Terminal-safe redexdump output

Makefile.am

@@ -40,6 +40,8 @@ AM_CPPFLAGS = \
 	-I$(top_srcdir)/opt/synth \
 	-I$(top_srcdir)/opt/unterface \
 	-I$(top_srcdir)/util \
+	-I$(top_srcdir)/tools/common \
+	-I$(top_srcdir)/tools/redexdump \
 	-I/usr/include/jsoncpp
 
 #
@@ -127,14 +129,23 @@ libredex_la_LIBADD = \
 #
 # redex-all: the main executable
 #
-bin_PROGRAMS = redex-all
+bin_PROGRAMS = redex-all redexdump
 
 redex_all_SOURCES = \
 	tools/redex-all/main.cpp \
 	tools/redex-all/Passes.cpp
 
 redex_all_LDADD = libredex.la
 
+redexdump_SOURCES = \
+	tools/redexdump/DumpTables.cpp \
+	tools/redexdump/PrintUtil.cpp \
+	tools/redexdump/RedexDump.cpp \
+	tools/common/DexCommon.cpp \
+	tools/common/Formatters.cpp
+
+redexdump_LDADD = libredex.la
+
 #
 # redex: Python driver script
 #

tools/redexdump/DumpTables.cpp

@@ -10,8 +10,10 @@
 #include "DexDebugInstruction.h"
 #include "Formatters.h"
 #include "PrintUtil.h"
+#include "Unicode.h"
 #include <RedexDump.h>
 #include <sstream>
+#include <vector>
 #include <string.h>
 #include <string>
 
@@ -432,14 +434,39 @@ static std::string get_debug_item(const uint8_t** pdebug_item) {
 }
 
 // Dump a string_data_item (i.e., an entry in the string data
-// section), advancing POS_INOUT over the item.  BUG: we dump the
-// MUTF8 string as if it were standard UTF8.  Shame on us.
+// section), advancing POS_INOUT over the item.
 static const char string_data_header[] = "u16len [contents]";
 static void dump_string_data_item(const uint8_t** pos_inout) {
   const uint8_t* pos = *pos_inout;
   uint32_t utf16_code_point_count = read_uleb128(&pos); // Not byte count!
   size_t utf8_length = strlen((char*) pos);
-  redump("%03u [%s]\n", (unsigned) utf16_code_point_count, pos);
+  std::string cleansed_data;
+  const char* string_to_print;
+  if (raw) { // Output whatever bytes we have
+    string_to_print = (char*) pos;
+  } else { // Translate to UTF-8; strip control characters
+    std::vector<char32_t> code_points;
+    const char* enc_pos = (char*) pos;
+    uint32_t cp;
+    while ((cp = mutf8_next_code_point(enc_pos))) {
+      if (cp < ' ' || cp == 255 /* DEL */) {
+        cp = '.';
+      }
+      code_points.push_back(cp);
+    }
+    ssize_t nr_utf8_bytes = utf32_to_utf8_length(
+      &code_points[0], code_points.size());
+    if (nr_utf8_bytes < 0 && utf8_length == 0) {
+      cleansed_data = "";
+    } else if (nr_utf8_bytes < 0) {
+      cleansed_data = "{invalid encoding?}";
+    } else {
+      cleansed_data.resize(nr_utf8_bytes);
+      utf32_to_utf8(&code_points[0], code_points.size(), &cleansed_data[0]);
+    }
+    string_to_print = cleansed_data.c_str();
+  }
+  redump("%03u [%s]\n", (unsigned) utf16_code_point_count, string_to_print);
   *pos_inout = pos + utf8_length + 1;
 }
 

tools/redexdump/PrintUtil.cpp

@@ -12,6 +12,7 @@
 #include <cstdarg>
 
 bool clean = false;
+bool raw = false;
 
 void redump(const char* format, ...) {
   va_list va;

tools/redexdump/PrintUtil.h

@@ -12,6 +12,7 @@
 #include <stdint.h>
 
 extern bool clean;
+extern bool raw;
 
 void redump(const char* format, ...);
 void redump(uint32_t off, const char* format, ...);

tools/redexdump/RedexDump.cpp

@@ -45,6 +45,7 @@ static const char ddump_usage_string[] =
     "-D, --ddebug=<addr>: disassemble debug info item at <addr>\n"
     "\nprinting options:\n"
     "--clean: does not print indexes or offsets making the output "
+    "--raw: print all bytes, even control characters "
     "usable for a diff\n";
 
 int main(int argc, char* argv[]) {
@@ -82,6 +83,7 @@ int main(int argc, char* argv[]) {
     { "debug", no_argument, nullptr, 'd' },
     { "ddebug", required_argument, nullptr, 'D' },
     { "clean", no_argument, (int*)&clean, 1 },
+    { "raw", no_argument, (int*)&raw, 1 },
     { "no-dump-map", no_argument, &no_dump_map, 1 },
     { "help", no_argument, nullptr, 'h' },
     { nullptr, 0, nullptr, 0 },