This is pretty much a complete overhaul of the step parsing process. We now leave it to the matching tree search to scan a set of step tokens for either words or arguments, depending on what a given part is interested in.
This keeps it flexible and extremely tolerant of people using the word don't, for example, in a step reference. That works fine provided the definition doesn't expect an argument at that point, at which point it will expect an argument.
Some extra work will be needed to allow argument validation messages to be generated inside ArgumentPart when an exact match has happened.
alistairjevans
commented
4 years ago
This is pretty much a complete overhaul of the step parsing process. We now leave it to the matching tree search to scan a set of step tokens for either words or arguments, depending on what a given part is interested in.
This keeps it flexible and extremely tolerant of people using the word
don't, for example, in a step reference. That works fine provided the definition doesn't expect an argument at that point, at which point it will expect an argument.Some extra work will be needed to allow argument validation messages to be generated inside
ArgumentPartwhen an exact match has happened.