Bug#983328: scipy 1.6.1 breaks sparse (COO) matrix construction

[Drew Parsons]

On 2021-03-02 20:19, Rebecca N. Palmer wrote:
> What, if any, action are you proposing before bullseye?  Or is this a
> request for an up-to-date pandas in experimental?
> 
> That workaround just disables the affected test, i.e. it will probably
> fix the FTBFS but do nothing to the underlying problem.  As noted in
> #983404, a real scipy fix also exists
> (https://github.com/scipy/scipy/pull/13586).  I have not tested either
> of them.
> 
> pandas and scipy are both key, so from tonight they need release team
> approval for any changes.



I think it's too dangerous to release scipy 1.6.1, even with its patch.

In that case there's nothing that needs to be done at the moment with 
pandas.  I posted up the pandas workaround in this bug report just to 
keep the record of what's going on and how the upstreams are addressing 
the problem.

Hopefully scipy 1.6.2 is released by the time we get out of freeze and 
need to update unstable. In any case the future pandas releases will 
contain the test workaround.

Drew