[ENH] Properly handle empty, yet existent files. Thanks to @jpaulLD and

@jmarshall.
arq5x · Dec 27, 2012 · ca50f59 · ca50f59
1 parent 2a16ea9
commit ca50f59
Show file tree

Hide file tree

Showing 8 changed files with 77 additions and 25 deletions.
diff --git a/src/utils/bedFile/bedFile.cpp b/src/utils/bedFile/bedFile.cpp
@@ -102,10 +102,14 @@ void BedFile::Open(void) {
             delete _bedStream;
             _bedStream = new igzstream(bedFile.c_str(), ios::in);
         }
-        if ( !(_bedStream->good()) ) {
-            cerr << "Error: The requested bed file (" 
-                 << bedFile 
-                 << ") could not be opened. Exiting!" << endl;
+        if ( _bedStream->fail() ) {
+            cerr << "Error: The requested file (" 
+                 << bedFile
+                 << ") " 
+                 << "could not be opened. "
+                 << "Error message: ("
+                 << strerror(errno)
+                 << "). Exiting!" << endl;
             exit (1);
         }
     }

diff --git a/src/utils/bedFilePE/bedFilePE.cpp b/src/utils/bedFilePE/bedFilePE.cpp
@@ -36,8 +36,14 @@ void BedFilePE::Open(void) {
             _bedStream = new igzstream(bedFile.c_str(), ios::in);
         }
         // can we open the file?
-        if ( !(_bedStream->good()) ) {
-            cerr << "Error: The requested bed file (" << bedFile << ") could not be opened. Exiting!" << endl;
+        if ( _bedStream->fail() ) {
+            cerr << "Error: The requested BEDPE file (" 
+                 << bedFile
+                 << ") " 
+                 << "could not be opened. "
+                 << "Error message: ("
+                 << strerror(errno)
+                 << "). Exiting!" << endl;
             exit (1);
         }
     }

diff --git a/src/utils/bedGraphFile/bedGraphFile.cpp b/src/utils/bedGraphFile/bedGraphFile.cpp
@@ -38,8 +38,14 @@ void BedGraphFile::Open() {
             _bedGraphStream = new igzstream(bedGraphFile.c_str(), ios::in);
         }
         // can we open the file?
-        if ( !(_bedGraphStream->good()) ) {
-            cerr << "Error: The requested bed file (" << bedGraphFile << ") could not be opened. Exiting!" << endl;
+        if ( _bedGraphStream->fail() ) {
+            cerr << "Error: The requested file (" 
+                 << bedGraphFile
+                 << ") " 
+                 << "could not be opened. "
+                 << "Error message: ("
+                 << strerror(errno)
+                 << "). Exiting!" << endl;
             exit (1);
         }
     }

diff --git a/src/utils/fileType/fileType.cpp b/src/utils/fileType/fileType.cpp
@@ -38,9 +38,7 @@ bool isRegularFile(const string& filename) {
 returns TRUE if the file has a GZIP header.
 Should only be run on regular files.
 */
-bool isGzipFile(istream *file) {
-    //see http://www.gzip.org/zlib/rfc-gzip.html#file-format
-
+bool isGzipFile(istream *file) {    
     /*
        11-Sep-2011: 
        We now only peek at the first byte and test for GZIPiness.
@@ -54,18 +52,19 @@ bool isGzipFile(istream *file) {
 //      unsigned char cm;
     } gzip_header;
 
-    if (!file->read((char*)&gzip_header, sizeof(gzip_header))) {
+    /*
+       26-Dec-2012:
+       Just peek at the first byte instead of reading it so that we don't
+       affect the istream's failbit.  This modification was wisely proposed
+       by John Marshall in response to Issue 30:
+       https://github.com/arq5x/bedtools/issues/30
+       */
+
+    // Check to see if the first byte matches expectations for a
+    // GZIP header.
+    // Deatils: http://www.gzip.org/zlib/rfc-gzip.html#file-format
+    if (file->peek() != 0x1f)
         return false;
-    }
-
-    if ( gzip_header.id1 == 0x1f )
-//       &&
-//       gzip_header.id2 == 0x8b
-//       &&
-//       gzip_header.cm == 8 )
-    {
+    else
         return true;
-    }
-    file->putback(gzip_header.id1);
-    return false;
 }
diff --git a/src/utils/tabFile/tabFile.cpp b/src/utils/tabFile/tabFile.cpp
@@ -36,8 +36,14 @@ void TabFile::Open(void) {
             delete _tabStream;
             _tabStream = new igzstream(_tabFile.c_str(), ios::in);
         }
-        if ( !(_tabStream->good()) ) {
-            cerr << "Error: The requested file (" << _tabFile << ") could not be opened. Exiting!" << endl;
+        if ( _tabStream->fail() ) {
+            cerr << "Error: The requested file (" 
+                 << _tabFile
+                 << ") " 
+                 << "could not be opened. "
+                 << "Error message: ("
+                 << strerror(errno)
+                 << "). Exiting!" << endl;
             exit (1);
         }
     }

diff --git a/test/general/empty.bed b/test/general/empty.bed
diff --git a/test/general/non-empty.bed.gz b/test/general/non-empty.bed.gz
diff --git a/test/general/test-general.sh b/test/general/test-general.sh
@@ -65,3 +65,34 @@ echo "It looks as though you have less than 3 columns at line: 1.  Are you sure
 check obs exp
 rm obs exp
 
+
+###########################################################
+#  Fail on non-existent files.
+###########################################################
+echo "    general.t06...\c"
+$BT merge -i idontexist.bed 2> obs
+echo "Error: The requested file (idontexist.bed) could not be opened. Error message: (No such file or directory). Exiting!" > exp
+check obs exp
+rm obs exp
+
+
+###########################################################
+#  Don't fail on existent, yet empty files.
+###########################################################
+echo "    general.t07...\c"
+$BT merge -i empty.bed > obs
+touch exp
+check obs exp
+rm obs exp
+
+
+###########################################################
+#  Process gzipped files.
+###########################################################
+echo "    general.t08...\c"
+$BT merge -i non-empty.bed.gz > obs
+echo "chr1	10	21" > exp
+check obs exp
+rm obs exp
+
+