PEP 659: Link to final version of Brunthaler paper

[gvanrossum]

No description provided.

[gvanrossum]

Let's wait until Brunthaler approves of the change. This is the highest-ranked link for the paper that I found through Google, but I don't know if the author has a preference.

[hugovk]

I don't have an ACM subscription and only see the abstract and references at https://dl.acm.org/doi/10.5555/1883978.1884008

Their website isn't very clear on how I could access it, after a few minutes of browsing looks like I'd need "Professional Membership plus ACM Digital Library: ($198 USD)".

Alternatively, googling the title, these are my top results:

1. https://publications.sba-research.org/publications/ecoop10.pdf
2. http://www.complang.tuwien.ac.at/kps09/pdfs/brunthaler.pdf
3. https://www.researchgate.net/publication/221496475_Inline_Caching_Meets_Quickening (has PDF download)
4. https://dl.acm.org/doi/10.5555/1883978.1884008
5. https://www.semanticscholar.org/paper/Inline-Caching-Meets-Quickening-Brunthaler/7a8705061003ae18bc0a783fa65b705d5d1d4bd7

[sbrunthaler]

Thanks a lot for the PR!

I agree with @hugovk regarding the rather expensive ACM membership fees. We (i.e., the authors) are allowed to host author versions of the papers on our web pages, I just hadn't really gotten around to doing this yet. The ECOOP'10 paper detailing the inline caching with quickening idea is now available via the following link: https://www.unibw.de/ucsrl/pubs/ecoop10.pdf/view

A more complete description, also w.r.t. quickening LOAD_GLOBAL instructions, and the optimized interpreter instruction format I have been using is available in my PhD thesis, which I've also put online: https://www.unibw.de/ucsrl/pubs/thesis.pdf/view
(N.B. There is an implementation error in the algorithm implementing partial stack-frame caching.)

All the best,
--stefan

[AA-Turner]

CI failed as this branch is before the preview support changes

A

[gvanrossum]

@hugovk I updated the PR with a new link, provided by the author, that's freely downloadable.

I would just merge at this point, but the RTD preview test failed. If that's expected, can you merge it for me? If that's due to a problem with the PR, could you help me debug it? The traceback in the build log at https://readthedocs.org/projects/pep-previews/builds/16342394/ doesn't have much of a clue, and I don't have all that it needs to attempt a local build.

[AA-Turner]

If that's expected, can you merge it for me?

I'm not Hugo, but as above it was expected due to timing of when things were merged.

A