Fix using default bounds for TPN's generated with HTN method

[tmarble]

Consistently throw exceptions in HTN generation using fatal-error Closes #98 Closes #108

Signed-off-by: Tom Marble tmarble@info9.net

[pmdoll]

In test/pamela/HTN/coverage.main.tpn.edn, line 225 changed from #{:costle-87 :tc-86 :rewardge-88} to #{:rewardge-89 :costle-88 :tc-87} . What caused the sort order to change? For consistency, I think edn should be sorted just before writing to the file. Is it not the case?

[tmarble]

Clearly the UID's are offset by one as there was probably a new constraint which was created since the previous rubric.

The EDN is sorted insomuch as the keys of maps are in order. However the values are not sorted. In particular sets do not have an order. If we start to have failures due simply to this class of problems we can re-examine the merits of doing literal file contents comparison.

[pmdoll]

Motivation behind literal file content comparison is to ensure and verify that code changes do not inadvertently introduce an unexpected change in produced artifacts which could have ripple effects.

For this PR, I do expect minor changes in UIDs but was wondering what caused the set values to be printed in different order than before. So it seems like it is definitely a change in code that caused the insertion order to be different.

[dcerys]

As mentioned above, sets, by default are unsorted. If the order was consistent beforehand, we were just lucky. Although we could address this by using sorted-set in cases like this, that seems like a slippery slope of code obfuscation. If necessary, I'd rather have the equivalence checker do the work.

[tmarble]

For reference:

pamela.cli> (= {:a 123 :b #{:third :first :second}} {:b #{:first :second :third} :a 123})
true
pamela.cli> (= "{:a 123 :b #{:third :first :second}}" "{:b #{:first :second :third} :a 123}")
false
pamela.cli> 

Our current checks are equivalent to the second kind of comparison.

[dcerys]

Since it doesn't fully close #98, I'll reopen it after closing this PR