Hello,

me again. Right now I am getting nasty bug. Time from time upon opening my patch the [gate] object from cyclone library is converted to normal Message object and disconnects of course all cords.
This behavior is completely random and I was wondering if anyone else witnessed it and / or
resolved it.

I am using Windows version of Pd 0.38.4-extended-RC8 and all these libraries are directly from the distribution. I also double checked for possible conflicts and they are none. Even if the patch is loaded improperly and I manually create the gate object, it works well. So it must be something with the loading the patch before all libraries are loaded?

Thanks for any hints.

Hi,

I have the following problem: I have a text file, which contains some heterogenous data. Let me show you some example:

LRSObjectFile;
0;
1;
test.ogg;
3;

The first Line tells PD that the file is intended for the patch. The second line
tells PD what type of object to load (may be a constant later down the road) and the
rest contains data specific for that object (I have many dynamic objects which I compose
on the fly through dyn~ external).

Now what I am doing to parse such a file is highly stupid and I'd like to find some more elegant solution:

I am using three textfile objects (each for one line). What I'd like to have is one textfile object and figure, how to route the reading logic so that the output from the textfile is routed to the branch I need. Here is some pseudocode through tmp status variable:

int status = 0;

switch(status)
{
case 0:
read_and_verify_header();
status = 1;
break;
case 1:
read_object_typ();
status = 2;
break;
case 2:
read_and_fill_object_data_to the_end_of_the_file();
}

But I found this very hard to construct in PD. Any ideas or pointers in completely different direction?

Thanks a lot for giving this a thought.

Edit: for some spelling.

Thanks for the link. I tried this but unfortunately it behaves in the same manner. Also what's strange is, that I have for example 4 gate objects in one abstraction and two of them loads well while the others load badly. And it's never the same. But I'll try to find some workaround.

Also another thing is, that if I open the abstraction directly, it always loads correctly! The bug occurs only when the abstraction is loaded from the whole patch structure. Since the system is modular, I have to heavily rely on abstraction logic so I don't want to go the subpatch way.

Thanks again.

thanks for the support! Yes. As soon as my project reaches beta stage, I'll set a site for it together with a download. I'll definitely look into those two objects, thanks for the advice.

Nevermind. I found a solution through a series of spigot objects.

Thanks again.

Edit: and after discovering cyclone [gate] object the solution is elegant in the way I wanted to have it.

Hi,

I'll try to make myself a little bit more clear:

I need to read a textfile and according to the contents of this file invoke different pd abstractions. The idea of this is that I already have a mechanism, which allows me to register an object which can be basically anything from ring modulation to playing a sample. My mechanism allows me to store this kind of object to the text file (database would have been much better, but I need to write some mysql external for that first )

Then when the patch enters the performance mode, I dynamically load the text files (according to the time queue) and construct the patch accordingly, but the parser doesn't know, what data's are inside. So first line is the header which describes, that the file indeed is intended for the system usage. Second line is so called object descriptor, which determines what kind of data will follow.
Then I invoke specific loader for each object, which already knows how to treat the data inside of the file.

My problem is with the textfile object itself. If I could get and object descriptor and use the same file on various places, the problem would have been solved for me, but right now it's very awkward.

Sorry for not making myself clear on the first occasion and thanks again. I hope this clarify a bit, what I'm trying to do.