Re: pypy stackless.py compatibility notes
Anselm Kruis <[email protected]> Mon, 28 Jul 2014 09:41:29 +0200
| Newsgroups | gmane.comp.python.stackless |
|---|---|
| Message-ID | <[email protected]> |
Hi, now I'm back from EP2014. Time for a short summary: Currently the compatibility between the stackless implementation of PyPy = and Stackless is - at best - limited. This is caused by two facts: - First, Stackless improved and enhanced its API during the last year = quite a bit. PyPy didn't catch up yet. - Secondly, the Stackless API by itself is not sufficiently well = defined. Obviously some functions/methods/attributes a clearly part of = the API (i.e. stackless.run) and others a clearly implementation = specific (i.e. stackless.cstack). But for others the situation is less = obvious. Therefore I started the project "stackless-testsuite". The aim is to create a test-suite, that covers/defines the "Stackless API" and = runs on every implementation: Stackless 2.7.x, Stackless 3.x, PyPy and = PyPy3. This test-suite can't test any implementation details or the internal = consistency of the implementation. But it can ensure that every API = exists and is functional to some degree. This way we'll get a sound = base-line to build compatible stackless implementations. As usual the project is hosted on Bitbucket: = https://bitbucket.org/stackless-dev/stackless-testsuite Currently it is in an very early state. For details see = https://bitbucket.org/stackless-dev/stackless-testsuite/issue/1. Cheers Anselm Am 19.07.2014 16:59, schrieb Christian Tismer: > Andrew, > > this was not a real request for help, but a reminder for us to work on Py= Py > compatibility issues, maybe at the EP2014 sprints. > > ciao - Chris > > > On 19.07.14 15:57, Andrew Francis wrote: >> Hi Ron: >> >> >> >> On Saturday, July 19, 2014 6:00 AM, "[email protected]" >> <[email protected]> wrote: >> >> Message: 1 >> Date: Fri, 18 Jul 2014 09:30:15 -0500 >> From: Rob Galanakis <[email protected] >> <mailto:[email protected]>> >> To: The Stackless Python Mailing List <[email protected] >> <mailto:[email protected]>> >> Subject: [Stackless] pypy stackless.py compatibility notes >> Message-ID: >> >> <CABaNhwr2fN9QhLz4YTnGMAg5nnf1vUQrNKwtUVVUB1qJ3CX3Mg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org >> <mailto:CABaNhwr2fN9QhLz4YTnGMAg5nnf1vUQrNKwtUVVUB1qJ3CX3Mg-JsoAwUIsXounXO2b/[email protected]= om>> >> Content-Type: text/plain; charset=3D"utf-8" >> >>> As requested by Herr Tismer on >>> https://github.com/rgalanakis/goless/issues/29 >> >>> stackless.py has no 'stackless.runcount' attribute (it has a >>> stackless.getruncount function). >> >>> I can swear there was some inconsistency between stackless.schedule() b= ut >>> cannot find exactly what it was right now. >> >> Off-hand, I think a major difference between Stackless Python and >> stackless.py stackless.schedule() is the internal representation of >> the runnable list. In Stackless, I recall it being a doubly-linked >> list accessible from the tasklet object. In stackless.py, it is a list. >> >> If I go through my notes, I can probably list other inconsistencies ..... >> >> As for deadlock detection (which the original poster was doing), I >> had some ideas on the subject: >> >> [Stackless] Requesting Comments for a Deadlock Detection Module >> <http://www.stackless.com/pipermail/stackless/2009-February/003956.html> >> >> >> Cheers, >> Andrew >> >> >> = >> >> = >> >> = >> >> = >> >> [Stackless] Requesting Comments for a Deadlock Detection Module >> <http://www.stackless.com/pipermail/stackless/2009-February/003956.html> >> [Stackless] Requesting Comments for a Deadlock Detection Module Andrew >> Francis andrewfr_ice at yahoo.com Fri Feb 27 20:28:14 CET 2009 >> >> View on www.stackless.com >> <http://www.stackless.com/pipermail/stackless/2009-February/003956.html> >> = >> Preview by Yahoo >> >> >> >> >> >> >> > > > > > Andrew, > > this was not a real request for help, but a reminder for us to work on Py= Py > compatibility issues, maybe at the EP2014 sprints. > > ciao - Chris > > > On 19.07.14 15:57, Andrew Francis wrote: >> Hi Ron: >> >> >> >> On Saturday, July 19, 2014 6:00 AM, "[email protected]" >> <[email protected]> wrote: >> >> Message: 1 >> Date: Fri, 18 Jul 2014 09:30:15 -0500 >> From: Rob Galanakis <[email protected] >> <mailto:[email protected]>> >> To: The Stackless Python Mailing List <[email protected] >> <mailto:[email protected]>> >> Subject: [Stackless] pypy stackless.py compatibility notes >> Message-ID: >> >> <CABaNhwr2fN9QhLz4YTnGMAg5nnf1vUQrNKwtUVVUB1qJ3CX3Mg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org >> <mailto:CABaNhwr2fN9QhLz4YTnGMAg5nnf1vUQrNKwtUVVUB1qJ3CX3Mg-JsoAwUIsXounXO2b/[email protected]= om>> >> Content-Type: text/plain; charset=3D"utf-8" >> >> >As requested by Herr Tismer on >> >https://github.com/rgalanakis/goless/issues/29 >> >> >stackless.py has no 'stackless.runcount' attribute (it has a >> >stackless.getruncount function). >> >> >I can swear there was some inconsistency between stackless.schedule() b= ut >> >cannot find exactly what it was right now. >> >> Off-hand, I think a major difference between Stackless Python and >> stackless.py stackless.schedule() is the internal representation of >> the runnable list. In Stackless, I recall it being a doubly-linked >> list accessible from the tasklet object. In stackless.py, it is a list. >> >> If I go through my notes, I can probably list other inconsistencies ..... >> >> As for deadlock detection (which the original poster was doing), I >> had some ideas on the subject: >> >> [Stackless] Requesting Comments for a Deadlock Detection Module >> <http://www.stackless.com/pipermail/stackless/2009-February/003956.html> >> >> >> Cheers, >> Andrew >> = >> = >> = >> = >> [Stackless] Requesting Comments for a Deadlock Detection Module >> <http://www.stackless.com/pipermail/stackless/2009-February/003956.html> >> [Stackless] Requesting Comments for a Deadlock Detection Module Andrew >> Francis andrewfr_ice at yahoo.com Fri Feb 27 20:28:14 CET 2009 >> >> View on www.stackless.com >> <http://www.stackless.com/pipermail/stackless/2009-February/003956.html> >> = >> Preview by Yahoo >> >> >> >> > > > -- > Christian Tismer :^)[email protected] > Software Consulting :http://www.stackless.com/ > Karl-Liebknecht-Str. 121 :http://www.pydica.net/ > 14482 Potsdam : GPG key -> 0xFB7BEE0E > phone +49 173 24 18 776 fax +49 (30) 700143-0023 > > > > _______________________________________________ > Stackless mailing list > [email protected] > http://www.stackless.com/mailman/listinfo/stackless > -- = Dipl. Phys. Anselm Kruis science + computing ag Senior Solution Architect Ingolst=E4dter Str. 22 email [email protected] 80807 M=FCnchen, Germany phone +49 89 356386 874 fax 737 www.science-computing.de -- = Vorstandsvorsitzender/Chairman of the board of management: Gerd-Lothar Leonhart Vorstand/Board of Management: Dr. Bernd Finkbeiner, Michael Heinrichs, Dr. Arno Steitz Vorsitzender des Aufsichtsrats/ Chairman of the Supervisory Board: Philippe Miltin Sitz/Registered Office: Tuebingen Registergericht/Registration Court: Stuttgart Registernummer/Commercial Register No.: HRB 382196