I posted some comments on github, I'll yet have to run the tests however.

Vinzent

2010/8/11 Christian Muise <christian.mu...@gmail.com>:
> Ok, so I went through the growing pains of figuring all this git stuff out,
> at we should be good to go with this branch:
> - http://github.com/haz/sympy/tree/soc-final-fixed
>   The commits are linear, all tests should be running clean (regardless of
> commit now, not just the head), and the trivial commits for fixes were
> squashed into the appropriate commits where the problem cropped up in the
> first place.
>   Merging the known fact forms will be post-SoC, and done with this issue:
> - http://code.google.com/p/sympy/issues/detail?id=2016
>   ...and getting the compiled forms to be changed online will be addressed
> with this issue:
> - http://code.google.com/p/sympy/issues/detail?id=2017
>   Please let me know what else I can / should fix up in the current branch
> for inclusion (that's my main goal for this final week).
>   On the topic of re-computing things online -- I should be able to detect
> and throw an error when an inconsistency is formed (since we have
> precompiled things already), and updating the known_facts_* will just be a
> form of belief revision...definitely doable. But it should be noted that
> these known facts are separate from default assumptions or assumption
> contexts. Both of which already exist. The default assumptions can go into
> global_assumptions (importable from sympy), and if you would rather use your
> own context, that's doable as well -- see this test case that demonstrates
> it:
> - http://github.com/haz/sympy/blob/soc-final-fixed/sympy/assumptions/tests/test_query.py#L947
>   The known facts are reserved for mathematical truths:
> - http://github.com/haz/sympy/blob/soc-final-fixed/sympy/assumptions/ask.py#L233
> ...and these aren't likely to change too much unless someone is introducing
> new predicates.
>   Thanks again for all the feedback everyone.
>   Cheers
>    Christian
>
> On Wed, Aug 11, 2010 at 1:14 PM, Aaron S. Meurer <asmeu...@gmail.com> wrote:
>>
>> On Aug 11, 2010, at 3:00 AM, Vinzent Steinberg wrote:
>>
>> > 2010/8/10 Aaron S. Meurer <asmeu...@gmail.com>:
>> >> Would it be possible to persistently cache the result, so that it's
>> >> computed once on an import, but then saved somewhere so that future 
>> >> imports
>> >> don't need to do it.  Maybe something that could be done by setup.py.
>> >
>> > I think this would be the cleanest solution, however we would need to
>> > introduce a writable directory for sympy. It's no problem if the first
>> > import of sympy takes 20 s (or every import where the facts were
>> > changed in the code), especially if this is done by setup.py, so the
>> > user won't even notice.
>> >
>> > The writable directory could also have the config file (another nice
>> > thing to introduce) and compiled code from codegen.
>> >
>> > Vinzent
>>
>> And files from preview().  And I suppose that if you configure it to not
>> be available, which might be necessary in certain environments, then it will
>> have to just do it at import time every time.
>>
>> Aaron Meurer
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "sympy-patches" group.
>> To post to this group, send email to sympy-patc...@googlegroups.com.
>> To unsubscribe from this group, send email to
>> sympy-patches+unsubscr...@googlegroups.com.
>> For more options, visit this group at
>> http://groups.google.com/group/sympy-patches?hl=en.
>>
>
> --
> You received this message because you are subscribed to the Google Groups
> "sympy-patches" group.
> To post to this group, send email to sympy-patc...@googlegroups.com.
> To unsubscribe from this group, send email to
> sympy-patches+unsubscr...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/sympy-patches?hl=en.
>

-- 
You received this message because you are subscribed to the Google Groups 
"sympy-patches" group.
To post to this group, send email to sympy-patc...@googlegroups.com.
To unsubscribe from this group, send email to 
sympy-patches+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sympy-patches?hl=en.

Reply via email to