Well, I need to get this release out - ther's stuff in there that's way overdue.

I have a not-quite-consistent policy to avoid making multichannel versions of stuff that has state.  I think readsf~ and writesf~ should be exceptions to that, but anything else that has state that has to be replicated for multiple channels sounds like trouble to me.  OTOH, clip~ has no state -it's a problem for a different reason - I can't come up with a consistent policy about dealing with all the possible cases.  But then I don't think it's terribly important since a clip~ can always be replaced by a min~ and max~ pair... so my suggestion would be to use those instead if doing multichannel stuff.

All this is worth continuing to think about.

Meanwhile, I'm not sure I'll ever have the bandwidth to properly test and maintain two different precisions in addition to everything else - I think I have to leave that for others to take on.

cheers

Miller

On 5/6/24 18:40, Alexandre Torres Porres wrote:

while we're at it I also hope we can finally start providing double precision builds from miller's site 'officially' and not as 'beta' tests anymore, huh?

Which brings me to the request, can we try and provide test releases with those builds as well?

thanks

Em seg., 6 de mai. de 2024 às 13:33, Alexandre Torres Porres porres@gmail.com escreveu:

Em seg., 6 de mai. de 2024 às 04:37, Miller Puckette
<mpuckette@cloud.ucsd.edu> escreveu:

    to pd develope -

    Of course there are lots of other things to do but... assuming
    it gets
    through CI I'm hoping to throw the current master out as 0.55
    test1...
    let me know if there are any last-minute requests or objections.


Hi, been focusing on docs and manual changes and no objections for
that as there'll be time for more revisions on the list.

I don't have any easy/quick request for the test release but I do
clearly have a great deal of hopes and expectations for several
PRs over there.

And I honestly don't see that many changes yet for the test
releases to start popping up and I wonder if you can be clear on
some of the things from the lots of other things to do yet.

I'll take the opportunity to reinforce that I hope we can focus on
more multichannel features and list PRs again.

- https://github.com/pure-data/pure-data/pull/2110 add secondary
signal inlets to clip~
- https://github.com/pure-data/pure-data/pull/2059 Add functions
to draw multichannel connections
- https://github.com/pure-data/pure-data/pull/2058 allow to change
the number of tables/channels in table DSP objects
- https://github.com/pure-data/pure-data/pull/2052 multichannel
support for readsf~ and writesf~
- https://github.com/pure-data/pure-data/pull/2049 multichannel
support for delay objects
- https://github.com/pure-data/pure-data/pull/1978 Some missing
multichannel objects

And also, something that's kinda bothering me but not related to
this is if we can 'fix' and have a more accurate sqrt~ and cosine
table for 0.55 (with a beloved compatibility flag)
https://github.com/pure-data/pure-data/issues/1906
cheers


    cheers

    Miller




    _______________________________________________
    Pd-dev mailing list
    Pd-dev@lists.iem.at
    https://lists.puredata.info/listinfo/pd-dev

Em seg., 6 de mai. de 2024 às 22:03, Alexandre Torres Porres < porres@gmail.com> escreveu:

Taking https://github.com/pure-data/pure-data/pull/2289 of course could also be a good last minute request by the way

oops, that'd actually be https://github.com/pure-data/pure-data/pull/2293 now

Yep, for sure :)

thanks

M

On 5/7/24 11:21, IOhannes m zmoelnig wrote:

On 5/6/24 09:36, Miller Puckette wrote:

...

to pd develope -

Of course there are lots of other things to do but... assuming it gets through CI I'm hoping to throw the current master out as 0.55 test1... let me know if there are any last-minute requests or objections.

today i managed to tackle down a crasher bug reported on raspberry pi.

it turns out that it only appears on arm32 architectures (and when compilers do not initialize values to 0, which they are allowed to do). while there, i build with "-Wuninitialized" and found a few more variables that might get used without being set to any value, which i also fixed.

https://github.com/pure-data/pure-data/pull/2295

sorry for being late, but i think this ought to go in.

gmfasdr IOhannes

---

Pd-dev mailing list Pd-dev@lists.iem.at https://lists.puredata.info/listinfo/pd-dev