Clean up legacy in Job class schema

[egede]

The Job class has some very old legacy fields in the schema. The fields inputsandbox, outputsandbox and outputdata can be removed.

The rest of the code base should be checked and places where these are used should be removed as it all deals with legacy ways of specifying a job.

The ForbidLegacyOutput config option should be removed as well. This last bit is related to #2109.

[vedanshbhartia]

Hi, I was looking into this issue. Should the ForbidLegacyInput Option be removed as well as part of this change?

[egede]

Yes, it should indeed. Make sure to keep the inputdata field in the schema though.

[Tanuj-rai]

I am working on this issue, but I would like to know what I should write in place of the field that is being removed. In conditional statements fields are used so how should I remove them?

[egede]

I am working on this issue, but I would like to know what I should write in place of the field that is being removed. In conditional statements fields are used so how should I remove them?

I believe that @vedanshbhartia is already working on this issue. As a reply to your question, the ForbidLegacyOutput output option can be regarded as always being True, so conditional statements should be removed to reflect this.

[vedanshbhartia]

Some components (such as the preparejob funtion in ganga/GangaCore/Lib/LCG/LCG.py) contain input_sandbox and output_sandbox fields. However, their usage is not wrapped around a check for ForbidLegacy[Input/Output]. Are these variables different from the sandboxes used in Job.py?

[egede]

That rather reflects that the complete LCG backend should be removed. Might be best to do that in a separate PR first in fact.