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_check_get_range (GstBaseSrc * basesrc);
78 static gboolean gst_split_file_src_get_size (GstBaseSrc * basesrc, guint64 * s);
79 static gboolean gst_split_file_src_unlock (GstBaseSrc * basesrc);
80 static GstFlowReturn gst_split_file_src_create (GstBaseSrc * basesrc,
81 guint64 offset, guint size, GstBuffer ** buffer);
83 static GstStaticPadTemplate gst_split_file_src_pad_template =
84 GST_STATIC_PAD_TEMPLATE ("src",
89 GST_DEBUG_CATEGORY_STATIC (splitfilesrc_debug);
90 #define GST_CAT_DEFAULT splitfilesrc_debug
92 GST_BOILERPLATE (GstSplitFileSrc, gst_split_file_src, GstBaseSrc,
96 gst_split_file_src_base_init (gpointer g_class)
98 GstElementClass *gstelement_class = GST_ELEMENT_CLASS (g_class);
100 GST_DEBUG_CATEGORY_INIT (splitfilesrc_debug, "splitfilesrc", 0,
101 "splitfilesrc element");
103 gst_element_class_add_pad_template (gstelement_class,
104 gst_static_pad_template_get (&gst_split_file_src_pad_template));
106 gst_element_class_set_details_simple (gstelement_class, "Split-File Source",
108 "Read a sequentially named set of files as if it was one large file",
109 "Tim-Philipp Müller <tim.muller@collabora.co.uk>");
113 #define WIN32_BLURB " Location string must be in UTF-8 encoding (on Windows)."
115 #define WIN32_BLURB /* nothing */
119 gst_split_file_src_class_init (GstSplitFileSrcClass * klass)
121 GstBaseSrcClass *gstbasesrc_class = GST_BASE_SRC_CLASS (klass);
122 GObjectClass *gobject_class = G_OBJECT_CLASS (klass);
124 gobject_class->set_property = gst_split_file_src_set_property;
125 gobject_class->get_property = gst_split_file_src_get_property;
126 gobject_class->finalize = gst_split_file_src_finalize;
128 g_object_class_install_property (gobject_class, PROP_LOCATION,
129 g_param_spec_string ("location", "File Location",
130 "Wildcard pattern to match file names of the input files. If "
131 "the location is an absolute path or contains directory components, "
132 "only the base file name part will be considered for pattern "
133 "matching. The results will be sorted." WIN32_BLURB,
134 DEFAULT_LOCATION, G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
136 gstbasesrc_class->start = GST_DEBUG_FUNCPTR (gst_split_file_src_start);
137 gstbasesrc_class->stop = GST_DEBUG_FUNCPTR (gst_split_file_src_stop);
138 gstbasesrc_class->create = GST_DEBUG_FUNCPTR (gst_split_file_src_create);
139 gstbasesrc_class->get_size = GST_DEBUG_FUNCPTR (gst_split_file_src_get_size);
140 gstbasesrc_class->unlock = GST_DEBUG_FUNCPTR (gst_split_file_src_unlock);
141 gstbasesrc_class->is_seekable =
142 GST_DEBUG_FUNCPTR (gst_split_file_src_can_seek);
143 gstbasesrc_class->check_get_range =
144 GST_DEBUG_FUNCPTR (gst_split_file_src_check_get_range);
148 gst_split_file_src_init (GstSplitFileSrc * splitfilesrc,
149 GstSplitFileSrcClass * g_class)
154 gst_split_file_src_finalize (GObject * obj)
156 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (obj);
158 g_free (src->location);
159 src->location = NULL;
161 G_OBJECT_CLASS (parent_class)->finalize (obj);
165 gst_split_file_src_can_seek (GstBaseSrc * basesrc)
171 gst_split_file_src_check_get_range (GstBaseSrc * basesrc)
177 gst_split_file_src_unlock (GstBaseSrc * basesrc)
179 /* This is not actually that useful, since all normal file
180 * operations are fully blocking anyway */
182 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
184 GST_DEBUG_OBJECT (src, "cancelling pending I/O operation if there is one");
185 /* g_cancellable_cancel (src->cancellable); */
186 GST_DEBUG_OBJECT (src, "done");
193 gst_split_file_src_get_size (GstBaseSrc * basesrc, guint64 * size)
195 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
197 *size = src->parts[src->num_parts - 1].stop + 1;
202 gst_split_file_src_set_property (GObject * object, guint prop_id,
203 const GValue * value, GParamSpec * pspec)
205 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (object);
209 GST_OBJECT_LOCK (src);
210 g_free (src->location);
211 src->location = g_value_dup_string (value);
213 if (!g_utf8_validate (src->location, -1, NULL)) {
214 g_warning ("splitfilesrc 'location' property must be in UTF-8 "
215 "encoding on Windows");
218 GST_OBJECT_UNLOCK (src);
221 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
227 gst_split_file_src_get_property (GObject * object, guint prop_id,
228 GValue * value, GParamSpec * pspec)
230 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (object);
234 GST_OBJECT_LOCK (src);
235 g_value_set_string (value, src->location);
236 GST_OBJECT_UNLOCK (src);
239 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
245 gst_split_file_src_array_sortfunc (gchar ** a, gchar ** b)
247 return strcmp (*a, *b);
251 gst_split_file_src_find_files (GstSplitFileSrc * src, const gchar * dirname,
252 const gchar * basename, GError ** err)
259 if (dirname == NULL || basename == NULL)
260 goto invalid_location;
262 GST_INFO_OBJECT (src, "checking in directory '%s' for pattern '%s'",
265 dir = g_dir_open (dirname, 0, err);
269 if (DEFAULT_PATTERN_MATCH_MODE == MATCH_MODE_UTF8 &&
270 !g_utf8_validate (basename, -1, NULL)) {
274 /* mode will be AUTO on linux/unix and UTF8 on win32 */
275 pspec = pattern_spec_new (basename, DEFAULT_PATTERN_MATCH_MODE);
277 files = g_ptr_array_new ();
279 while ((name = g_dir_read_name (dir))) {
280 GST_TRACE_OBJECT (src, "check: %s", name);
281 if (pattern_match_string (pspec, name)) {
282 GST_DEBUG_OBJECT (src, "match: %s", name);
283 g_ptr_array_add (files, g_build_filename (dirname, name, NULL));
290 g_ptr_array_sort (files, (GCompareFunc) gst_split_file_src_array_sortfunc);
291 g_ptr_array_add (files, NULL);
293 pattern_spec_free (pspec);
296 return (gchar **) g_ptr_array_free (files, FALSE);
301 g_set_error_literal (err, G_FILE_ERROR, G_FILE_ERROR_INVAL,
302 "No filename specified.");
308 g_set_error_literal (err, G_FILE_ERROR, G_FILE_ERROR_INVAL,
309 "Filename pattern must be UTF-8 on Windows.");
314 pattern_spec_free (pspec);
316 g_set_error_literal (err, G_FILE_ERROR, G_FILE_ERROR_NOENT,
317 "Found no files matching the pattern.");
323 gst_split_file_src_start (GstBaseSrc * basesrc)
325 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
326 GCancellable *cancel;
327 gboolean ret = FALSE;
330 gchar *basename = NULL;
331 gchar *dirname = NULL;
335 GST_OBJECT_LOCK (src);
336 if (src->location != NULL && src->location[0] != '\0') {
337 basename = g_path_get_basename (src->location);
338 dirname = g_path_get_dirname (src->location);
340 GST_OBJECT_UNLOCK (src);
342 files = gst_split_file_src_find_files (src, dirname, basename, &err);
344 if (files == NULL || *files == NULL)
347 src->num_parts = g_strv_length (files);
348 src->parts = g_new0 (GstFilePart, src->num_parts);
350 cancel = src->cancellable;
353 for (i = 0; i < src->num_parts; ++i) {
354 GFileInputStream *stream;
359 file = g_file_new_for_path (files[i]);
360 stream = g_file_read (file, cancel, &err);
361 g_object_unref (file);
364 goto open_read_error;
366 info = g_file_input_stream_query_info (stream, "standard::*", NULL, &err);
368 g_object_unref (stream);
369 goto query_info_error;
372 size = g_file_info_get_size (info);
373 g_object_unref (info);
375 src->parts[i].stream = stream;
376 src->parts[i].path = g_strdup (files[i]);
377 src->parts[i].start = offset;
378 src->parts[i].stop = offset + size - 1;
380 GST_DEBUG ("[%010" G_GUINT64_FORMAT "-%010" G_GUINT64_FORMAT "] %s",
381 src->parts[i].start, src->parts[i].stop, src->parts[i].path);
386 GST_INFO ("Successfully opened %u file parts for reading", src->num_parts);
390 src->cancellable = g_cancellable_new ();
405 if (err->code == G_IO_ERROR_CANCELLED)
408 GST_ELEMENT_ERROR (src, RESOURCE, OPEN_READ, ("%s", err->message),
409 ("Failed to find files in '%s' for pattern '%s'",
410 GST_STR_NULL (dirname), GST_STR_NULL (basename)));
415 if (err->code == G_IO_ERROR_CANCELLED)
418 GST_ELEMENT_ERROR (src, RESOURCE, OPEN_READ, ("%s", err->message),
419 ("Failed to open file '%s' for reading", files[i]));
424 if (err->code == G_IO_ERROR_CANCELLED)
427 GST_ELEMENT_ERROR (src, RESOURCE, OPEN_READ, ("%s", err->message),
428 ("Failed to query info for file '%s'", files[i]));
433 GST_DEBUG_OBJECT (src, "I/O operation cancelled from another thread");
439 gst_split_file_src_stop (GstBaseSrc * basesrc)
441 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
444 for (i = 0; i < src->num_parts; ++i) {
445 if (src->parts[i].stream != NULL)
446 g_object_unref (src->parts[i].stream);
447 g_free (src->parts[i].path);
453 g_object_unref (src->cancellable);
454 src->cancellable = NULL;
460 gst_split_file_src_find_part_for_offset (GstSplitFileSrc * src, guint64 offset,
466 /* TODO: could use gst_util_array_binary_search() here */
468 for (i = 0; i < src->num_parts; ++i) {
469 if (offset >= part->start && offset <= part->stop) {
480 gst_split_file_src_create (GstBaseSrc * basesrc, guint64 offset, guint size,
483 GstSplitFileSrc *src = GST_SPLIT_FILE_SRC (basesrc);
484 GstFilePart cur_part;
485 GInputStream *stream;
486 GCancellable *cancel;
494 cur_part = src->parts[src->cur_part];
495 if (offset < cur_part.start || offset > cur_part.stop) {
496 if (!gst_split_file_src_find_part_for_offset (src, offset, &src->cur_part))
497 return GST_FLOW_UNEXPECTED;
498 cur_part = src->parts[src->cur_part];
501 GST_LOG_OBJECT (src, "current part: %u (%" G_GUINT64_FORMAT " - "
502 "%" G_GUINT64_FORMAT ", %s)", src->cur_part, cur_part.start,
503 cur_part.stop, cur_part.path);
505 buf = gst_buffer_new_and_alloc (size);
507 GST_BUFFER_OFFSET (buf) = offset;
509 data = GST_BUFFER_DATA (buf);
511 cancel = src->cancellable;
514 guint64 bytes_to_end_of_part;
517 /* we want the offset into the file part */
518 read_offset = offset - cur_part.start;
520 GST_LOG ("Reading part %03u from offset %" G_GUINT64_FORMAT " (%s)",
521 src->cur_part, read_offset, cur_part.path);
523 /* FIXME: only seek when needed (hopefully gio is smart) */
524 seekable = G_SEEKABLE (cur_part.stream);
525 if (!g_seekable_seek (seekable, read_offset, G_SEEK_SET, cancel, &err))
528 GST_LOG_OBJECT (src, "now: %" G_GUINT64_FORMAT, g_seekable_tell (seekable));
530 bytes_to_end_of_part = (cur_part.stop - cur_part.start) + 1 - read_offset;
531 to_read = MIN (size, bytes_to_end_of_part);
533 GST_LOG_OBJECT (src, "reading %u bytes from part %u (bytes to end of "
534 "part: %u)", to_read, src->cur_part, (guint) bytes_to_end_of_part);
536 stream = G_INPUT_STREAM (cur_part.stream);
538 /* NB: we won't try to read beyond EOF */
539 if (!g_input_stream_read_all (stream, data, to_read, &read, cancel, &err))
542 GST_LOG_OBJECT (src, "read %u bytes", (guint) read);
552 GST_LOG_OBJECT (src, "%u bytes left to read for this chunk", size);
554 /* corner case, this should never really happen (assuming basesrc clips
555 * requests beyond the file size) */
556 if (read < to_read) {
557 if (src->cur_part == src->num_parts - 1) {
558 /* last file part, stop reading and truncate buffer */
559 GST_BUFFER_SIZE (buf) = offset - GST_BUFFER_OFFSET (buf);
562 goto file_part_changed;
567 cur_part = src->parts[src->cur_part];
570 GST_BUFFER_OFFSET_END (buf) = offset;
573 GST_LOG_OBJECT (src, "read %u bytes into buf %p", GST_BUFFER_SIZE (buf), buf);
579 if (err->code == G_IO_ERROR_CANCELLED)
582 GST_ELEMENT_ERROR (src, RESOURCE, SEEK, (NULL),
583 ("Seek to %" G_GUINT64_FORMAT " in %s failed", read_offset,
586 gst_buffer_unref (buf);
587 return GST_FLOW_ERROR;
591 if (err->code == G_IO_ERROR_CANCELLED)
594 GST_ELEMENT_ERROR (src, RESOURCE, READ, ("%s", err->message),
595 ("Read from %" G_GUINT64_FORMAT " in %s failed", read_offset,
598 gst_buffer_unref (buf);
599 return GST_FLOW_ERROR;
603 GST_ELEMENT_ERROR (src, RESOURCE, READ,
604 ("Read error while reading file part %s", cur_part.path),
605 ("Short read in file part, file may have been modified since start"));
606 gst_buffer_unref (buf);
607 return GST_FLOW_ERROR;
611 GST_DEBUG_OBJECT (src, "I/O operation cancelled from another thread");
613 gst_buffer_unref (buf);
614 return GST_FLOW_WRONG_STATE;