Diego Elio Pettenò authored 12 years ago and Timothy B. Terriberry committed 12 years ago

Includes
- A make debug target that disables optimizations and enables
   assertions,
- Proper ./configure switches for the optional features,
- A configuration summary,
- libtool versioning information,
- Visibility and warning flags,
- API documentation, and
- Support for out-of-tree builds.

Signed-off-by: Diego Elio Pettenò <flameeyes@flameeyes.eu>

* s/op_read_stereo_float/op_read_float_stereo/ for the fixed-point
   API.
* Fix compiler warnings exposed when optimizations are enabled.
* Fix opusfile_example to work with --enable-fixed-point
   --disable-float
* Fix seeking_example to not re-define OP_FIXED_POINT if it's
   already been defined.

GNU make provides a default definition for this variable.

I've no idea what's special about mingw32-ar vs native ar,
but it seems to be required.

The most important for the https backend:
- Implement RFC 6066 Server Name Indication.
- Implement SSL session resumption.
  This gives less of a speed improvement than you might think.

It's served its purpose, and isn't worth the time it takes now.
Leaving the code around to make it easy to re-enable if necessary.

Splits the documentation into sections and provides an overview.
Also explicitly document which functions can and cannot be called on
 partially-open streams, and fix-up a few things to behave as
 documented.

There are now a couple of places where we need to know the actual
 stream position indicator, not just the offset of the data we've
 read out of the ogg_sync_state.
Make a function to compute this, document it, and use it.

end_offset should be the _start_ of the last Opus page in a link,
 to guarantee we seek before it to have enough information to do
 end-trimming.
After the changes in 007ec4e4, it would be set to the start of the
 next link if we cached the last page granule position.

Also add more comments.

1) Remember the granule position of the last page we've seen from
    the current link and save the first page of the next link as
    long as we're scanning forward.
   This knocks almost 10% off the number of seeks for large links.
   For smaller links the improvement is much larger.
2) Only use pairs of close-by serial numbers to estimate link
    start locations (assuming they're above our start threshold).
   This gives a minor (<2%) improvement, which might be in the
    noise, but as it doesn't appear to hurt and is faster, might as
    well.
3) Eliminate a redundant check in op_pcm_seek_page_impl().

If we encounter the first page after the target while scanning
 forward, we no longer seek back and read it again after we exit the
 bisection search, but just hold on to the copy we already have.

Teach me to change something right before committing.

This gives us a better idea when to re-use a connection.

This uses two sightings of a valid granule position from the same
 future stream to extrapolate where the start of that stream is
 during link enumeration.
This can cut out more than 20% of the seeks required to open large
 files with lots of long links.

2df5f786 did not quite save enough.
The ogg_stream_state and initial ogg_packet list also need to be
 saved, and the starting granule position and pre-skip need to be
 restored.
Otherwise, if the stream has more than one link, the first packets
 decoded will come from the wrong one.

This means that in the normal case, after identifying the chain
 boundaries, we will seek back exactly where we were before.
With the http backend, that gives us a good chance of re-using our
 initial connection to continue reading the stream.
Currently the LRU connection expiration algorithm means that is
 unlikely with lots of links, however.

Try to guess that the next link will be approximately the average
 size of all previous links, for files with many links.
This cuts off 6-17% of the seeks.

Also remove a variable that was left unused after 5e36109d.

Reduce the number of places we modify 'offset' so that
 op_seek_helper() can always skip seeks to the current offset.
The checks we were doing before already covered all the places
 where this was useful in the normal case, but this lets us
 centralize that logic.

This commit also includes a few minor follow-ups to 9b57b0c2:
* Use a smaller type for ret_size and initialize it.
* Verify 'end' is at least as large as data we've already read.

1) We were treating EOF in op_get_next_page() as a read error when
    called from op_get_prev_page_serial().
2) We also assumed op_get_prev_page_serial() stopped scanning at the
    end of the page it returned, in order to compute the size of
    that page.
   Return the page size explicitly instead.
3) Finally, once we discover where the last page is, there is no
    reason to ever look at data past it.
   Update 'end' once we find it, and always pass that to
    op_get_next_page().

Actually close the old socket.
Don't initialize SSL more than once.

Also fixes some indenting in opusfile_example.

This is pretty preliminary for now.
Seeking performance isn't great, and there's no caching.
I'm sure there's also lots of bugs.

When a match fails, the index of the position in the link needs to
 be 64 bits.

op_find_initial_pcm_offset() would swallow any holes at the start
 of a link, but seekable streams did not use it when crossing a
 link boundary, so they would report a hole.
This unifies the hole reporting between the two, and also
 simplifies the seeking code (which also needs to ignore holes) a
 bit.

1) Check for allocation failure and fall back to merely scanning a
 file for consistent timestamps instead of loading it to RAM.
2) Report holes, but do not quit scanning/loading unless they cause
 a discontinuity in the timestamps.

The simple git-version script asks git for the current
revision, and falls back to 'unknown' if the result is
empty. This will have little effect until we have tags
in the repository.

We use the FILE_VERSION_FILTER option on doxygen to
embed this in the documentation with each file reference.

Unfortunately, there doesn't seem to be a way to
set PROJECT_NUMBER globally in the same way. This
commit unsets that option to avoid confusing references
in the generated documentation.

Just runs doxygen for us, an optionally latex.
The 'clean' target removes the doxygen output
directories.

This is just to help us validate the api docs as they're
added. By using a minimal Doxyfile we save space, make
it easier to read, and avoid warnings from deprecated
and unknown options across different doxygen versions.

Apparently just declaring the functions in the header is not good
 enough.
Currently completely untested.

3f18c55f moved it out of opus/, so strip the prefix.

If a source file moves (as in the previous commit), any stale
dependencies will block execution of 'make' because they
define a dependency relationship to an non-existant file.

The obvious thing to do is 'make clean' to get rid of the
dependency files, but of course that fails the same way.
To support this recovery, skip the include directive if
the GNU make variable MAKECMDGOALS is 'clean', exempting
this target from dependency checks.

Based on an example in the GNU make manual.

This matches the convention used by opus itself, where
headers are included without a prefix, and the build
must do -I/path/to/opus to find the appropriate versions.

More consistent formatting, s/file/stream/g, and finish adding
 function warning attributes.

The actual problem was that opus-uninstalled.pc was forcing static
 linking, but not including all the relevant libraries.