This page is out of date, for information on the coercion model, please see the main coercion page.

Merge up at (pull on top of 3.0.3 release candidate) Should build and start without problems.

Hint: If something doesn't work, try typing

sage: coercion_traceback()

which will give a listing of all errors caught in the coercion model.



In Progress

To Do

This should fail: 
            sage: x, y = var('x,y')
            sage: parent(ZZ[x][y](1)*vector(QQ[y],[1,2]))
            sage: parent(ZZ[x](1)*vector(QQ[y],[1,2]))

Files to fix (Make all parents conform to the spec)

Useful script to list all parents in a given module that still need work: list_parents.sage

Nothing to do


In progress

Little to do

Lots to do

Base classes

Coercion doctesting

Comment from William:

teragon:sage was$ sage -coverage categories/action.pyx
structure/element.pyx structure/coerce.pyx |grep SCORE
SCORE categories/action.pyx: 0% (0 of 22)
SCORE structure/element.pyx: 18% (28 of 153)
SCORE structure/coerce.pyx: 2% (1 of 39)

I tried to understand and use the coercion model code to trac down the issue with #2079 and
was amazed at how nonexistent the doctesting and documentation of
functions is there.
In order for people to write lots of coercion code all of Sage, it's critical that they can
read the coercion model code so they can track down -- for themselves -- what is going
wrong when they run into trouble.  I would put getting the coverage of the above files
(and whatever else is related to coercion) up to 100% as the first step in your coercion
model stuff.   Seriously.   And don't say it can't be doctested, since even though there are
a lot of things not easily accessible now from the interpreter, such as the coercion model

Categories to implement

Old notes

To start, get a copy of Sage 2.10.1, install the latest cython spkg (available at, pull from and build.

Import fix-random.patch to sage-scripts to ignore random tests.

days7/coercion/todo (last edited 2009-07-29 19:16:15 by robertwb)