1 /* GStreamer Split File Source
2 * Copyright (C) 2011 Collabora Ltd. <tim.muller@collabora.co.uk>
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Library General Public
6 * License as published by the Free Software Foundation; either
7 * version 2 of the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Library General Public License for more details.
14 * You should have received a copy of the GNU Library General Public
15 * License along with this library; if not, write to the
16 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
17 * Boston, MA 02111-1307, USA.
20 * SECTION:element-splitfilesrc
21 * @see_also: #GstFileSrc, #GstMultiFileSrc
23 * Reads data from multiple files, presenting those files as one continuous
24 * file to downstream elements. This is useful for reading a large file that
25 * had to be split into multiple parts due to filesystem file size limitations,
28 * The files to select are chosen via the location property, which supports
29 * (and expects) shell-style wildcards (but only for the filename, not for
30 * directories). The results will be sorted.
33 * <title>Example launch line</title>
35 * gst-launch splitfilesrc location="/path/to/part-*.mpg" ! decodebin ! ... \
36 * ]| Plays the different parts as if they were one single MPEG file.
43 * - implement splitfile:// URI handler?
50 #include "gstsplitfilesrc.h"
51 #include "patternspec.h"
56 #define DEFAULT_PATTERN_MATCH_MODE MATCH_MODE_UTF8
58 #define DEFAULT_PATTERN_MATCH_MODE MATCH_MODE_AUTO
66 #define DEFAULT_LOCATION NULL
68 static void gst_split_file_src_set_property (GObject * object, guint prop_id,
69 const GValue * value, GParamSpec * pspec);
70 static void gst_split_file_src_get_property (GObject * object, guint prop_id,
71 GValue * value, GParamSpec * pspec);
72 static void gst_split_file_src_finalize (GObject * obj);
74 static gboolean gst_split_file_src_start (GstBaseSrc * basesrc);
75 static gboolean gst_split_file_src_stop (GstBaseSrc * basesrc);
76 static gboolean gst_split_file_src_can_seek (GstBaseSrc * basesrc);
77 static gboolean gst_split_file_src_get_size (GstBaseSrc * basesrc, guint64 * s);
78 static gboolean gst_split_file_src_unlock (GstBaseSrc * basesrc);
79 static GstFlowReturn gst_split_file_src_create (GstBaseSrc * basesrc,
80 guint64 offset, guint size, GstBuffer ** buffer);
82 static GstStaticPadTemplate gst_split_file_src_pad_template =
83 GST_STATIC_PAD_TEMPLATE ("src",
88 GST_DEBUG_CATEGORY_STATIC (splitfilesrc_debug);
89 #define GST_CAT_DEFAULT splitfilesrc_debug
91 G_DEFINE_TYPE (GstSplitFileSrc, gst_split_file_src, GST_TYPE_BASE_SRC);
94 #define WIN32_BLURB " Location string must be in UTF-8 encoding (on Windows)."
96 #define WIN32_BLURB /* nothing */
100 gst_split_file_src_class_init (GstSplitFileSrcClass * klass)
102 GstBaseSrcClass *gstbasesrc_class = GST_BASE_SRC_CLASS (klass);
103 GstElementClass *gstelement_class = GST_ELEMENT_CLASS (klass);
104 GObjectClass *gobject_class = G_OBJECT_CLASS (klass);
106 gobject_class->set_property = gst_split_file_src_set_property;
107 gobject_class->get_property = gst_split_file_src_get_property;
108 gobject_class->finalize = gst_split_file_src_finalize;
110 g_object_class_install_property (gobject_class, PROP_LOCATION,
111 g_param_spec_string ("location", "File Location",
112 "Wildcard pattern to match file names of the input files. If "
113 "the location is an absolute path or contains directory components, "
114 "only the base file name part will be considered for pattern "
115 "matching. The results will be sorted." WIN32_BLURB,
116 DEFAULT_LOCATION, G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
118 gstbasesrc_class->start = GST_DEBUG_FUNCPTR (gst_split_file_src_start);
119 gstbasesrc_class->stop = GST_DEBUG_FUNCPTR (gst_split_file_src_stop);
120 gstbasesrc_class->create = GST_DEBUG_FUNCPTR (gst_split_file_src_create);
121 gstbasesrc_class->get_size = GST_DEBUG_FUNCPTR (gst_split_file_src_get_size);
122 gstbasesrc_class->unlock = GST_DEBUG_FUNCPTR (gst_split_file_src_unlock);
123 gstbasesrc_class->is_seekable =
124 GST_DEBUG_FUNCPTR (gst_split_file_src_can_seek);
126 GST_DEBUG_CATEGORY_INIT (splitfilesrc_debug, "splitfilesrc", 0,
127 "splitfilesrc element");
129 gst_element_class_add_pad_template (gstelement_class,
130 gst_static_pad_template_get (&gst_split_file_src_pad_template));
132 gst_element_class_set_details_simple (gstelement_class, "Split-File Source",
134 "Read a sequentially named set of files as if it was one large file",
135 "Tim-Philipp Müller <tim.muller@collabora.co.uk>");
139 gst_split_file_src_init (GstSplitFileSrc * splitfilesrc)
144 gst_split_file_src_finalize (GObject * obj)
146 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (obj);
148 g_free (src->location);
149 src->location = NULL;
151 G_OBJECT_CLASS (gst_split_file_src_parent_class)->finalize (obj);
155 gst_split_file_src_can_seek (GstBaseSrc * basesrc)
161 gst_split_file_src_unlock (GstBaseSrc * basesrc)
163 /* This is not actually that useful, since all normal file
164 * operations are fully blocking anyway */
166 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
168 GST_DEBUG_OBJECT (src, "cancelling pending I/O operation if there is one");
169 /* g_cancellable_cancel (src->cancellable); */
170 GST_DEBUG_OBJECT (src, "done");
177 gst_split_file_src_get_size (GstBaseSrc * basesrc, guint64 * size)
179 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
181 *size = src->parts[src->num_parts - 1].stop + 1;
186 gst_split_file_src_set_property (GObject * object, guint prop_id,
187 const GValue * value, GParamSpec * pspec)
189 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (object);
193 GST_OBJECT_LOCK (src);
194 g_free (src->location);
195 src->location = g_value_dup_string (value);
197 if (!g_utf8_validate (src->location, -1, NULL)) {
198 g_warning ("splitfilesrc 'location' property must be in UTF-8 "
199 "encoding on Windows");
202 GST_OBJECT_UNLOCK (src);
205 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
211 gst_split_file_src_get_property (GObject * object, guint prop_id,
212 GValue * value, GParamSpec * pspec)
214 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (object);
218 GST_OBJECT_LOCK (src);
219 g_value_set_string (value, src->location);
220 GST_OBJECT_UNLOCK (src);
223 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
229 gst_split_file_src_array_sortfunc (gchar ** a, gchar ** b)
231 return strcmp (*a, *b);
235 gst_split_file_src_find_files (GstSplitFileSrc * src, const gchar * dirname,
236 const gchar * basename, GError ** err)
243 if (dirname == NULL || basename == NULL)
244 goto invalid_location;
246 GST_INFO_OBJECT (src, "checking in directory '%s' for pattern '%s'",
249 dir = g_dir_open (dirname, 0, err);
253 if (DEFAULT_PATTERN_MATCH_MODE == MATCH_MODE_UTF8 &&
254 !g_utf8_validate (basename, -1, NULL)) {
258 /* mode will be AUTO on linux/unix and UTF8 on win32 */
259 pspec = pattern_spec_new (basename, DEFAULT_PATTERN_MATCH_MODE);
261 files = g_ptr_array_new ();
263 while ((name = g_dir_read_name (dir))) {
264 GST_TRACE_OBJECT (src, "check: %s", name);
265 if (pattern_match_string (pspec, name)) {
266 GST_DEBUG_OBJECT (src, "match: %s", name);
267 g_ptr_array_add (files, g_build_filename (dirname, name, NULL));
274 g_ptr_array_sort (files, (GCompareFunc) gst_split_file_src_array_sortfunc);
275 g_ptr_array_add (files, NULL);
277 pattern_spec_free (pspec);
280 return (gchar **) g_ptr_array_free (files, FALSE);
285 g_set_error_literal (err, G_FILE_ERROR, G_FILE_ERROR_INVAL,
286 "No filename specified.");
292 g_set_error_literal (err, G_FILE_ERROR, G_FILE_ERROR_INVAL,
293 "Filename pattern must be UTF-8 on Windows.");
298 pattern_spec_free (pspec);
300 g_set_error_literal (err, G_FILE_ERROR, G_FILE_ERROR_NOENT,
301 "Found no files matching the pattern.");
307 gst_split_file_src_start (GstBaseSrc * basesrc)
309 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
310 GCancellable *cancel;
311 gboolean ret = FALSE;
314 gchar *basename = NULL;
315 gchar *dirname = NULL;
319 GST_OBJECT_LOCK (src);
320 if (src->location != NULL && src->location[0] != '\0') {
321 basename = g_path_get_basename (src->location);
322 dirname = g_path_get_dirname (src->location);
324 GST_OBJECT_UNLOCK (src);
326 files = gst_split_file_src_find_files (src, dirname, basename, &err);
328 if (files == NULL || *files == NULL)
331 src->num_parts = g_strv_length (files);
332 src->parts = g_new0 (GstFilePart, src->num_parts);
334 cancel = src->cancellable;
337 for (i = 0; i < src->num_parts; ++i) {
338 GFileInputStream *stream;
343 file = g_file_new_for_path (files[i]);
344 stream = g_file_read (file, cancel, &err);
345 g_object_unref (file);
348 goto open_read_error;
350 info = g_file_input_stream_query_info (stream, "standard::*", NULL, &err);
352 g_object_unref (stream);
353 goto query_info_error;
356 size = g_file_info_get_size (info);
357 g_object_unref (info);
359 src->parts[i].stream = stream;
360 src->parts[i].path = g_strdup (files[i]);
361 src->parts[i].start = offset;
362 src->parts[i].stop = offset + size - 1;
364 GST_DEBUG ("[%010" G_GUINT64_FORMAT "-%010" G_GUINT64_FORMAT "] %s",
365 src->parts[i].start, src->parts[i].stop, src->parts[i].path);
370 GST_INFO ("Successfully opened %u file parts for reading", src->num_parts);
374 src->cancellable = g_cancellable_new ();
389 if (err->code == G_IO_ERROR_CANCELLED)
392 GST_ELEMENT_ERROR (src, RESOURCE, OPEN_READ, ("%s", err->message),
393 ("Failed to find files in '%s' for pattern '%s'",
394 GST_STR_NULL (dirname), GST_STR_NULL (basename)));
399 if (err->code == G_IO_ERROR_CANCELLED)
402 GST_ELEMENT_ERROR (src, RESOURCE, OPEN_READ, ("%s", err->message),
403 ("Failed to open file '%s' for reading", files[i]));
408 if (err->code == G_IO_ERROR_CANCELLED)
411 GST_ELEMENT_ERROR (src, RESOURCE, OPEN_READ, ("%s", err->message),
412 ("Failed to query info for file '%s'", files[i]));
417 GST_DEBUG_OBJECT (src, "I/O operation cancelled from another thread");
423 gst_split_file_src_stop (GstBaseSrc * basesrc)
425 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
428 for (i = 0; i < src->num_parts; ++i) {
429 if (src->parts[i].stream != NULL)
430 g_object_unref (src->parts[i].stream);
431 g_free (src->parts[i].path);
437 g_object_unref (src->cancellable);
438 src->cancellable = NULL;
444 gst_split_file_src_find_part_for_offset (GstSplitFileSrc * src, guint64 offset,
450 /* TODO: could use gst_util_array_binary_search() here */
452 for (i = 0; i < src->num_parts; ++i) {
453 if (offset >= part->start && offset <= part->stop) {
464 gst_split_file_src_create (GstBaseSrc * basesrc, guint64 offset, guint size,
467 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
468 GstFilePart cur_part;
469 GInputStream *stream;
470 GCancellable *cancel;
479 cur_part = src->parts[src->cur_part];
480 if (offset < cur_part.start || offset > cur_part.stop) {
481 if (!gst_split_file_src_find_part_for_offset (src, offset, &src->cur_part))
483 cur_part = src->parts[src->cur_part];
486 GST_LOG_OBJECT (src, "current part: %u (%" G_GUINT64_FORMAT " - "
487 "%" G_GUINT64_FORMAT ", %s)", src->cur_part, cur_part.start,
488 cur_part.stop, cur_part.path);
490 buf = gst_buffer_new_allocate (NULL, size, 0);
492 GST_BUFFER_OFFSET (buf) = offset;
494 gst_buffer_map (buf, &map, GST_MAP_WRITE);
497 cancel = src->cancellable;
500 guint64 bytes_to_end_of_part;
503 /* we want the offset into the file part */
504 read_offset = offset - cur_part.start;
506 GST_LOG ("Reading part %03u from offset %" G_GUINT64_FORMAT " (%s)",
507 src->cur_part, read_offset, cur_part.path);
509 /* FIXME: only seek when needed (hopefully gio is smart) */
510 seekable = G_SEEKABLE (cur_part.stream);
511 if (!g_seekable_seek (seekable, read_offset, G_SEEK_SET, cancel, &err))
514 GST_LOG_OBJECT (src, "now: %" G_GUINT64_FORMAT, g_seekable_tell (seekable));
516 bytes_to_end_of_part = (cur_part.stop - cur_part.start) + 1 - read_offset;
517 to_read = MIN (size, bytes_to_end_of_part);
519 GST_LOG_OBJECT (src, "reading %u bytes from part %u (bytes to end of "
520 "part: %u)", to_read, src->cur_part, (guint) bytes_to_end_of_part);
522 stream = G_INPUT_STREAM (cur_part.stream);
524 /* NB: we won't try to read beyond EOF */
525 if (!g_input_stream_read_all (stream, data, to_read, &read, cancel, &err))
528 GST_LOG_OBJECT (src, "read %u bytes", (guint) read);
538 GST_LOG_OBJECT (src, "%u bytes left to read for this chunk", size);
540 /* corner case, this should never really happen (assuming basesrc clips
541 * requests beyond the file size) */
542 if (read < to_read) {
543 if (src->cur_part == src->num_parts - 1) {
544 /* last file part, stop reading and truncate buffer */
545 gst_buffer_set_size (buf, offset - GST_BUFFER_OFFSET (buf));
548 goto file_part_changed;
553 cur_part = src->parts[src->cur_part];
556 GST_BUFFER_OFFSET_END (buf) = offset;
558 gst_buffer_unmap (buf, &map);
561 GST_LOG_OBJECT (src, "read %" G_GSIZE_FORMAT " bytes into buf %p",
562 gst_buffer_get_size (buf), buf);
568 if (err->code == G_IO_ERROR_CANCELLED)
571 GST_ELEMENT_ERROR (src, RESOURCE, SEEK, (NULL),
572 ("Seek to %" G_GUINT64_FORMAT " in %s failed", read_offset,
575 gst_buffer_unref (buf);
576 return GST_FLOW_ERROR;
580 if (err->code == G_IO_ERROR_CANCELLED)
583 GST_ELEMENT_ERROR (src, RESOURCE, READ, ("%s", err->message),
584 ("Read from %" G_GUINT64_FORMAT " in %s failed", read_offset,
587 gst_buffer_unref (buf);
588 return GST_FLOW_ERROR;
592 GST_ELEMENT_ERROR (src, RESOURCE, READ,
593 ("Read error while reading file part %s", cur_part.path),
594 ("Short read in file part, file may have been modified since start"));
595 gst_buffer_unref (buf);
596 return GST_FLOW_ERROR;
600 GST_DEBUG_OBJECT (src, "I/O operation cancelled from another thread");
602 gst_buffer_unref (buf);
603 return GST_FLOW_WRONG_STATE;