#20241: Separate Sage-specific components from generic C-interface in
PariInstance
-------------------------------------+-------------------------------------
Reporter: defeo | Owner:
Type: PLEASE CHANGE | Status: new
Priority: major | Milestone: sage-7.2
Component: interfaces | Resolution:
Keywords: | Merged in:
Authors: Luca De Feo | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
u/defeo/trac/u/defeo/pari_instance | 799f2555869086a4a302313545a9b74d8e5597e8
Dependencies: #20217 | Stopgaps:
-------------------------------------+-------------------------------------
Changes (by defeo):
* commit: => 799f2555869086a4a302313545a9b74d8e5597e8
* dependencies: => #20217
Comment:
I'm pushing a failed attempt at splitting `PariInstance` into a generic
`PariInstance` cython type, and a Sage-specific `SagePariInstance`.
This is problematic because of the global
`sage.libs.pari.pari_instance.pari` variable: it is not clear where this
variable should go.
Some concrete suggestions: `PariInstance` serves three main purposes:
- As a wrapper around the pari stack,
- As a collection of conversion methods to and from pari types,
- As a namespace for a handful of functions, for lack of a better place
(e.g. `primes`, `genus2_reduction`, ...)
Of these, only the first one really needs a Cython type. I suggest:
- To make a proper abstraction `PariStack` around the pari stack. This
should not be handled as a global variable, but rather as a singleton
class (we cannot have more than one pari stack, because of global
variables).
- Move conversion functions to a separate namespace, as #20226 started
doing.
- Have `PariInstance` hold a pointer to a `PariStack`. Have `gen`'s hold a
pointer to a `PariInstance`. (Maybe this double wrapping is superfluous?)
I don't know quite well what to do with utility functions such as
`primes`. They may stay as methods of `PariInstance`, or maybe go to a
different namespace as pure functions (they should take a `PariInstance`
or `PariStack` as parameter, then).
----
Last 10 new commits:
||[http://git.sagemath.org/sage.git/commit/?id=dce67fca49e515ce610539263e1e7a6f26c7bc69
dce67fc]||{{{Move memory functions to cysignals}}}||
||[http://git.sagemath.org/sage.git/commit/?id=4bb8337295ed82c93d1a9c9f7173a4c36f97151d
4bb8337]||{{{Rename sage_malloc -> sig_malloc and friends}}}||
||[http://git.sagemath.org/sage.git/commit/?id=5ab73c1c098478da766a061109a2a248d741e49e
5ab73c1]||{{{Get rid of factorint_withproof_sage in PARI interface}}}||
||[http://git.sagemath.org/sage.git/commit/?id=c0ed97a0b2a036c3a4677d8e5dbabf6b2a6d54fc
c0ed97a]||{{{Stop using deprecated PARI factoring features}}}||
||[http://git.sagemath.org/sage.git/commit/?id=edc5ce28655f157a13f27d2d94ae2e5d5d731bd6
edc5ce2]||{{{Merge branch
't/20205/get_rid_of_factorint_withproof_sage_in_pari_interface' into
HEAD}}}||
||[http://git.sagemath.org/sage.git/commit/?id=5fb408dedb22e4147ee68ae5656e643b4f1df06c
5fb408d]||{{{Replace pari_catch_sig_on by sig_on}}}||
||[http://git.sagemath.org/sage.git/commit/?id=d5c934c5164255e352cee9ddbc780d0759b6e185
d5c934c]||{{{Deprecate PARI nth_prime and prime_list}}}||
||[http://git.sagemath.org/sage.git/commit/?id=d7d2d7df24760cdecf0d6585972d2bf6140c1123
d7d2d7d]||{{{Remove redundant functions from pari_instance.pyx}}}||
||[http://git.sagemath.org/sage.git/commit/?id=282b1878dd25528debef017d13057b68c05b3b5b
282b187]||{{{Failed attempt to split PariInstance}}}||
||[http://git.sagemath.org/sage.git/commit/?id=799f2555869086a4a302313545a9b74d8e5597e8
799f255]||{{{Split pxd file too}}}||
--
Ticket URL: <http://trac.sagemath.org/ticket/20241#comment:2>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
