Convert all webrtc code to not access EncodedImage::_size directly.

Read using capacity() method, write using set_buffer() method. This is
a preparation for making the member private, and renaming it to
capacity_.

Bug: webrtc:9378
Change-Id: I2f96679d052a83fe81be40301bd9863c87074640
Reviewed-on: https://webrtc-review.googlesource.com/c/113520
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25934}

modules/video_coding/codecs/h264/h264_decoder_impl.cc

@@ -250,7 +250,8 @@ int32_t H264DecoderImpl::Decode(const EncodedImage& input_image,
 
   // FFmpeg requires padding due to some optimized bitstream readers reading 32
   // or 64 bits at once and could read over the end. See avcodec_decode_video2.
-  RTC_CHECK_GE(input_image._size, input_image._length +
+  RTC_CHECK_GE(input_image.capacity(),
+               input_image._length +
                    EncodedImage::GetBufferPaddingBytes(kVideoCodecH264));
   // "If the first 23 bits of the additional bytes are not 0, then damaged MPEG
   // bitstreams could cause overread and segfault." See