Mark Conway

This is pretty close to what I have come up with myself. I went about the problem from two sides, one being to extend jsmn_t, the second adding an extra...

I also like leaving the combinations (minus JSMN_ALL_TYPE) implicit, but I could also extend jsmntype_t to include the explicit classes you have, though I would suggest letting bitwise logic set...

Let me finish up cleaning up a few bits so all of the previous tests pass again and I will at least put in a link to my fork. I...

@dominickpastore I'm forked over at https://github.com/themobiusproject/jsmn. I am going to start moving pieces to their own branches to try some pull requests. master, right now, only has two tests left...

I was looking at it previously because I like what I have come up with and would be happy to share it with others, especially through the official repo. Since...

@dominickpastore @pt300 I am still hanging around here as well. Both @dominickpastore's PR #194, #195, and #196 and my PR #197 are all valid options. I had once tried to...

> In fact, those two relaxations are all I would provide in non-strict mode. To be more specific: > * Allow primitives to be made of any character that doesn't...

@atomsnc - Sorry, I saw your post and then life hit me for a bit. Very good catch. I don't know how I wasn't testing for that case previously. I...

@pt300 I appreciate the comments. > A massive pr It certainly is. I swiped the next_sibling idea from #68 and shorter tokens and unsigned ints from #93 years ago when...

@pt300 > After the last sweep through your code it's been mostly few smaller things like the use of defined() in preprocessor macros instead of #ifdef and #ifndef. There was...