[Csnd-dev] develop branch badly broken for me

classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

[Csnd-dev] develop branch badly broken for me

Dave Seidel
Hi all,

I need to report that the current develop code is badly broken for me as I try to run my Implication Organ code (though some older, simpler scripts seem to work). I went back to commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working again. Sorry, I don't have time to try to narrow down the issues. No runtime errors, but very bad sonic results using scanned synthesis and other stuff. 

I understand that you guys are cleaning up git history, but please use great care and be sure to test it thoroughly, and on my part I will not pull again until there's some indication that things have settled down. As this is my only free day to prepare for a gig on Saturday, I'm kind of freaked out, but relieved to be able to get things working by going back to an earlier commit (can't just go to the 6.11 release as I'm now relying on John's recent OSC changes).

- Dave
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

jpff
You should be able to checkout s particular commit of your choosing.  If that does not work I still have code here...

Sent from TypeApp
On 4 Sep 2018, at 18:34, Dave Seidel <[hidden email]> wrote:
Hi all,

I need to report that the current develop code is badly broken for me as I try to run my Implication Organ code (though some older, simpler scripts seem to work). I went back to commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working again. Sorry, I don't have time to try to narrow down the issues. No runtime errors, but very bad sonic results using scanned synthesis and other stuff. 

I understand that you guys are cleaning up git history, but please use great care and be sure to test it thoroughly, and on my part I will not pull again until there's some indication that things have settled down. As this is my only free day to prepare for a gig on Saturday, I'm kind of freaked out, but relieved to be able to get things working by going back to an earlier commit (can't just go to the 6.11 release as I'm now relying on John's recent OSC changes).

- Dave
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

Victor Lazzarini-2
In reply to this post by Dave Seidel
We have not touched the remote repo in our git work. What you report is something else.

Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy
Maynooth University
Ireland

> On 4 Sep 2018, at 18:34, Dave Seidel <[hidden email]> wrote:
>
> Hi all,
>
> I need to report that the current develop code is badly broken for me as I try to run my Implication Organ code (though some older, simpler scripts seem to work). I went back to commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working again. Sorry, I don't have time to try to narrow down the issues. No runtime errors, but very bad sonic results using scanned synthesis and other stuff.
>
> I understand that you guys are cleaning up git history, but please use great care and be sure to test it thoroughly, and on my part I will not pull again until there's some indication that things have settled down. As this is my only free day to prepare for a gig on Saturday, I'm kind of freaked out, but relieved to be able to get things working by going back to an earlier commit (can't just go to the 6.11 release as I'm now relying on John's recent OSC changes).
>
> - Dave
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

Steven Yi
We can bisect against 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 to
figure out what the issue is.
On Tue, Sep 4, 2018 at 2:54 PM Victor Lazzarini <[hidden email]> wrote:

>
> We have not touched the remote repo in our git work. What you report is something else.
>
> Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy
> Maynooth University
> Ireland
>
> > On 4 Sep 2018, at 18:34, Dave Seidel <[hidden email]> wrote:
> >
> > Hi all,
> >
> > I need to report that the current develop code is badly broken for me as I try to run my Implication Organ code (though some older, simpler scripts seem to work). I went back to commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working again. Sorry, I don't have time to try to narrow down the issues. No runtime errors, but very bad sonic results using scanned synthesis and other stuff.
> >
> > I understand that you guys are cleaning up git history, but please use great care and be sure to test it thoroughly, and on my part I will not pull again until there's some indication that things have settled down. As this is my only free day to prepare for a gig on Saturday, I'm kind of freaked out, but relieved to be able to get things working by going back to an earlier commit (can't just go to the 6.11 release as I'm now relying on John's recent OSC changes).
> >
> > - Dave
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

Victor Lazzarini-2
In reply to this post by jpff
Looking at the commit messages, there is some algebraic optimisation mentioned, which could be the issue. Other stuff seems unrelated: some changes to the score, faust opcodes, fixing a parsing crash, and printing fixes.

Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy
Maynooth University
Ireland

On 4 Sep 2018, at 19:50, John ff <[hidden email]> wrote:

You should be able to checkout s particular commit of your choosing.  If that does not work I still have code here...

Sent from TypeApp
On 4 Sep 2018, at 18:34, Dave Seidel <[hidden email]> wrote:
Hi all,

I need to report that the current develop code is badly broken for me as I try to run my Implication Organ code (though some older, simpler scripts seem to work). I went back to commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working again. Sorry, I don't have time to try to narrow down the issues. No runtime errors, but very bad sonic results using scanned synthesis and other stuff. 

I understand that you guys are cleaning up git history, but please use great care and be sure to test it thoroughly, and on my part I will not pull again until there's some indication that things have settled down. As this is my only free day to prepare for a gig on Saturday, I'm kind of freaked out, but relieved to be able to get things working by going back to an earlier commit (can't just go to the 6.11 release as I'm now relying on John's recent OSC changes).

- Dave
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

jpff
The algebraic optimisations affect aritmetic with 0 or 1 as one of the
arguments.  Easy to remove if it is the problem.  Only hit 1 case in te
regression tests, and none elsewhere

On Tue, 4 Sep 2018, Victor Lazzarini wrote:

> Looking at the commit messages, there is some algebraic optimisation
> mentioned, which could be the issue. Other stuff seems unrelated: some changes
> to the score, faust opcodes, fixing a parsing crash, and printing fixes.
>
> Victor Lazzarini Dean of Arts, Celtic Studies, and Philosophy
> Maynooth University
> Ireland
>
> On 4 Sep 2018, at 19:50, John ff <[hidden email]> wrote:
>
>       You should be able to checkout s particular commit of your
>       choosing.  If that does not work I still have code here...
>
> Sent from TypeApp
> On 4 Sep 2018, at 18:34, Dave Seidel <[hidden email]> wrote:
>       Hi all,
>
> I need to report that the current develop code is badly broken for
> me as I try to run my Implication Organ code (though some older,
> simpler scripts seem to work). I went back to
> commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working
> again. Sorry, I don't have time to try to narrow down the issues.
> No runtime errors, but very bad sonic results using scanned
> synthesis and other stuff. 
>
> I understand that you guys are cleaning up git history, but please
> use great care and be sure to test it thoroughly, and on my part I
> will not pull again until there's some indication that things have
> settled down. As this is my only free day to prepare for a gig on
> Saturday, I'm kind of freaked out, but relieved to be able to get
> things working by going back to an earlier commit (can't just go
> to the 6.11 release as I'm now relying on John's recent OSC
> changes).
>
> - Dave
>
>
>
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

Dave Seidel
When I have time (Sunday or after), I will try to bisect and narrow down the issue, In the meantime, the commit I mentioned above is working for me. My script is fairly complex (MIDI, OSC, scanned synth, etc.) so I won't attempt to get to a minimal example until I can narrow down to a "bad" commit. I may be doing something to trigger an edge case somewhere. In case anyone is curious, my code is here: https://github.com/DaveSeidel/music-src/tree/master/implication-organ, but I don't expect anyone else to grovel through it looking for issues, there's just too much to eliminate.

Of course you're right Victor, this has nothing to do with the cleanup effort, it is some other unintended consequence from a code change.

- Dave

On Tue, Sep 4, 2018 at 3:12 PM jpff <[hidden email]> wrote:
The algebraic optimisations affect aritmetic with 0 or 1 as one of the
arguments.  Easy to remove if it is the problem.  Only hit 1 case in te
regression tests, and none elsewhere

On Tue, 4 Sep 2018, Victor Lazzarini wrote:

> Looking at the commit messages, there is some algebraic optimisation
> mentioned, which could be the issue. Other stuff seems unrelated: some changes
> to the score, faust opcodes, fixing a parsing crash, and printing fixes.
>
> Victor Lazzarini Dean of Arts, Celtic Studies, and Philosophy
> Maynooth University
> Ireland
>
> On 4 Sep 2018, at 19:50, John ff <[hidden email]> wrote:
>
>       You should be able to checkout s particular commit of your
>       choosing.  If that does not work I still have code here...
>
> Sent from TypeApp
> On 4 Sep 2018, at 18:34, Dave Seidel <[hidden email]> wrote:
>       Hi all,
>
> I need to report that the current develop code is badly broken for
> me as I try to run my Implication Organ code (though some older,
> simpler scripts seem to work). I went back to
> commit 9f170cdc73bc9ce50157a6d9e8e96c8a2a44bcc6 and it is working
> again. Sorry, I don't have time to try to narrow down the issues.
> No runtime errors, but very bad sonic results using scanned
> synthesis and other stuff. 
>
> I understand that you guys are cleaning up git history, but please
> use great care and be sure to test it thoroughly, and on my part I
> will not pull again until there's some indication that things have
> settled down. As this is my only free day to prepare for a gig on
> Saturday, I'm kind of freaked out, but relieved to be able to get
> things working by going back to an earlier commit (can't just go
> to the 6.11 release as I'm now relying on John's recent OSC
> changes).
>
> - Dave
>
>
>


--
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

jpff
I just committed code with the suspect assignment optimisation removed,
pending an example.
==John ff
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

Dave Seidel
I'm happy to report that after getting a fresh clone of develop, the issue I had seems to be resolved -- I am able to run the Implication Organ without difficulty. Thanks, John.

BTW, I had my first public outing with the Implication Organ, in a show where I collaborated with my friend Matt Samolis, who played his beautiful bowed-metal instrument (which is entirely acoustic). I used a meta-slendro scale on the IO. Although we'd never played together before, our respective instruments blended beautifully.

- Dave

On Tue, Sep 4, 2018 at 3:36 PM jpff <[hidden email]> wrote:
I just committed code with the suspect assignment optimisation removed,
pending an example.
==John ff



--
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

joachim heintz
i find an oddity in the multiplication of a k-indexed i-array with an
i-variable:

giDurations[] fillarray 1, 2, 3, 2, 1

instr 1

  kDurIndx = 0
  printk 0, giDurations[kDurIndx] ; returns 1
  printk 0, giDurations[kDurIndx]*2 ; returns 0 instead of 2
  turnoff

endin
schedule 1, 0, 1


i am not sure whether it is related to the optimization issue, but
giDurations[kDurIndx]*2 should return 2, not?

        joachim




On 09/09/18 17:53, Dave Seidel wrote:

> I'm happy to report that after getting a fresh clone of develop, the
> issue I had seems to be resolved -- I am able to run the Implication
> Organ without difficulty. Thanks, John.
>
> BTW, I had my first public outing with the Implication Organ, in a show
> where I collaborated with my friend Matt Samolis, who played his
> beautiful bowed-metal instrument (which is entirely acoustic). I used a
> meta-slendro scale on the IO. Although we'd never played together
> before, our respective instruments blended beautifully.
>
> - Dave
>
> On Tue, Sep 4, 2018 at 3:36 PM jpff <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     I just committed code with the suspect assignment optimisation removed,
>     pending an example.
>     ==John ff
>
>
>
> --
> http://mysterybear.net
> https://soundcloud.com/mysterybear
> https://mysterybear.bandcamp.com
> https://www.youtube.com/channel/UCo0--3uN5ycq_aupZXskBgA
>
> http://recordings.irritablehedgehog.com/album/dave-seidel-60-hz
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

Mauro Giubileo

I've filed a bug report.

Mauro


Il 2018-09-10 08:45 joachim heintz ha scritto:

i find an oddity in the multiplication of a k-indexed i-array with an i-variable:

giDurations[] fillarray 1, 2, 3, 2, 1

instr 1

 kDurIndx = 0
 printk 0, giDurations[kDurIndx] ; returns 1
 printk 0, giDurations[kDurIndx]*2 ; returns 0 instead of 2
 turnoff

endin
schedule 1, 0, 1


i am not sure whether it is related to the optimization issue, but giDurations[kDurIndx]*2 should return 2, not?

    joachim




On 09/09/18 17:53, Dave Seidel wrote:
I'm happy to report that after getting a fresh clone of develop, the
issue I had seems to be resolved -- I am able to run the Implication
Organ without difficulty. Thanks, John.

BTW, I had my first public outing with the Implication Organ, in a show
where I collaborated with my friend Matt Samolis, who played his
beautiful bowed-metal instrument (which is entirely acoustic). I used a
meta-slendro scale on the IO. Although we'd never played together
before, our respective instruments blended beautifully.

- Dave

On Tue, Sep 4, 2018 at 3:36 PM jpff <[hidden email]
<mailto:[hidden email]>> wrote:

    I just committed code with the suspect assignment optimisation removed,
    pending an example.
    ==John ff



--
http://mysterybear.net
https://soundcloud.com/mysterybear
https://mysterybear.bandcamp.com
https://www.youtube.com/channel/UCo0--3uN5ycq_aupZXskBgA

http://recordings.irritablehedgehog.com/album/dave-seidel-60-hz
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

jpff
In reply to this post by joachim heintz
Fixed in but for one-dimensional array access

⁣Sent from TypeApp ​

On Sep 10, 2018, 07:56, at 07:56, joachim heintz <[hidden email]> wrote:

>i find an oddity in the multiplication of a k-indexed i-array with an
>i-variable:
>
>giDurations[] fillarray 1, 2, 3, 2, 1
>
>instr 1
>
>  kDurIndx = 0
>  printk 0, giDurations[kDurIndx] ; returns 1
>  printk 0, giDurations[kDurIndx]*2 ; returns 0 instead of 2
>  turnoff
>
>endin
>schedule 1, 0, 1
>
>
>i am not sure whether it is related to the optimization issue, but
>giDurations[kDurIndx]*2 should return 2, not?
>
> joachim
>
>
>
>
>On 09/09/18 17:53, Dave Seidel wrote:
>> I'm happy to report that after getting a fresh clone of develop, the
>> issue I had seems to be resolved -- I am able to run the Implication
>> Organ without difficulty. Thanks, John.
>>
>> BTW, I had my first public outing with the Implication Organ, in a
>show
>> where I collaborated with my friend Matt Samolis, who played his
>> beautiful bowed-metal instrument (which is entirely acoustic). I used
>a
>> meta-slendro scale on the IO. Although we'd never played together
>> before, our respective instruments blended beautifully.
>>
>> - Dave
>>
>> On Tue, Sep 4, 2018 at 3:36 PM jpff <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>>     I just committed code with the suspect assignment optimisation
>removed,
>>     pending an example.
>>     ==John ff
>>
>>
>>
>> --
>> http://mysterybear.net
>> https://soundcloud.com/mysterybear
>> https://mysterybear.bandcamp.com
>> https://www.youtube.com/channel/UCo0--3uN5ycq_aupZXskBgA
>>
>> http://recordings.irritablehedgehog.com/album/dave-seidel-60-hz
Reply | Threaded
Open this post in threaded view
|

Re: [Csnd-dev] develop branch badly broken for me

joachim heintz
yes works now — thanks.
        j



On 11/09/18 17:36, John ff wrote:

> Fixed in but for one-dimensional array access
>
> ⁣Sent from TypeApp ​
>
> On Sep 10, 2018, 07:56, at 07:56, joachim heintz <[hidden email]> wrote:
>> i find an oddity in the multiplication of a k-indexed i-array with an
>> i-variable:
>>
>> giDurations[] fillarray 1, 2, 3, 2, 1
>>
>> instr 1
>>
>>  kDurIndx = 0
>>  printk 0, giDurations[kDurIndx] ; returns 1
>>  printk 0, giDurations[kDurIndx]*2 ; returns 0 instead of 2
>>  turnoff
>>
>> endin
>> schedule 1, 0, 1
>>
>>
>> i am not sure whether it is related to the optimization issue, but
>> giDurations[kDurIndx]*2 should return 2, not?
>>
>> joachim
>>
>>
>>
>>
>> On 09/09/18 17:53, Dave Seidel wrote:
>>> I'm happy to report that after getting a fresh clone of develop, the
>>> issue I had seems to be resolved -- I am able to run the Implication
>>> Organ without difficulty. Thanks, John.
>>>
>>> BTW, I had my first public outing with the Implication Organ, in a
>> show
>>> where I collaborated with my friend Matt Samolis, who played his
>>> beautiful bowed-metal instrument (which is entirely acoustic). I used
>> a
>>> meta-slendro scale on the IO. Although we'd never played together
>>> before, our respective instruments blended beautifully.
>>>
>>> - Dave
>>>
>>> On Tue, Sep 4, 2018 at 3:36 PM jpff <[hidden email]
>>> <mailto:[hidden email]>> wrote:
>>>
>>>     I just committed code with the suspect assignment optimisation
>> removed,
>>>     pending an example.
>>>     ==John ff
>>>
>>>
>>>
>>> --
>>> http://mysterybear.net
>>> https://soundcloud.com/mysterybear
>>> https://mysterybear.bandcamp.com
>>> https://www.youtube.com/channel/UCo0--3uN5ycq_aupZXskBgA
>>>
>>> http://recordings.irritablehedgehog.com/album/dave-seidel-60-hz
>