Why not test with a proven client on some PC to get reference data?
Aside this, I am curious how the current Solaris client code (->CVS) performs.
Take Winstons results
> id=Ya4FaFDnbkWD.XOCRKQH4KTvBL8-
> tracks=12
> toc=1+12+998498+278+83118+149558+232398+315950+410938+515698+596418+669490+739278+838430+919510
and apply the transformation I posted earlier:
Track 1: 278 -> (278 - 150) / 4 + 150 = 182
.
.
Track 12: 919510 -> (919510 - 150) / 4 + 150 = 229990
Lead out: 998498 -> (998498 - 150) / 4 + 150 = 249737
Using
http://www.cdindex.org/cgi-bin/cdi/xget.pl?albumid=58
one gets this stored TOC
<Title>Achtung Baby</Title>
<NumTracks>12</NumTracks>
<Id>C7Nxdg0Z5YpJFZ3B9LVHF.dnPf0-</Id>
<TOC First="1" Last="12">
<Offset Num="0">249737</Offset>
<Offset Num="1">182</Offset>
<Offset Num="2">20892</Offset>
<Offset Num="3">37502</Offset>
<Offset Num="4">58212</Offset>
<Offset Num="5">79100</Offset>
<Offset Num="6">102847</Offset>
<Offset Num="7">129037</Offset>
<Offset Num="8">149217</Offset>
<Offset Num="9">167485</Offset>
<Offset Num="10">184932</Offset>
<Offset Num="11">209720</Offset>
<Offset Num="12">229990</Offset>
</TOC>
compare it and you see why I believe that it should work now.
Regards,
Marc
P.S.
Winston, was no additional -libfoo statement necessary?