This may seem like a trivial issue, and it may be limited to Windows (I haven't tested it on Linux or Mac OS), but when I have a canvass in an abstraction (in order to give it a coloured background), the inlets and outlets aren't visible in the parent patch (see screen shots). If I make the canvass smaller than the size of the box displayed in the parent patch, or offset it to the right or below, then they are displayed, but the coloured background only covers part of the box shown in the parent patch. And if I remove the canvass, then they're displayed, but I can't give it a coloured background.

It doesn't seem to make any difference whether I add the inlets and outlets before or after I add the canvass, the little tabs in the corner don't get displayed in the parent patch either way.

I know this is purely cosmetic and doesn't affect the behaviour in any way, but am I doing something wrong? Or is this just the way Pure Data works?

Visible:

Not visible:

@dfkettle said:

and it may be limited to Window

Not only a windows problem...that's one of so many bad UI design and annoyance of puredata vanilla...
To fix that, you can create fake inlets using veryyyy small canvas on top of background (ex: canvas size=5, height=2 and selection square should be size 2 also). Then you will have to zoom one time to properly select them. Good luck...

@60hz Thanks. I thought about doing something like that, but it's too much work. I guess I'll just live with it.

@dfkettle Maybe it was a conscious choice by Miller?
Coloured canvasses through GOP windows are most often used (maybe only used?) to build marginally prettier GUI's, and hiding inlets and outlets behind the coloured background certainly helps with that.
Direct connections (better for minimising block latency) can be totally hidden behind the canvasses that make up the patch...... so sends and receives are unnecessary.
http://www.pdpatchrepo.info/
David.

@whale-av I'm not sure why you're asking me, I wouldn't know why he made the decision (if it was him).

Just to be clear, I'm talking about the appearance in the parent patch. If you have a canvass in the sub-patch that covers the entire rectangle displayed in the parent, the inlets and outlets at the top and bottom of the object aren't displayed. I'm not sure what it has to do with direct connections vs. sends/receives. The point is, a feature known as a "canvass" was provided which, if you choose to use it can conceal the inlets and outlets, but only if it covers the entire visible area in the parent patch. As I said, it's purely a cosmetic issue. Maybe canvasses are totally unnecessary, they just make the patch "marginally prettier", as you say (although making Pure Data a little prettier might not be a bad thing). But if you're going to provide a feature, shouldn't it behave consistently? Maybe there's some technical reason for this, that I'm not aware of, though.

@dfkettle I also think that it is the indented feature to hide the inlets and outlets so they would not interfere with the "prettiness" of the GUI.

I also use dummy names in the send/receive fields of GUI elements to hide their inlets and outlets often times.

@dfkettle Sorry........ I was not asking you. I meant to suggest that it is by design.
The results do visually improve a patch and I can think of no other use for a coloured canvas.
David.

@ingox @whale-av OK, I get that, and I know that numbers, toggles, etc., work that way. But I don't want to hide the inlets and the outlets of my abstraction in the parent patch. Here's a screen shot of what I'm working on (a "work in progress"), just to better illustrate what I'm trying to do. The abstraction encapsulates a MIDI sequencer, but I want to be able to chain multiple sequences, both sequentially and in parallel. So I want the inlets and outlets to be displayed, to indicate to the user that they can be chained. But if I give them a canvass background (even if it's white), they don't display. Like I said in my original post, it's rather trivial and I can live with the way it works. I just thought it might be something I was doing wrong, but it seems that it was meant to work this way.

This is what the abstraction looks like if you click on the "Edit" button.

@dfkettle Nice and yes.

@dfkettle Please don't ask (because so far I have no idea..... it's a hack )....... hmm.zip
It seems to stick, but might need [initbang] as even a [loadbang] + [delay] wil not automatically activate it.
David.

Hmm

??????????????????????????????????

@whale-av Amazing ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !