[Mlir-commits] [mlir] [MLIR][Tensor] Canonicalize fully covering slice insertions into tensors with unit prefixes (PR #92912)

[Stella Laurenzo]

stellaraccident wrote:

> > Minor correction: this thread is full of rebuttals from three different contributors in good standing to this dialect, but they are not passing your bar for listening and letting the conversation transpire.
> 
> I don't parse your sentence, are you saying that in this thread, someone rebutted one or multiple of these three statements and I've been blind to it?
> 
> * these are two operations from the same dialect,
> * operating at the same level of abstraction,
> * and the semantics of the two forms seems completely equivalent

I'm sorry you don't parse my statement, but I think that just reinforces that this topic needs to be upleveled. People are rebutting your conclusions. They may not be agreeing with you but are struggling with the form of the argument as you are approaching it. To be blunt, you have a keen mind for rigorous logic. This is an asset because it lets you see inconsistencies at the outset. But not everyone can debate in that way in real time on a deeply nested issue. My experience is that most people need a minute, a zoom out and a helping hand to get to the best outcomes. A well developed ability to identify inconsistencies like you have might be a bit less alarming to you when you see less structured reasoning if you focus more on guiding people to eventual consistency of their rationale vs requiring it on each step. At least that's a lesson I had to learn in order to be able to integrate broader perspectives. 

Sorry if this is off topic. I do get very strong feedback on the impact when we go into laser focus, raid fire mode on this stuff, and I'm trying to make us better. I apologize in advance if any of it doesn't land right and an willing to discuss more in person or on an upleveled thread.

https://github.com/llvm/llvm-project/pull/92912