Thanks very much, Anthony.
I hope your changes are incorporated into the final release of 5.08.
Presumably this problem may have been partly responsible for the lack of use
of <CsVersion>, as john ff observed.
This correction should be particularly helpful to musician/performers
presenting live works by others, and who are not otherwise Csound
practitioners. (I'm sure this group of performers is not large; hopefully,
with care on our part [dealing with user issues like this one], the group
will multiply in the future. At least, this is my hope.)
----- Original Message -----
From: "Anthony Kozar" <mailing-lists-1001@...>
To: "Csound Developer list" <csound-devel@...>
Sent: Monday, January 28, 2008 11:20 PM
Subject: Re: [Cs-dev] <CsVersion>
> Art Hunkins wrote on 1/28/08 9:38 PM:
>> I've been exploring the <CsVersion> tag in WinXP, and it does not seem to
>> behave correctly. Could a developer please check it out?
> Yes. It was broken but I have fixed it now. (Of course, versions of
> from 5.00 to 5.07 will fail to check this tag correctly, potentially
> it less useful).
> There were two problems. The first was an internal calculation of the
> version number. The second was that "Before x.y" and "After x.y" were
> OK with version x.y where the manual says that they should reject it. I
> have made the check match what the manual says the behavior should be.
>> This is what I've found:
>> I'm running Csound 5.06.
>> The CsVersion code I tested is:
>> After x.x
>> Csd's run fine with the following values for x.x:
>> 4.99, 5.06, 5.07, 5.08, 5.09, 5.6, 5.60, 5.7, 5.8, 5.9
>> The following values stop Csound with a "Decode failed...stopping"
>> 5.70, 5.80, 5.85, 5.89, 5.90, 5.99
> Note that 5.80 and 5.8 are not the same. 5.8 is interpreted the same as
> 5.08. The minor version number of Csound is always supposed to be two
> digits long.
>> (Please note as well that a Windows version 5.04.2 exists. Does CsVersion
>> handle this second decimal correctly?)
> No. I am not sure how to modify the check for this in a way that will
> it compatible with the currently documented behavior. (Not that it could
> not be done ... but I don't want to potentially break it again :)
>> I'd also recommend that the somewhat cryptic "decode failed" message be
>> changed to:
>> "A version of Csound > x.x is required"
>> or alternately:
>> "A version of Csound < x.x is required".
> I have improved the error reporting for this and I will commit all of my
> changes after making some more needed improvements to messages about CSD
> "decoding" errors.
> Anthony Kozar
> mailing-lists-1001 AT anthonykozar DOT net
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> Csound-devel mailing list