I received a message from C.R.Helmrich (he doesn't want to subscribe
just to post a single message).
The following message was posted by C.R.Helmrich
[begin]
Hi,
I tested my own Win32 compile of pre1 and pre2 one or two weeks ago
(flac.exe
through foobar2000 for transcoding), and the results were as expected: no
errors,
roughly same speed on my Intel Core i...something mobile, very slightly
improved
compression performance in comparison with 1.3.1). I didn't test all
-1..-8 modes,
but I diffed the source code of libFLAC between 1.3.1 and 1.3.2, and it
seems there
were no relevant changes to the internal compression routine aside from
the LPC
calculation, so I don't think this was/is necessary. Other than that:
In libFLAC/src/stream_encoder.c, the following C++ style comment was added
recently:
// FLAC__ASSERT(samples <= blocksize);
Since all other comments in this file (and, possibly, the entire lib) are
C-style /* */,
I suggest to remove this line again.
Best,
Chris
[end]
P.S. The line was added in the commit
<http://git.xiph.org/?p=flac.git;a=commitdiff;h=c06a44969c1145242a22f75fc8fb2e8b54c55303>
P.P.S. actually there are more cpp-style comments in libFLAC.
Many of them are inside SSE intrinsic functions... (sigh)
but some are much older, for example:
<http://git.xiph.org/?p=flac.git;a=commitdiff;h=f299910ddb90a33b2866d0ae27e687de8bbc154e>
It seems that the oldest cpp-style comment in libFLAC is the following
line from src/libFLAC/include/protected/stream_encoder.h:
#endif // #ifndef FLAC__INTEGER_ONLY_LIBRARY
--it exists since FLAC 1.1.3 (see commit
<http://git.xiph.org/?p=flac.git;a=commitdiff;h=bf0f52c21e439aefdbcf71551621023b0bc9cef0#patch16>)
_______________________________________________
flac-dev mailing list
flac-dev@xiph.org
http://lists.xiph.org/mailman/listinfo/flac-dev
