On Wed, 3 Sep 2014, Mika Raento wrote:
This is a non-standard file that maps the MSS segment names to offsets
in the ISMV file. This can be used to build a custom MSS streaming
server without splitting the ISMV into separate files.
---
tools/ismindex.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++---------
1 file changed, 68 insertions(+), 13 deletions(-)
diff --git a/tools/ismindex.c b/tools/ismindex.c
index a6a9763..47846bd 100644
--- a/tools/ismindex.c
+++ b/tools/ismindex.c
@@ -23,7 +23,9 @@
* avconv <normal input/transcoding options> -movflags frag_keyframe foo.ismv
* ismindex -n foo foo.ismv
* This step creates foo.ism and foo.ismc that is required by IIS for
- * serving it.
+ * serving it. It also creates foo.ismf, which maps fragment names to
+ * start-end offsets in the ismv, for use in your own streaming server
+ * (the .ismf file is not created if -split is given).
*
I'd rather have a separate option for doing this, instead of always
producing a non-standard file that most people won't use. Like -ismf or
so?
* By adding -path-prefix path/, the produced foo.ism will refer to the
* files foo.ismv as "path/foo.ismv" - the prefix for the generated ismc
@@ -118,6 +120,21 @@ static int copy_tag(AVIOContext *in, AVIOContext *out,
int32_t tag_name)
return 0;
}
+static int skip_tag(AVIOContext *in, int32_t tag_name)
+{
+ int64_t pos = avio_tell(in);
This has got some unnecessary spaces before the = - you probably copied it
from somewhere where it was vertically aligned with something above/below.
+ int32_t size, tag;
+
+ size = avio_rb32(in);
+ tag = avio_rb32(in);
+ if (tag != tag_name) {
+ fprintf(stderr, "wanted tag %d, got %d", tag_name, tag);
Missing \n, and this could perhaps also write the literal tags similarly
to copy_tag (perhaps that part could be split to a separate function, like
expect_tag, which does the check and prints an error)
+ return -1;
+ }
+ avio_seek(in, pos + size, SEEK_SET);
+ return 0;
+}
+
static int write_fragment(const char *filename, AVIOContext *in)
{
AVIOContext *out = NULL;
@@ -139,25 +156,62 @@ static int write_fragment(const char *filename,
AVIOContext *in)
return ret;
}
+static int skip_fragment(AVIOContext *in)
+{
+ return (skip_tag(in, MKBETAG('m', 'o', 'o', 'f')) ||
+ skip_tag(in, MKBETAG('m', 'd', 'a', 't')));
+}
As you noted yourself, this could be written slightly differently
+
static int write_fragments(struct Tracks *tracks, int start_index,
- AVIOContext *in, const char *output_prefix)
+ AVIOContext *in, const char *basename,
+ int split, const char* output_prefix)
{
- char dirname[2048], filename[2048];
- int i, j;
+ char dirname[2048], filename[2048], idxname[2048];
+ int i, j, ret, fragment_ret;
+ FILE* out;
+ out = NULL;
+ ret = 0;
+ if (!split) {
+ snprintf(idxname, sizeof(idxname), "%s%s.ismf", output_prefix,
basename);
+ out = fopen(idxname, "w");
+ if (!out) {
+ ret = AVERROR(errno);
+ perror(idxname);
+ goto fail;
+ }
+ }
for (i = start_index; i < tracks->nb_tracks; i++) {
struct Track *track = tracks->tracks[i];
const char *type = track->is_video ? "video" : "audio";
snprintf(dirname, sizeof(dirname), "%sQualityLevels(%d)", output_prefix,
track->bitrate);
- mkdir(dirname, 0777);
+ if (split) {
+ if (mkdir(dirname, 0777) == -1) {
+ ret = AVERROR(errno);
+ perror(dirname);
+ goto fail;
+ }
+ }
for (j = 0; j < track->chunks; j++) {
snprintf(filename, sizeof(filename), "%s/Fragments(%s=%"PRId64")",
dirname, type, track->offsets[j].time);
avio_seek(in, track->offsets[j].offset, SEEK_SET);
- write_fragment(filename, in);
+ if (split) {
+ fragment_ret = write_fragment(filename, in);
+ } else {
+ fprintf(out, "%s %"PRId64, filename, avio_tell(in));
+ fragment_ret = skip_fragment(in);
+ fprintf(out, " %"PRId64"\n", avio_tell(in));
+ }
+ if (fragment_ret != 0) {
+ fprintf(stderr, "failed fragment");
Missing \n
+ ret = fragment_ret;
+ }
}
}
- return 0;
+fail:
+ if (out) fclose(out);
This is mostly a nitpick, but within the libav codebase, we mostly write
conditions like this with the fclose() on a separate line (unless it
significantly impairs readability, for some table-like things).
The rest of it looks quite good though!
// Martin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
