Support reading subrip subtitles from a string; remove UTF-16 test.

diff --git a/src/subrip_reader.cc b/src/subrip_reader.cc
index a8b41aec9f9e0d0157af191985cdac1f0e88346d..97d28c546658644c77b54c3444b3d2f908de03f2 100644 (file)
--- a/src/subrip_reader.cc
+++ b/src/subrip_reader.cc
@@ -22,6 +22,7 @@
 #include <boost/algorithm/string.hpp>
 #include <boost/lexical_cast.hpp>
 #include <boost/regex.hpp>
+#include <boost/bind.hpp>
 #include <cstdio>
 #include <vector>
 #include <iostream>
@@ -31,11 +32,52 @@ using std::vector;
 using std::list;
 using std::cout;
 using std::hex;
+using std::stringstream;
 using boost::lexical_cast;
 using boost::to_upper;
+using boost::optional;
+using boost::function;
 using namespace sub;
 
+/** @param s Subtitle string encoded in UTF-8 */
+SubripReader::SubripReader (string const & s)
+{
+       stringstream str (s);
+       this->read (boost::bind (&SubripReader::get_line_stringstream, this, &str));
+}
+
+/** @param f Subtitle file encoded in UTF-8 */
 SubripReader::SubripReader (FILE* f)
+{
+       this->read (boost::bind (&SubripReader::get_line_file, this, f));
+}
+
+optional<string>
+SubripReader::get_line_stringstream (stringstream* str) const
+{
+       string s;
+       getline (*str, s);
+       if (!str->good ()) {
+               return optional<string> ();
+       }
+
+       return s;
+}
+
+optional<string>
+SubripReader::get_line_file (FILE* f) const
+{
+       char buffer[256];
+       char* r = fgets (buffer, sizeof (buffer), f);
+       if (r == 0 || feof (f)) {
+               return optional<string> ();
+       }
+
+       return string (buffer);
+}
+
+void
+SubripReader::read (function<optional<string> ()> get_line)
 {
        enum {
                COUNTER,
@@ -43,38 +85,35 @@ SubripReader::SubripReader (FILE* f)
                CONTENT
        } state = COUNTER;
 
-       char buffer[256];
-
        Time from;
        Time to;
 
        string line;
        int line_number = 0;
 
-       while (!feof (f)) {
-               char* r = fgets (buffer, sizeof (buffer), f);
-               if (r == 0 || feof (f)) {
+       while (true) {
+               optional<string> line = get_line ();
+               if (!line) {
                        break;
                }
 
-               line = string (buffer);
-               trim_right_if (line, boost::is_any_of ("\n\r"));
+               trim_right_if (*line, boost::is_any_of ("\n\r"));
 
                if (
-                       line.length() >= 3 &&
-                       static_cast<unsigned char> (line[0]) == 0xef &&
-                       static_cast<unsigned char> (line[1]) == 0xbb &&
-                       static_cast<unsigned char> (line[2]) == 0xbf
+                       line->length() >= 3 &&
+                       static_cast<unsigned char> (line.get()[0]) == 0xef &&
+                       static_cast<unsigned char> (line.get()[1]) == 0xbb &&
+                       static_cast<unsigned char> (line.get()[2]) == 0xbf
                        ) {
 
                        /* Skip Unicode byte order mark */
-                       line = line.substr (3);
+                       line = line->substr (3);
                }
 
                switch (state) {
                case COUNTER:
                {
-                       if (line.empty ()) {
+                       if (line->empty ()) {
                                /* a blank line at the start is ok */
                                break;
                        }
@@ -85,9 +124,9 @@ SubripReader::SubripReader (FILE* f)
                case METADATA:
                {
                        vector<string> p;
-                       boost::algorithm::split (p, line, boost::algorithm::is_any_of (" "));
+                       boost::algorithm::split (p, *line, boost::algorithm::is_any_of (" "));
                        if (p.size() != 3 && p.size() != 7) {
-                               throw SubripError (line, "a time/position line");
+                               throw SubripError (*line, "a time/position line");
                        }
 
                        from = convert_time (p[0]);
@@ -99,11 +138,11 @@ SubripReader::SubripReader (FILE* f)
                        break;
                }
                case CONTENT:
-                       if (line.empty ()) {
+                       if (line->empty ()) {
                                state = COUNTER;
                                line_number = 0;
                        } else {
-                               convert_line (line, line_number, from, to);
+                               convert_line (*line, line_number, from, to);
                                line_number++;
                        }
                        break;

diff --git a/src/subrip_reader.h b/src/subrip_reader.h
index a010f4a9dc2a5caf695c1134479c3d723afa7721..f515356e4511c9a0aed0020c35a6644eb3f757d6 100644 (file)
--- a/src/subrip_reader.h
+++ b/src/subrip_reader.h
@@ -21,6 +21,7 @@
 #define LIBSUB_SUBRIP_READER_H
 
 #include "reader.h"
+#include <boost/function.hpp>
 
 struct subrip_reader_convert_line_test;
 struct subrip_reader_convert_time_test;
@@ -31,16 +32,20 @@ class SubripReader : public Reader
 {
 public:
        SubripReader (FILE* f);
+       SubripReader (std::string const & subs);
 
 private:
        /* For tests */
        friend struct ::subrip_reader_convert_line_test;
        friend struct ::subrip_reader_convert_time_test;
        SubripReader () {}
-       
+
        static Time convert_time (std::string t);
        void convert_line (std::string t, int line_number, Time from, Time to);
        void maybe_content (RawSubtitle& p);
+       boost::optional<std::string> get_line_stringstream (std::stringstream* str) const;
+       boost::optional<std::string> get_line_file (FILE* file) const;
+       void read (boost::function<boost::optional<std::string> ()> get_line);
 };
 
 }

diff --git a/test/subrip_reader_test.cc b/test/subrip_reader_test.cc
index 79bda51c28f42aaf33d1973c81639b65438fe3ad..0e7f50db2d72781c89b68d9b2c52506f9e42ee43 100644 (file)
--- a/test/subrip_reader_test.cc
+++ b/test/subrip_reader_test.cc
@@ -348,5 +348,4 @@ BOOST_AUTO_TEST_CASE (subrip_read_test)
        test ("sintel_en.srt");
        test ("sintel_fr.srt");
        test ("Fight.Club.1999.720p.BRRip.x264-x0r.srt");
-       test ("Chudo.Berna.2003.DUAL.BDRip.XviD.AC3.-HELLYWOOD.ENG.srt");
 }