diff options
author | Andrée Ekroth <andree.ekroth@gmail.com> | 2014-01-13 22:25:41 +0100 |
---|---|---|
committer | Andrée Ekroth <andree.ekroth@gmail.com> | 2014-01-13 22:28:29 +0100 |
commit | cd5817b67e89d4f47546809ad0299bdec1823147 (patch) | |
tree | cf5e12dc9887fc6faec4306f627308d991f48d20 /src/encoder | |
parent | 56a7fcf189b904a9d1f6e96f7efa0233c742190a (diff) |
encoder/shine: fix segfault workaround
Initializing and closing the Shine library without
writing any data results in a segmentation fault.
The current workaround writes zeroes if there was
no actual data.
Diffstat (limited to 'src/encoder')
-rw-r--r-- | src/encoder/ShineEncoderPlugin.cxx | 22 |
1 files changed, 17 insertions, 5 deletions
diff --git a/src/encoder/ShineEncoderPlugin.cxx b/src/encoder/ShineEncoderPlugin.cxx index 3097810b9..39e400a58 100644 --- a/src/encoder/ShineEncoderPlugin.cxx +++ b/src/encoder/ShineEncoderPlugin.cxx @@ -57,7 +57,7 @@ struct ShineEncoder { bool Setup(Error &error); - bool WriteChunk(bool flush, Error &error); + bool WriteChunk(bool flush); }; static constexpr Domain shine_encoder_domain("shine_encoder"); @@ -142,7 +142,9 @@ shine_encoder_open(Encoder *_encoder, AudioFormat &audio_format, Error &error) encoder->stereo[0] = new int16_t[encoder->frame_size]; encoder->stereo[1] = new int16_t[encoder->frame_size]; - encoder->input_pos = 0; + /* workaround for bug: + https://github.com/savonet/shine/issues/11 */ + encoder->input_pos = SHINE_MAX_SAMPLES + 1; encoder->output_buffer.Construct(BUFFER_INIT_SIZE); @@ -154,6 +156,12 @@ shine_encoder_close(Encoder *_encoder) { ShineEncoder *encoder = (ShineEncoder *)_encoder; + if (encoder->input_pos > SHINE_MAX_SAMPLES) { + /* write zero chunk */ + encoder->input_pos = 0; + encoder->WriteChunk(true); + } + shine_close(encoder->shine); delete[] encoder->stereo[0]; delete[] encoder->stereo[1]; @@ -161,7 +169,7 @@ shine_encoder_close(Encoder *_encoder) } bool -ShineEncoder::WriteChunk(bool flush, gcc_unused Error &error) +ShineEncoder::WriteChunk(bool flush) { if (flush || input_pos == frame_size) { long written; @@ -195,6 +203,10 @@ shine_encoder_write(Encoder *_encoder, length /= sizeof(*data) * encoder->audio_format.channels; size_t written = 0; + if (encoder->input_pos > SHINE_MAX_SAMPLES) { + encoder->input_pos = 0; + } + /* write all data to de-interleaved buffers */ while (written < length) { for (; @@ -207,7 +219,7 @@ shine_encoder_write(Encoder *_encoder, encoder->stereo[1][encoder->input_pos] = data[base + 1]; } /* write if chunk is filled */ - encoder->WriteChunk(false, error); + encoder->WriteChunk(false); } return true; @@ -220,7 +232,7 @@ shine_encoder_flush(Encoder *_encoder, gcc_unused Error &error) long written; /* flush buffers and flush shine */ - encoder->WriteChunk(true, error); + encoder->WriteChunk(true); const uint8_t *data = shine_flush(encoder->shine, &written); if (written > 0) |