/*
- Copyright (C) 2014-2018 Carl Hetherington <cth@carlh.net>
+ Copyright (C) 2014-2020 Carl Hetherington <cth@carlh.net>
This file is part of DCP-o-matic.
*/
-#include "string_text_file.h"
+
#include "cross.h"
#include "exceptions.h"
+#include "string_text_file.h"
#include "string_text_file_content.h"
-#include <sub/subrip_reader.h>
-#include <sub/ssa_reader.h>
+#include <dcp/file.h>
#include <sub/collect.h>
+#include <sub/exceptions.h>
+#include <sub/ssa_reader.h>
+#include <sub/stl_binary_reader.h>
+#include <sub/subrip_reader.h>
+#include <sub/web_vtt_reader.h>
#include <unicode/ucsdet.h>
#include <unicode/ucnv.h>
#include <iostream>
#include "i18n.h"
-using std::vector;
+
using std::cout;
+using std::shared_ptr;
using std::string;
-using boost::shared_ptr;
+using std::vector;
using boost::scoped_array;
using boost::optional;
-using dcp::Data;
+using dcp::ArrayData;
+using namespace dcpomatic;
+
StringTextFile::StringTextFile (shared_ptr<const StringTextFileContent> content)
{
- 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);
-
- /* Fix OS X line endings */
- size_t utf8_len = strlen (utf8.get ());
- for (size_t i = 0; i < utf8_len; ++i) {
- if (utf8[i] == '\r' && ((i == utf8_len - 1) || utf8[i + 1] != '\n')) {
- utf8[i] = '\n';
- }
- }
+ string ext = content->path(0).extension().string();
+ transform (ext.begin(), ext.end(), ext.begin(), ::tolower);
- ucsdet_close (detector);
- ucnv_close (to_utf16);
- ucnv_close (to_utf8);
+ std::unique_ptr<sub::Reader> reader;
- sub::Reader* reader = 0;
+ if (ext == ".stl") {
+ dcp::File f(content->path(0), "rb");
+ if (!f) {
+ throw OpenFileError (f.path(), errno, OpenFileError::READ);
+ }
+ try {
+ reader.reset(new sub::STLBinaryReader(f.get()));
+ } catch (...) {
+ throw;
+ }
- string ext = content->path(0).extension().string();
- transform (ext.begin(), ext.end(), ext.begin(), ::tolower);
+ } else {
+ /* Text-based file; sort out its character encoding before we try to parse it */
+
+ ArrayData in (content->path (0));
+
+ UErrorCode status = U_ZERO_ERROR;
+ UCharsetDetector* detector = ucsdet_open (&status);
+ ucsdet_setText (detector, reinterpret_cast<const char *>(in.data()), 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()), 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);
+
+ /* Fix OS X line endings */
+ size_t utf8_len = strlen (utf8.get ());
+ for (size_t i = 0; i < utf8_len; ++i) {
+ if (utf8[i] == '\r' && ((i == utf8_len - 1) || utf8[i + 1] != '\n')) {
+ utf8[i] = '\n';
+ }
+ }
- if (ext == ".srt") {
- reader = new sub::SubripReader (utf8.get());
- } else if (ext == ".ssa" || ext == ".ass") {
- reader = new sub::SSAReader (utf8.get());
+ ucsdet_close (detector);
+ ucnv_close (to_utf16);
+ ucnv_close (to_utf8);
+
+ if (ext == ".srt") {
+ try {
+ reader.reset(new sub::SubripReader(utf8.get()));
+ } catch (sub::SubripError& subrip_error) {
+ /* Sometimes files are have the .srt extension but are really WEBVTT... */
+ try {
+ reader.reset(new sub::WebVTTReader(utf8.get()));
+ } catch (sub::WebVTTHeaderError&) {
+ /* ...but in this case there isn't even a WebVTT error */
+ throw subrip_error;
+ }
+ }
+ } else if (ext == ".ssa" || ext == ".ass") {
+ reader.reset(new sub::SSAReader(utf8.get()));
+ } else if (ext == ".vtt") {
+ reader.reset(new sub::WebVTTReader(utf8.get()));
+ }
}
if (reader) {
- _subtitles = sub::collect<vector<sub::Subtitle> > (reader->subtitles ());
+ _subtitles = sub::collect<vector<sub::Subtitle>>(reader->subtitles());
}
-
- delete reader;
}
/** @return time of first subtitle, if there is one */
StringTextFile::first () const
{
if (_subtitles.empty()) {
- return optional<ContentTime>();
+ return {};
}
return ContentTime::from_seconds(_subtitles[0].from.all_as_seconds());
StringTextFile::length () const
{
if (_subtitles.empty ()) {
- return ContentTime ();
+ return {};
}
return ContentTime::from_seconds (_subtitles.back().to.all_as_seconds ());
projects / dcpomatic.git / blobdiff