keyRel2

[tom-wa]

Key relation types: Below, DirectBelow, Silbling, and Nephew
DirectBelow and Silbling are boolean tests
relation suffixes:
SN: the namespaces have to match
IN: ignoring namespaces completely
CN: cascading namespace matches all namespaces

currently returns:
-1 for invalid keys / usage error
0 for failed tests
1 if a boolean relation test succeeded

0 if non-boolean relation tests succeeded - the value is the number of levels the check key is below

[markus2330]

should be in proposal first. And please suffix the names.

[tom-wa]

just wanted to know if it's basically ok (return values/relationship checks/etc) before i finish it up
any suggestions for other checks ?

[markus2330]

Its always good if you write what you want to know directly in the PR description :-) Again, its hard to say what you wanted to achieve. Please put proper API docu in the source.

The description of the return values seem to be flawed, -1 is used for two different purposes. Why do you think is the isSame test so important? I would rather go for -1: usage error, 0: failed, 1 succeeded (or even >1 for numbers and not only boolean tests)

And I need to get used to the enum values... Maybe a bit longer names would be better.

(comment was edited multiple times)

[markus2330]

I would avoid this, users can check with keyCmp themselves.

[tom-wa]

you are right about the return values. not even sure if checking if the keys are the same is needed here.
and 0 for failed check and dropping the -1 would better anyway i guess.

[tom-wa]

where should i put the testcases for keyRel2 ?

[tom-wa]

nvm, just found test_proposal.c

i've updated it now, and used longer names for the enum values

[markus2330]

please run code formatter

[tom-wa]

@markus2330 done

[markus2330]

Thank you, well done!