I'm having a go at the nearly 12-year-old bug. There is a fix in mpg123 trunk now. At least you get an error message now when the file cannot be flushed to disk. It results in mpg123 aborting with non-zero exit code only when the file header (or at least a single byte at the beginning) cannot be flushed. When the out-of-disk occurs later, you get the error message from a final flush on closing the file, but the return value of that is not handled (would be some PITA to add that).
Perhaps I'll consider dropping usage of STDIO from the file writer, since adding fflush() all around to detect the condition earlier kindof defeats the idea of buffered streams. Anyhow, the user is informed with the next major mpg123 release and that is hopefully enough to close this bug. Alrighty then, Thomas
signature.asc
Description: PGP signature
