Detect input file = output file (fixes #29)

ChangeLog

@@ -1,5 +1,8 @@
 2017-07-29  Jay Berkenbilt  <ejb@ql.org>
 
+	* Detect when input file and output file are the same and exit to
+	avoid overwriting and losing input file. Fixes #29.
+
 	* When passing multiple inspection arguments, run --check first,
 	and defer exit until after all the checks have been run. This
 	makes it possible to force operations such as --show-xref to be

include/qpdf/QUtil.hh

@@ -74,6 +74,9 @@ namespace QUtil
     QPDF_DLL
     qpdf_offset_t tell(FILE* stream);
 
+    QPDF_DLL
+    bool same_file(char const* name1, char const* name2);
+
     QPDF_DLL
     char* copy_string(std::string const&);
 

libqpdf/QUtil.cc

@@ -24,6 +24,7 @@
 #include <io.h>
 #else
 #include <unistd.h>
+#include <sys/stat.h>
 #endif
 
 std::string
@@ -188,6 +189,55 @@ QUtil::tell(FILE* stream)
 #endif
 }
 
+bool
+QUtil::same_file(char const* name1, char const* name2)
+{
+    if ((name1 == 0) || (strlen(name1) == 0) ||
+        (name2 == 0) || (strlen(name2) == 0))
+    {
+        return false;
+    }
+#ifdef _WIN32
+    HANDLE fh1 = CreateFile(name1, GENERIC_READ, FILE_SHARE_READ,
+                            NULL, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL);
+    HANDLE fh2 = CreateFile(name2, GENERIC_READ, FILE_SHARE_READ,
+                            NULL, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL);
+    BY_HANDLE_FILE_INFORMATION fi1;
+    BY_HANDLE_FILE_INFORMATION fi2;
+    bool same = false;
+    if ((fh1 != INVALID_HANDLE_VALUE) &&
+        (fh2 != INVALID_HANDLE_VALUE) &&
+        GetFileInformationByHandle(fh1, &fi1) &&
+        GetFileInformationByHandle(fh2, &fi2) &&
+        (fi1.dwVolumeSerialNumber == fi2.dwVolumeSerialNumber) &&
+        (fi1.nFileIndexLow == fi2.nFileIndexLow) &&
+        (fi1.nFileIndexHigh == fi2.nFileIndexHigh))
+    {
+        same = true;
+    }
+    if (fh1 != INVALID_HANDLE_VALUE)
+    {
+        CloseHandle(fh1);
+    }
+    if (fh2 != INVALID_HANDLE_VALUE)
+    {
+        CloseHandle(fh2);
+    }
+    return same;
+#else
+    struct stat st1;
+    struct stat st2;
+    if ((stat(name1, &st1) == 0) &&
+        (stat(name2, &st2) == 0) &&
+        (st1.st_ino == st2.st_ino) &&
+        (st1.st_dev == st2.st_dev))
+    {
+        return true;
+    }
+#endif
+    return false;
+}
+
 char*
 QUtil::copy_string(std::string const& str)
 {

libtests/qtest/qutil/other-file

@@ -0,0 +1 @@
+test

libtests/qtest/qutil/qutil.out

@@ -35,3 +35,10 @@ quack1
 quack2
 quack3
 quack4
+----
+file1: -qutil.out-, file2: -./qutil.out-; same: 1: PASS
+file1: -qutil.out-, file2: -qutil.out-; same: 1: PASS
+file1: -qutil.out-, file2: -other-file-; same: 0: PASS
+file1: -qutil.out-, file2: --; same: 0: PASS
+file1: -qutil.out-, file2: -(null)-; same: 0: PASS
+file1: --, file2: -qutil.out-; same: 0: PASS

libtests/qutil.cc

@@ -140,6 +140,36 @@ void get_whoami_test()
     print_whoami("a\\b\\c\\quack4.exe");
 }
 
+void assert_same_file(char const* file1, char const* file2, bool expected)
+{
+    bool actual = QUtil::same_file(file1, file2);
+    std::cout << "file1: -" << (file1 ? file1 : "(null)") << "-, file2: -"
+              << (file2 ? file2 : "(null)") << "-; same: "
+              << actual << ": " << ((actual == expected) ? "PASS" : "FAIL")
+              << std::endl;
+}
+
+void same_file_test()
+{
+    try
+    {
+        fclose(QUtil::safe_fopen("qutil.out", "r"));
+        fclose(QUtil::safe_fopen("other-file", "r"));
+    }
+    catch (std::exception)
+    {
+        std::cout << "same_file_test expects to have qutil.out and other-file"
+            " exist in the current directory\n";
+        return;
+    }
+    assert_same_file("qutil.out", "./qutil.out", true);
+    assert_same_file("qutil.out", "qutil.out", true);
+    assert_same_file("qutil.out", "other-file", false);
+    assert_same_file("qutil.out", "", false);
+    assert_same_file("qutil.out", 0, false);
+    assert_same_file("", "qutil.out", false);
+}
+
 int main(int argc, char* argv[])
 {
     try
@@ -155,6 +185,8 @@ int main(int argc, char* argv[])
 	to_utf8_test();
 	std::cout << "----" << std::endl;
 	get_whoami_test();
+	std::cout << "----" << std::endl;
+	same_file_test();
     }
     catch (std::exception& e)
     {

qpdf/qpdf.cc

@@ -1361,6 +1361,12 @@ int main(int argc, char* argv[])
 	usage("no output file may be given for this option");
     }
 
+    if (QUtil::same_file(infilename, outfilename))
+    {
+        QTC::TC("qpdf", "qpdf same file error");
+        usage("input file and output file are the same; this would cause input file to be lost");
+    }
+
     try
     {
 	QPDF pdf;

qpdf/qpdf.testcov

@@ -284,3 +284,4 @@ QPDFWriter preserve unreferenced standard 0
 QPDFObjectHandle non-stream in parsecontent 0
 QPDFObjectHandle errors in parsecontent 0
 QPDF stream with non-space 0
+qpdf same file error 0

qpdf/qtest/qpdf.test

@@ -206,7 +206,7 @@ $td->runtest("remove page we don't have",
 show_ntests();
 # ----------
 $td->notify("--- Miscellaneous Tests ---");
-$n_tests += 93;
+$n_tests += 94;
 
 $td->runtest("qpdf version",
 	     {$td->COMMAND => "qpdf --version"},
@@ -641,6 +641,10 @@ $td->runtest("dump corrected bad xref",
               $td->EXIT_STATUS => 3},
              $td->NORMALIZE_NEWLINES);
 
+$td->runtest("don't overwrite self",
+             {$td->COMMAND => "qpdf a.pdf a.pdf"},
+             {$td->REGEXP => "input file and output file are the same.*",
+              $td->EXIT_STATUS => 2});
 
 show_ntests();
 # ----------