Complete implementation of JSON Pointer and better errors. Fixes #94. #95
Conversation
Replaced `#toString()` with `#asText()` on `PathUtils#getPath()`.
| @@ -11,10 +11,10 @@ | |||
| "node": { "x": "a" } | |||
| }, | |||
| { | |||
| "op": [{ "op": "replace", "path": "non-existing-path", "value": "some-value"}], | |||
| "op": [{ "op": "replace", "path": "/non-existing-path", "value": "some-value"}], | |||
There was a problem hiding this comment.
The path in the original test was invalid per RFC 6901
|
First of all, you did a trully wonderful job there. The code looks a lot more cleaner and the "Path" logic is all centralized now. I took a quick look and everything looks fine, I will review it more carefully now, this is quite a big change indeed, so I think it will take a while to review it all, but I can do it before the weekend. |
|
Well I went back to the parser and looked at /** ...
* {@code StringTokenizer} is a legacy class that is retained for
* compatibility reasons although its use is discouraged in new code. It is
* recommended that anyone seeking this functionality use the {@code split}
* method of {@code String} or the java.util.regex package instead.
*/If anyone wants to rewrite this in what they consider a simpler fashion, go for it. This isn't a really good parser anyway, there's absolutely no validation (Unicode character ranges etc.). |
|
I am really really sorry @holograph , I meant |
|
Ah, that amounts to writing my own parser, which I'm too lazy to do without good reason :-) |
|
I too came back to write the parser and I got a question: The current implementation parses any "wrong" input as a empty/Root JsonPointer, by wrong I mean inputs without |
|
That's what I get for being lazy. Fixed. @vishwakarma Ball's in your court :-) |
Yup, this is a pretty big change, following up on @LouizFC's work on #90. It fixes #94 and generally clears up a lot of crufty code, although it doesn't yet add the compatibility flag discussed in #90.
What started as a small fix ended up being a comprehensive refactoring, so I apologize in advance for the size of this PR. A few important notes for reviewers:
Let me know if you need any help, or alternatively have some ideas on how to break this down.