[issue10399] AST Optimization: inlining of function calls

Change by Paulo Henrique Silva : -- nosy: +phsilva ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue10399] AST Optimization: inlining of function calls

Change by Batuhan Taskaya : -- nosy: +BTaskaya ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue10399] AST Optimization: inlining of function calls

Change by Yonatan Goldschmidt : -- nosy: +Yonatan Goldschmidt ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue10399] AST Optimization: inlining of function calls

Changes by Mateusz Bysiek : -- nosy: +mbdevpl ___ Python tracker ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Changes by Mark Lawrence : -- nosy: -BreamoreBoy ___ Python tracker ___ ___

[issue10399] AST Optimization: inlining of function calls

Changes by irdb : -- nosy: +irdb ___ Python tracker ___ ___ Python-bugs-list mailing

[issue10399] AST Optimization: inlining of function calls

STINNER Victor added the comment: "Can we resurrect this, perhaps by taking it up on python-dev?" I created a new "FAT Python" project to reimplement such kind of optimizations with a similar design (similar but different ;-)): https://faster-cpython.readthedocs.org/fat_python.html --

[issue10399] AST Optimization: inlining of function calls

Mark Lawrence added the comment: I get the impression that there's a lot to be gained here, especially when compared to some of the micro-optimizations that are committed. Can we resurrect this, perhaps by taking it up on python-dev? Also msg121082 refers to a Globals cache patch posted on

[issue10399] AST Optimization: inlining of function calls

Changes by Brett Cannon br...@python.org: -- nosy: -brett.cannon ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Changes by Nick Coghlan ncogh...@gmail.com: -- nosy: +ncoghlan ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Changes by Eric Snow ericsnowcurren...@gmail.com: -- nosy: +eric.snow ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Changes by Giampaolo Rodola' g.rod...@gmail.com: -- nosy: +giampaolo.rodola ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___

[issue10399] AST Optimization: inlining of function calls

Changes by Gregory P. Smith g...@krypto.org: -- nosy: +gregory.p.smith ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___

[issue10399] AST Optimization: inlining of function calls

Changes by Meador Inge mead...@gmail.com: -- nosy: +meador.inge ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Changes by John O'Connor tehj...@gmail.com: -- nosy: +jcon ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list mailing

[issue10399] AST Optimization: inlining of function calls

Changes by Pas pasthe...@gmail.com: -- nosy: +pas ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list mailing list

[issue10399] AST Optimization: inlining of function calls

Changes by Armin Rigo ar...@users.sourceforge.net: -- nosy: -arigo ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Changes by Ryan Kelly r...@rfk.id.au: -- nosy: +rfk ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list mailing list

[issue10399] AST Optimization: inlining of function calls

Changes by Dave Malcolm dmalc...@redhat.com: -- assignee: - dmalcolm ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Armin Rigo ar...@users.sourceforge.net added the comment: But this seems to me like a contrived example: how often in real code do people pass around these builtins, rather than calling them directly? From experience developing PyPy, every argument that goes this theoretically breaks

[issue10399] AST Optimization: inlining of function calls

Dave Malcolm dmalc...@redhat.com added the comment: If this a work in progress, you could create an SVN branch in the sandbox (you can then use svnmerge to avoid diverging too much from mainline) or an hg repo. Good idea; I've created branch dmalcolm-ast-optimization-branch (of py3k) on

[issue10399] AST Optimization: inlining of function calls

Dave Malcolm dmalc...@redhat.com added the comment: py3k-ast-pyoptimize-2010-11-19-006.patch fixed up and committed to the branch as r86715; I'll work on that branch for the time being. -- ___ Python tracker rep...@bugs.python.org

[issue10399] AST Optimization: inlining of function calls

Alex alex.gay...@gmail.com added the comment: There are a couple places you mention not doing the optimization when specific functions are used (e.g. dir, globals, locals), how exactly do you verify that, given those functions could be bound to any name? --

[issue10399] AST Optimization: inlining of function calls

Dave Malcolm dmalc...@redhat.com added the comment: Thanks for reading through this. There are a couple places you mention not doing the optimization when specific functions are used (e.g. dir, globals, locals), how exactly do you verify that, given those functions could be bound to any

[issue10399] AST Optimization: inlining of function calls

Neal Norwitz nnorw...@gmail.com added the comment: There is some precedent for allowing minor differences in -O mode. http://mail.python.org/pipermail/python-dev/2008-February/077193.html I think we should push this to see how practical we can make it. Dave, this is great work, I'm really

[issue10399] AST Optimization: inlining of function calls

Dave Malcolm dmalc...@redhat.com added the comment: Third, for that Graphviz output, was anything special required? If so, I would toss the code into Tools for others to benefit from. It's merely the to_dot function from Lib/__optimizer__.py (which turns an AST into .dot source code),

[issue10399] AST Optimization: inlining of function calls

Brett Cannon br...@python.org added the comment: No, it's rather Linux and tool specific to go into ast.py. But adding it to the Tools/ directory makes sense. -- ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399

[issue10399] AST Optimization: inlining of function calls

Dave Malcolm dmalc...@redhat.com added the comment: Sorry again for another epic-length comment... I'm attaching the latest work-in-progress on this. The code is still fairly messy (embarrasingly so in places), but it's better to have it out in public in this tracker than stuck on my hard

[issue10399] AST Optimization: inlining of function calls

Antoine Pitrou pit...@free.fr added the comment: Sorry again for another epic-length comment... I'm attaching the latest work-in-progress on this. If this a work in progress, you could create an SVN branch in the sandbox (you can then use svnmerge to avoid diverging too much from mainline)

[issue10399] AST Optimization: inlining of function calls

Changes by Senthil Kumaran orsent...@gmail.com: -- nosy: +orsenthil ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___ ___ Python-bugs-list

[issue10399] AST Optimization: inlining of function calls

Brett Cannon br...@python.org added the comment: While I have nothing to say directly about the inline optimization, I do have some stuff to say about moving to AST optimizations. First, doing in Python is a good thing. It not only makes prototyping easier, but it allows other VMs to use the

[issue10399] AST Optimization: inlining of function calls

New submission from Dave Malcolm dmalc...@redhat.com: In msg#120541 of issue#1346238 Raymond suggested to aim high, so here goes... I'm opening this as a separate bug as it's a very different approach to the patches in that bug; adding those on the nosy list for that bug. Sorry in advance

[issue10399] AST Optimization: inlining of function calls

Changes by Dave Malcolm dmalc...@redhat.com: Added file: http://bugs.python.org/file19584/before.png ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___

[issue10399] AST Optimization: inlining of function calls

Changes by Dave Malcolm dmalc...@redhat.com: Added file: http://bugs.python.org/file19585/after.png ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___

[issue10399] AST Optimization: inlining of function calls

Antoine Pitrou pit...@free.fr added the comment: iAs I understand it, functions are name-resolved before the arguments are evaluated,/i There is no difference between resolution of function names and of other names. For example, global names (LOAD_GLOBAL) are resolved entirely at runtime

[issue10399] AST Optimization: inlining of function calls

Antoine Pitrou pit...@free.fr added the comment: Globals cache patch posted on issue10401. -- ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue10399 ___

[issue10399] AST Optimization: inlining of function calls

Alex alex.gay...@gmail.com added the comment: Just a thought: it's possible to cover all the cases properly with inlining (locals, globals (even with cross module inlining), tracebacks, sys._getframe(), etc.), however I think you're going to find that manually managing all of those is going