What does pipe() mean?

[yyk808]

I mean pipe() not pipeline().
To be more precise, what are the differences between pipe() and to()?

They both input a pipeline layout. And it seems like they both redirect( is this correct? ) the data to somewhere. I got confused when reading code like https://github.com/flomesh-io/ztm/blob/main/agent/mesh.js#L150-L151.
As the question is not about ztm itself, I make a discussion here.

Anyone can help me?

[pajama-coder]

@yyk808 Hi, it's a good question. Although the issue is closed, I still feel the need to clarify the confusing usage of to(). Unlike pipe(), an invocation of to() doesn't add any filters to the pipeline, instead, it should be looked at as a part of the filter that was just added prior to it. Filters that need to construct subpipelines all need a way for users to specify a subpipeline template, by calling to() is one way. The other is giving a direct function argument like we do for pipe().

In short, pipe() is to create a pipe filter, but to() is for providing more info to the filter created before it.

[yyk808]

@pajama-coder Thank you for your clarification, now I got it.

However, there's still something I cannot understand completely about its usage. Here it is.
If I want to use pipe(), it must be in a declared pipeline.

var pp = pipeline($=>$.dump())
pipeline($=>$.onStart(new Message('Msg')).pipe(pp)).spawn() // works fine.
pipy().task().onStart(new Message()).pipe(pp) // not work: not a function.

I can't directly call it from pipy(). Is this a intentional design?