It seems the root of the bug is some misunderstanding between
the parsing and the decoding code about samples/frames per block.
What is parsed as "samplesPerBlock" in the .wav file is
a _frame_ count in reality, if we follow what seem to
be libaudiofile's conventions.
If you just decode less samples as in the proposed patch
you'll always skip the second half of each buffer.
The appended patch works better for me. It still doesn't
protect against buffer overflows.
best regards
Matthias
-------------------------------------------------------------------
-------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDir'in Baerbel Brumme-Bothe
Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Prof. Dr. Harald Bolt,
Dr. Sebastian M. Schmidt
-------------------------------------------------------------------
-------------------------------------------------------------------
$NetBSD$
--- ./libaudiofile/wave.c.orig 2004-03-06 07:39:23.000000000 +0100
+++ ./libaudiofile/wave.c
@@ -220,7 +220,8 @@ static status ParseFormat (AFfilehandle
extraByteCount = LENDIAN_TO_HOST_INT16(extraByteCount);
af_fread(&samplesPerBlock, 1, 2, fp);
- samplesPerBlock =
LENDIAN_TO_HOST_INT16(samplesPerBlock);
+ samplesPerBlock = LENDIAN_TO_HOST_INT16(samplesPerBlock)
+ * track->f.channelCount;
af_fread(&numCoefficients, 1, 2, fp);
numCoefficients =
LENDIAN_TO_HOST_INT16(numCoefficients);
