Arthur Taylor

Looked into this, and it is fixable. The reason Ogg encapsulated files currently require a seek is because the container code in src/ogg.c loads the first packet of the first...

May have been fixed inadvertently by #457, need to retest.

Rounding is something that we should address and solve. > I very much doubt that setting and restoring the rounding mode with SSE2 intrinsics would noticably influence performance I concur....

> You mean it does? I'm confused, the spec seems to suggest lrint honours the rounding mode It honours but doesn't change the rounding mode. A C int cast of...

When a test fails, the audio fail used in the test is not deleted. Could you check your working directory for a *.mp3 file that the failed test created and...

Okay, I have a theory. SD2 files use a Classic Mac OS "resource fork" to store file metadata, including all headers, while the actual file is just raw PCM. (On...

A fix for this issue would be to move the attempt to identify a file as MP3 based on a MPEG sync frame to after looking for a resource fork,...

mat4.c ``` int rows, cols, imag ; ... psf->dataend = psf->dataoffset + rows * cols * psf->bytewidth ; ``` So clang is complaining that the result of `rows * cols`...

Thanks for your report! As seeking is implemented per-subformat, I've edited the title to reflect that this is an issue with the FLAC subformat handling. Would it be possible for...

This behaviour should be changed to be changed to be consistent with other formats. When the ogg code hits a zero-length-read for the first time, it sets the end-of-stream bit...