Hello there,
reporting as per subject, fluid.plotter is broken in the latest Pd 0.54, at least on my machine, Linux Mint 21.
The object is created fine, but when one clicks with the mouse on it, the mouse draws a standard Pd array, rather than navigating the datapoints (datapoints disappear in the background of the array).
Unfortunately I’m in a production rush (got a premiere very soon), so I had to revert back to Pd 0.52.1 to keep working. I can try to help debug by installing flucoma on an old laptop with the new Pd version, if it’s helpful (cannot ensure when though, but do let me know, I’ll try my best).
ok I can reproduce, watch this space, it seems there is an issue with my xy.pd hack. that object needs recoding in C and that will happen but for now I’ll have to find what changed…
I just need to troubleshoot xy.pd but that is already a nightmare - I reckon the donecanvasdialog order of magic numbers has changed (which would mean a nightmare-ish checking of version for our thing to work)
I need to check ELSE’s code to see what was updated there
ok I’ve been in touch with Porres via the pd-discord and also the list. Nobody could confirm but I can confirm there is a change of behaviour in the superimposed structures getting (or not) the click message.
I’m doing a full bug report now, it messed my brain but I got it in the end I don’t know if we can do a fix until the fix Pd (or if that was a fix on their side that messed up behaviour forever)
ok I’ve recoded it all natively. It is not much faster but at least it should work on all oses. I added a few things too. still no zoom though.
If you are not set to compile on your platform, I can provide binaries. any testing is welcome. It is a drop-in replacement but you have to set the size (yes it is scalable. oops I forgot to add the arguments in the help but use the properties editor)
@tremblap sorry for the delay (life in the way), I tested the new binary and seems to work nicely with the latest pd 54.1 on Linux Mint!
Thank you!
there’s only an unimportant graphical glitch, whereby highlighted dots very close to the edges of the graph do not disappear once highlighted. See screenshot. But it’s only a graphic hiccup, and I know wish and tcl don’t really refresh graphics too well.
Note for others who may replace the old plotter with these new ones in their patches: on my linux computer and Pd, I had to go to the patches I had created with the old plotter, select plotter properties and indicate a larger dimension. For some reasons, the object resetted its dimension to 10x10 and was so small it took me a while to understand what was going on.
The new help file works fine though, so it may only be retroactive, or maybe it was just in my case…
I cannot get that to behave. I would need to mask the graph drawing on the boundaries and I just cannot find a way to do that!
indeed that is a breaking change - the object can now have any size so old code will need either to add x y (as in [fluid.plotter 300 200]) or right-click and open the property inspector and set them there. it also has send and receive and can draw its frame or not.
hey @tremblap thanks for the notes! I’ve been waiting to reply so I could test it more intensively and I can confirm it works super fine
The graphic hiccup is not a problem at all. At least I’m using one plotter in real time in a patch and I see no problems (sometimes pd wish begins stuttering when graphics are too heavy or misbehave).
I don’t know if we can do a fix until the fix Pd (or if that was a fix on their side that messed up behaviour forever)
