search

sparkslabs / pyxie

Little Python to C++ compiler
Apache License 2.0
91 stars 23 forks source link

Not accepting Issues at present (please read before opening an issue) #31

Open sparkslabs opened 2 years ago

sparkslabs commented 2 years ago

Please Read this

Hi,

Hopefully you're reading this because you've looked at Pyxie and are considering opening an issue. If you do, I'll likely close it.

Why?

This isn't the day job, and frankly, I don't have the time to accept and manage issues at present. The code is here for you to use as you see fit within the license, but opening an issue here is not likely to result in the solution you're after. I'm mentioning this to avoid you wasting your time right now.

Comments and general observations and so on I'm happy to see as comments on this issue, but please don't be offended if your query or comment languishes for a bit. Similarly, I'd find pull requests interesting, but unless your approach aligns with the approach I'm interested in taking, please note I'll give you hopefully constructive comments and positive feedback, but at present I'm unlikely to merge your PR. (at present)

I'll probably close this issue when life changes (likely to be after the pandemic).

sparkslabs commented 10 months ago

If you see this and think "this issue is 2 years old, have things changed" - a minor update.

This issue is still current. I think in hindsight, the comment about "I'll probably just close it", while true might seem a bit harsh. However, it has stopped spurious PR's coming through and I think that means it's saved people time and hair.

What I wasn't aware of when I opened this issue was that life would actually get much harder on a personal level barely 10 days later. Since then life has remained extremely complex and difficult. At some point I'd like to be able to close off this issue, but sadly it's not now. I'll hopefully revisit this at some point within 6-9 months.

sparkslabs commented 7 months ago

2-3 months on, revisiting this again. I'm minded to close out this issue. I need a more general discussion about project status and so on, but I've got some ideas about moving this forward, in particular in enabling guild style systems to run:

And so doing that re-opens this conceptually at least. Furthermore, there some specific ideas in this outlined roughly in a mini-guild implementation hashed out in Summer holiday '22, which I think would be an interesting set of test cases for my code.

Specifically this also has some discrete steps here:

These were written about for a blog that was never written, but would be useful as implementation tests.

There's still no good way to compile generators to C++ in a performant way in other projects, which is a shame, despite me having a surprisingly good way from around 20 years ago in Kamaelia.

This approach IS used in pyxie and results in readable generated C++ code as well as performs well.

Extending this to guild style systems and so on could be interesting, but also potentially very useful.

I'll not close off this issue until I'm ready to start at least writing some design notes/etc and that's very dependent on time and frankly the issues noted in the ticket above of which can be summarised as "things hitting the fan" so to speak in 2021 (as well as 2020) have still continued on some levels - so I would not like to mislead anyone about activity levels, but I do aspire to come back to this since I think there's some real value to be had with a useful properly compile-able subset of python - especially one that can be used in constrained environments. The obvious one there is embedded systems, but web assembly is another potential target.

Enabling pyxie to target other development languages too might be useful.