTF-16LE$ ]];
+  then
+    cp $SRC/xml_UTF_16LE.dict $OUT/${fuzzer_name}.dict
+  elif [[ ${fuzzer_name} =~ ^.*UTF-16BE$ ]];
+  then
+    cp $SRC/xml_UTF_16BE.dict $OUT/${fuzzer_name}.dict
   else
-    cp $SRC/xml.dict  $OUT/${fuzz_target_name}.dict
+    cp $SRC/xml.dict $OUT/${fuzzer_name}.dict
   fi
-
-  # Generate .option files for each fuzzer.
-  echo -en "[libfuzzer]\ndict = ${fuzz_target_name}.dict\nmax_len = 1024\n" \
-      > $OUT/${fuzz_target_name}.options
 done
diff --git a/projects/expat/parse_fuzzer.cc b/projects/expat/parse_fuzzer.cc
deleted file mode 100644
index ff4cccd0..00000000
--- a/projects/expat/parse_fuzzer.cc
+++ /dev/null
@@ -1,47 +0,0 @@
-// Copyright 2016 The Chromium Authors. All rights reserved.
-// Use of this source code is governed by a BSD-style license that can be
-// found in the LICENSE file.
-
-#include <stddef.h>
-#include <stdint.h>
-
-#include "expat.h"
-
-#include <functional>
-#include <string>
-
-const char* kEncoding =
-#if defined(ENCODING_UTF_16)
-"UTF-16"
-#elif defined(ENCODING_UTF_8)
-"UTF-8"
-#elif defined(ENCODING_ISO_8859_1)
-"ISO-8859-1"
-#elif defined(ENCODING_US_ASCII)
-"US-ASCII"
-#elif defined(ENCODING_UTF_16BE)
-"UTF-16BE"
-#elif defined(ENCODING_UTF_16LE)
-"UTF-16LE"
-#else
-#error Encoding type is not specified.
-#endif
-;
-
-// Entry point for LibFuzzer.
-extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) {
-  std::string input(reinterpret_cast<const char*>(data), size);
-  auto hash_salt = std::hash<std::string>()(input);
-
-  for (int use_ns = 0; use_ns <= 1; ++use_ns) {
-    XML_Parser parser =
-        use_ns ? XML_ParserCreateNS(kEncoding, '\n') :
-                 XML_ParserCreate(kEncoding);
-
-    // Set a hash salt to prevent MSan from crashing on random bytes generation.
-    XML_SetHashSalt(parser, hash_salt);
-    XML_Parse(parser, input.c_str(), input.size(), true);
-    XML_ParserFree(parser);
-  }
-  return 0;
-}
diff --git a/projects/expat/project.yaml b/projects/expat/project.yaml
index 987a5de8..0022a91b 100644
--- a/projects/expat/project.yaml
+++ b/projects/expat/project.yaml
@@ -4,6 +4,7 @@ auto_ccs:
  - "rhodri@kynesim.co.uk"
  - "hanno@hboeck.de"
  - "webmaster@hartwork.org"
+ - "bshas3@gmail.com"
 vendor_ccs:
  - "cdiehl@mozilla.com"
 sanitizers: