diff options
Diffstat (limited to 'src/lib/subrip.cc')
| -rw-r--r-- | src/lib/subrip.cc | 39 |
1 files changed, 34 insertions, 5 deletions
diff --git a/src/lib/subrip.cc b/src/lib/subrip.cc index f19867952..a707d1f9f 100644 --- a/src/lib/subrip.cc +++ b/src/lib/subrip.cc @@ -21,22 +21,51 @@ #include "cross.h" #include "exceptions.h" #include "subrip_content.h" +#include "data.h" #include <sub/subrip_reader.h> #include <sub/collect.h> +#include <unicode/ucsdet.h> +#include <unicode/ucnv.h> +#include <iostream> #include "i18n.h" using std::vector; +using std::cout; +using std::string; using boost::shared_ptr; +using boost::scoped_array; SubRip::SubRip (shared_ptr<const SubRipContent> content) { - FILE* f = fopen_boost (content->path (0), "r"); - if (!f) { - throw OpenFileError (content->path (0)); - } + Data in (content->path (0)); + + UErrorCode status = U_ZERO_ERROR; + UCharsetDetector* detector = ucsdet_open (&status); + ucsdet_setText (detector, reinterpret_cast<const char *> (in.data().get()), in.size(), &status); + + UCharsetMatch const * match = ucsdet_detect (detector, &status); + char const * in_charset = ucsdet_getName (match, &status); + + UConverter* to_utf16 = ucnv_open (in_charset, &status); + /* This is a guess; I think we should be able to encode any input in 4 times its input size */ + scoped_array<uint16_t> utf16 (new uint16_t[in.size() * 2]); + int const utf16_len = ucnv_toUChars ( + to_utf16, reinterpret_cast<UChar*>(utf16.get()), in.size() * 2, + reinterpret_cast<const char *> (in.data().get()), in.size(), + &status + ); + + UConverter* to_utf8 = ucnv_open ("UTF-8", &status); + /* Another guess */ + scoped_array<char> utf8 (new char[utf16_len * 2]); + ucnv_fromUChars (to_utf8, utf8.get(), utf16_len * 2, reinterpret_cast<UChar*>(utf16.get()), utf16_len, &status); + + ucsdet_close (detector); + ucnv_close (to_utf16); + ucnv_close (to_utf8); - sub::SubripReader reader (f); + sub::SubripReader reader (utf8.get()); _subtitles = sub::collect<vector<sub::Subtitle> > (reader.subtitles ()); } |