Use mathics scanner release

[rocky]

The PossibleZeroQ failure I think was a break from before and another branch may fix.

The RecursionLimit failure I'll investigate later.

[mmatera]

• I fixed the PossibleZeroQ in Adding an option for use Asymptote to build svg graphics #1145
• I restore SingleLine.send_messages in method in mathics-scanner.feeder, get the CI passing again.

[rocky]

• I fixed the PossibleZeroQ in Adding an option for use Asymptote to build svg graphics #1145

Would you please pull out the PossibleZeroQ and put that in its own PR. We'll merge that soon then

• I restore SingleLine.send_messages in method in mathics-scanner.feeder, get the CI passing again.

This PR addresses that. So that change can be reverted. Thanks.

I am merging this in now. Feel free to keep commenting on it or improve it though.

Merging in will allow mathicsscript and mathics-django to move closer to a release.

[mmatera]

• I fixed the PossibleZeroQ in Adding an option for use Asymptote to build svg graphics #1145

Would you please pull out the PossibleZeroQ and put that in its own PR. We'll merge that soon then

See PR #1147

• I restore SingleLine.send_messages in method in mathics-scanner.feeder, get the CI passing again.

This PR addresses that. So that change can be reverted. Thanks.
Maybe now the simple way would be to make another PR removing the method in mathics-scanner. On the other hand, wouldn't be a good idea to keep this method by now, for backward compatibility?

I am merging this in now. Feel free to keep commenting on it or improve it though.

Merging in will allow mathicsscript and mathics-django to move closer to a release.

[rocky]

See comment on that PR.

[mmatera]

regarding send_messages in mathics-scanner, you did this to avoid introduce (implicit) references to mathics in mathics-scanner?

[rocky]

Right - this was something that @GarkGarcia originally noticed.

My goal for this weekend though is to get out a release of whatever it is we can. If I had to do again I probably would have totally left in with the Mathics dependencies (as you report doing) and fixed up in a later release.

However since I went down this path, best not to go backwards, but instead go forward it in the places that use mathics scanner and not deal with backwards compatibilty of something that was never released.