Why don't you become a Csound developer? I would prefer that. If there's
some reason why this is impossible, feel free to send me the sources.

I now have -B and ksmps now working independently of each other, so if you
could review my fix in pa_blocking.c in CVS and update your proposed code
for full duplex, I can put that in as well. The reason the code was not
working is that the paBlockingWrite call is supposed to be passed a count
of samples (not bytes, not sample frames, but samples), but it was being
passed a count of bytes; and maxLag is a count of sample frames, but it was
being interpreted as a count of samples. As a result the buffer reads were
in invalid memory causing a crash.

I appreciate your work, and I certainly hope you keep contributing.

Original Message:
-----------------
From: Victor Lazzarini Victor.Lazzarini@nuim.ie
Date: Tue, 11 Jan 2005 13:28:45 +0000
To: csound-dev@eartha.mills.edu
Subject: [CSOUND-DEV:5671] GEN01 bug fixes



I noticed today that csound5 GEN01 was twice broken:
(a) did not load all the contents of a file
(b) did not properly create a deferred allocation table.

The first bug was easy to spot and fix, as I saw the same
thing before: there was a mix-up between counting bytes
and counting samples. Done.

The second one was a little more involved, as the gen loader
ftgens was not calling the normalise/display function. So the
rescaling was all messed up. What I did was to call it
from gen01raw(). It works.

Now the same probably needs to be done by the other gens that
use deferred allocation, but I haven't done that yet.

To John Ffitch:
can I send you the fixed sources?

Victor

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 


--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .

John Ffitch's has given me a developer status; however I am battling
with my LAN administrators in the uni to get CVS to work; currently
it doesn't. So I have to rely on someone else's help to check in my
changes.

I'll certainly have a look at your improvements. I'll get the new pa_blocking
code from the CVS browser. In the meantime, I'll send you the fixes I
have.

Victor



At 14:36 11/01/2005, you wrote:
>Why don't you become a Csound developer? I would prefer that. If there's
>some reason why this is impossible, feel free to send me the sources.
>
>I now have -B and ksmps now working independently of each other, so if you
>could review my fix in pa_blocking.c in CVS and update your proposed code
>for full duplex, I can put that in as well. The reason the code was not
>working is that the paBlockingWrite call is supposed to be passed a count
>of samples (not bytes, not sample frames, but samples), but it was being
>passed a count of bytes; and maxLag is a count of sample frames, but it was
>being interpreted as a count of samples. As a result the buffer reads were
>in invalid memory causing a crash.
>
>I appreciate your work, and I certainly hope you keep contributing.
>
>Original Message:
>-----------------
>From: Victor Lazzarini Victor.Lazzarini@nuim.ie
>Date: Tue, 11 Jan 2005 13:28:45 +0000
>To: csound-dev@eartha.mills.edu
>Subject: [CSOUND-DEV:5671] GEN01 bug fixes
>
>
>
>I noticed today that csound5 GEN01 was twice broken:
>(a) did not load all the contents of a file
>(b) did not properly create a deferred allocation table.
>
>The first bug was easy to spot and fix, as I saw the same
>thing before: there was a mix-up between counting bytes
>and counting samples. Done.
>
>The second one was a little more involved, as the gen loader
>ftgens was not calling the normalise/display function. So the
>rescaling was all messed up. What I did was to call it
>from gen01raw(). It works.
>
>Now the same probably needs to be done by the other gens that
>use deferred allocation, but I haven't done that yet.
>
>To John Ffitch:
>can I send you the fixed sources?
>
>Victor
>
>Victor Lazzarini
>Music Technology Laboratory
>Music Department
>National University of Ireland, Maynooth
>
>
>--------------------------------------------------------------------
>mail2web - Check your email from the web at
>http://mail2web.com/ .

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 

Got the latest snapshot and updated the pa_blocking/rtpa
code from the CVS browser.

I tested the new csound5 pa_blocking.c code and I got
constant dropous (at every buffer it seems). What's the
suggestion for -B (do we also need -b=ksmps?). I tried
a number of configurations, but have been unlucky so far.
Csound also seems to crash on exit.
I'll keep trying and see if I can discover anything.

I also noticed that FLTK now seems to be broken; csound
crashes when trying to run FLTK orcs. Looks like it
happen after the latest major changes (the csound arg
addition to opcodes). I have an earlier, stable, code which
is running fine. Can anyone check this?

Victor


Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 

I have some new opcodes, which I would like to send in.
Now it seems that cs4.23 and cs5 have more
than just trivial differences between the opcode API.
So, I have prepared different versions for the two.
I expect then that cs5 ones will of course go to the opcodes
directory in the CVS, but what about cs4.23?

Is anyone keeping a repository of new opcodes for 4.23?

Victor


Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth