Is there a channel to register issues with deken packages?

I chose (unwisely, as it happens) to release an abstraction that depends on PDContainer.

In Linux, the most recent upload is from 2015, and it ships separate binaries for each object: [declare -path pdcontainer].

Then I found later that the latest Windows upload is from 2018, shipped as a library: [declare -lib PDContainer] (note also that the case of the folder name is inconsistent with the other upload, in a not useful additional source of confusion).

So, OK, I can do [declare -path pdcontainer -lib PDContainer] and accept that I will get a spurious error message in Linux.

But it would be better if these could be made consistent... but I'm not sure where to take that request. The original developer of PDContainer is Georg Holzmann; the Linux uploader is chr15m; and the Windows uploader is lucardo -- so I guess it would be pointless to approach the original developer... but then, where to go?

hjh

@ddw_music I suppose Deken issues should be filed here....... https://github.com/pure-data/deken/issues

Lucarda has been hyperactive updating Pd external packages for 64-bit during the last few years (thank you @Lucarda). You could send a pull request here....... https://github.com/Lucarda/pure-data/pulls

The source code is here..... http://grh.mur.at/software/pdcontainer.html so maybe you can compile the library, or include the tarball with a 64-bit makefile in your patch folder if different flavours of Linux require it.
David.

@whale-av said:

The source code is here..... http://grh.mur.at/software/pdcontainer.html so maybe you can compile the library, or include the tarball with a 64-bit makefile in your patch folder if different flavours of Linux require it.

Thanks for the suggestions. I guess building for Linux (and Mac -- discovered that the 2018 upload seems to be only for Windows) is the best way... if I can, or have time.

hjh

@ddw_music Did you find anything out about this? I am thinking about adopting this one as I have started to fix a few irritations I have with them, mainly sick of them filling the pd window with error messages when the right outlet gives us what is needed if we want such messages but also to make them conform to pd conventions with a right inlet for setting name, and your issue. Seems an abandoned project from the extended days but I have not been able to find anything regarding if @Lucarda or someone else is maintaining them or just compiling them for different architectures as needed or what.

@oid said:

@ddw_music Did you find anything out about this?

I didn't (busy week). If you have some other fixes in mind, I would certainly not mind leaving it to you

I am thinking about adopting this one as I have started to fix a few irritations I have with them, mainly sick of them filling the pd window with error messages when the right outlet gives us what is needed if we want such messages but also to make them conform to pd conventions with a right inlet for setting name, and your issue.

Yeah, the posting seems over the top. In my scheduler, I avoided the error messages by "pop"-ping only at the moment of outputting a message... which can happen only when a message was retrieved by "top," so I'm certain that the queue is not empty.

Another good fix to make is that (I guess) it used to be called h_priority_queue, and was later renamed to h_prioqueue without updating the help patch. So there's h_priority_queue-help.pd in which "h_priority_queue $0-aetsch... couldn't create." Fair enough if the author got tired of maintaining it, but still, it isn't tidy.

It would be a pity for these to fall into disrepair. PDContainer does include some data structures I've grown accustomed to in SuperCollider but missed in Pd.

hjh

@ddw_music I am sort of thinking I might just fork it so I can fix all my issues with pdcontainer, especially make it follow the object per function standard of the other named objects like [array] and [text], creates neater patches most of the time. Still will fix those two bugs in pdcontainer and the log issue, only a 5 minute job.