#10187: Update ecl and maxima
-------------------------------------------------------------------------------+
Reporter: vbraun
| Owner: tbd
Type: defect
| Status: needs_work
Priority: major
| Milestone: sage-4.6.1
Component: packages
| Keywords:
Author: Volker Braun, David Kirkby
| Upstream: Workaround found; Bug reported upstream.
Reviewer: Karl-Dieter Crisman, David Kirkby, Volker Braun, Leif Leonhardy
| Merged:
Work_issues:
|
-------------------------------------------------------------------------------+
Changes (by drkirkby):
* status: needs_review => needs_work
Comment:
Some comments, having spent some time looking at the code others have
written.
=== My ECL package ===
* I don't see anything wrong with the way CXXFLAGS/CFLAGS is handled in
the ECL package. It looks OK to me. It might not be 100% necessary, but in
general we should set these flags and hope the package follows this.
----
=== Looking at the doc tests: ===
* '''trac_10187_general_display_prefix_workaround.patch''' looks OK to
me. So I'll positively review that part.
* '''trac_10187_mark_some_doctests_random_until_9880_gets_merged.patch'''
looks OK to me, So I'll positively review that part.
* '''maxima-5.22.1.patch''' SPKG.txt looks OK, but you will need to find
someone that know more about Lisp than me to tell you if that's correct.
Ask on sage-devel to get someone who understands it, since I don't.
* '''trac_10187_fix_easy_doctests.patch'''
1) There are these numerical results. Has anyone independently verified
this are correct?
{{{
[2.6789385347..., -8.3905259853..., 26.662447494..., -80.683148377...]
}}}
On what basis was the number of significant digits selected?
As a matter of interest, do we know if anyone every verify the original
horrid looking symbolic integral is correct?
I have a dislike for tests where is appears we don't verify the result is
correct, but assume it must be. Please justify why those results are
correct. That should be in a comment in the code, not displayed.
2) Why was the doctest:
{{{
integral(abs(x), x, 0, a)
}}}
changed to:
{{{
integral(abs(x)*x, x, 0, a)
}}}
It rather strikes me that we should show the example of how to do this
calculation in the way Maxima now wants, rather than to simply change the
integral to one which gives a simpler answer.
It so happens that Mathematica 7.0.1 can do the first inegral:
{{{
In[7]:= Integrate[Abs[x],{x,0,a}]
a Abs[a]
Out[7]= --------
2
}}}
I think it would be better to show the original integral, show the
failure. We could even point out how to do the integral in Maxima, and get
the hopefully correct result.
Is rather strikes me that inputting something that was quite reasonable
looking does not work. So we should explain why that reasonable thing is
wrong, rather than just pick another example.
The Mathematica documentation often shows a section marked '''Possible
issues''' and would explain problems one might get. Here it seems we
should do this, rather than just pick another integral.
'''sage/plot/plot3d/transform.pyx''' Do we know if that horrible looking
result is correct? I see it is a simple change from the previous horrible
looking result, but how do we know if the previous one is correct? If, as
I suspect, this has not been verified by hand or with another package,
then we should write that fact.
== Summary ==
* The portability issues appears to have been resolved.
* Someone needs to check my ECL package
* '''trac_10187_general_display_prefix_workaround.patch''' is OK
* '''trac_10187_mark_some_doctests_random_until_9880_gets_merged.patch'''
is OK
* '''trac_10187_fix_easy_doctests.patch''' Needs explanation.
* '''maxima-5.22.1.patch''' Needs someone to review with a skill set
different to me.
I'm changing to needs work, as some of these tests changes needs
explanation.
Dave
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/10187#comment:99>
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 post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en.
