Add __str__ method for SolverVarFormBlock

[Ig-dolci]

Description

Adding __str__ method in the SolverVarFormBlock.

The __str__ is useful to visualise the tape DAG.

The ufl2unicode(self.lhs) is going to fail raised by a ufl bug. I have opened a UFL PR to fix it.

Obs:
Thank you Connor for helping me with the UFL fail.:)

Addressed.

[connorjward]

I will mark this PR as draft for now since we want to land the UFL PR first.

[connorjward]

We need to update our UFL fork now. This might be complicated since there is already a longstanding PR to do this. @nbouziani what is the status of that?

[nbouziani]

We need to update our UFL fork now. This might be complicated since there is already a longstanding PR to do this. @nbouziani what is the status of that?

That PR was for the dual changes and is out-of-date given that we merged UFL upstream after that. However, I recently noticed that the recent UFL changes with the new finite element interface have caused some of our tests to fail. There seem to be minor things to change first in firedrake and tsfc to update our UFL fork successfully.

[connorjward]

We need to update our UFL fork now. This might be complicated since there is already a longstanding PR to do this. @nbouziani what is the status of that?

That PR was for the dual changes and is out-of-date given that we merged UFL upstream after that. However, I recently noticed that the recent UFL changes with the new finite element interface have caused some of our tests to fail. There seem to be minor things to change first in firedrake and tsfc to update our UFL fork successfully.

Is someone willing to take that on? @nbouziani @Ig-dolci?

[Ig-dolci]

We need to update our UFL fork now. This might be complicated since there is already a longstanding PR to do this. @nbouziani what is the status of that?

That PR was for the dual changes and is out-of-date given that we merged UFL upstream after that. However, I recently noticed that the recent UFL changes with the new finite element interface have caused some of our tests to fail. There seem to be minor things to change first in firedrake and tsfc to update our UFL fork successfully.

Is someone willing to take that on? @nbouziani @Ig-dolci?

I will work on this today.

[Ig-dolci]

We need to update our UFL fork now. This might be complicated since there is already a longstanding PR to do this. @nbouziani what is the status of that?

That PR was for the dual changes and is out-of-date given that we merged UFL upstream after that. However, I recently noticed that the recent UFL changes with the new finite element interface have caused some of our tests to fail. There seem to be minor things to change first in firedrake and tsfc to update our UFL fork successfully.

Is someone willing to take that on? @nbouziani @Ig-dolci?

I will work on this today.

I just figured out that these PRs 3166 and 302 are aiming to update firedrake and tsfc for the latest ufl (upstream) interface.

[dham]

Need to cast the rhs to ufl.

[Ig-dolci]

This PR is incorrect. I did not realise that super has the __str__ method. I will close this PR, but I opened the PR 3222 aiming to fix the error for the cases in which rhs is a Cofunction type.