of::urn user template deduction SWIG syntax error (ofxLua)

[danomatika]

From checking ofxLua with the of_v20250319_osx_release nightly:

In ofUtils.h lines 455 - 456, the of::urn user template deduction "->" breaks SWIG completely with a syntax error. I have to comment out those two lines in order to generate the Lua bindings successfully. There is support for C++14 and C++17, but I can't find info that SWIG currently handles this part of C++17 (yet).

Can of::urn be moved to a separate header which is included by ofUtils.h, in which case SWIG would not have to parse that section since it only parses headers given to it explicitly.

My only other option is to maintain a copy of basically the entire header myself without the breaking parts to feed SWIG. Obviously this is not ideal.

[artificiel]

@danomatika yes ofUrn can be moved out, at the moment it is not required by anything core (it was part of the brainstorm/experiments that was triggered by ofShuffle being invalidated which led to the new random stuff). so moving it out (or commenting the problematic lines) is OK as i presume nobody is using it yet and the template deductions are also experimental in facilitating the use by requiring less information from the user. it may also be a freestanding ofUrn.h which is not included by ofUtils and only on demand (which is not a bad pattern per se).

[danomatika]

Ok, this seems like a feature more for a 0.13.0 release, however I assume things got a bit mixed in after 0.12. I suppose another option is an experimental define etc but I think that's overkill.

It can still be included in ofUtils.h, just being in a separate header solves my issue.

[danomatika]

As an aside, I was thinking of::urn didn't look quite finalized as the snake case naming on functions and variables doesn't really match the overall OF camel case usage. If it's released, there is no easy way to change it for people in the future...

[ofTheo]

@artificiel want to do a PR to either comment out or remove it for now?

[artificiel]

not near a dev env right now nor for the next 48h — I guess I can cut it out from Github's editor, but maybe better for someone to do it while being able to test it?

also #8182 should probably be merged before removing urn as it updates it, and the rest of the touch-ups are important to of::random (and it won't be a clean merge if urn is removed before).

[ofTheo]

@artificiel - maybe when you are back at the computer you could update #8182 with the of::urn removal and then we can just merge that in?

[artificiel]

OK!

[ofTheo]

closed by #8182