Nodes containing /
#169
Comments
|
Yeah ... that can ruin some internals ... especially UTs ...
The first link is a bit incorrect. Not all. All but the control chars. They must be escaped. Meaning that they will not present in a string. So the first idea is to use tab. But it will ruin all UTs and actually all existing addressing. Maybe in opposite, the |
|
You are right, the |
|
That means unsupporting a valid JSON ... not ideal. I'm thinking on autoreplacing |
I added tests for handling of special characters - like slashes, quotes, and unicode - in element names, paths, and values. Everything - except of one case - was working but now we know for sure.
The special case is a slash in an element name that has sub-objects: `"contains/slash": "test"` is ok, but `"contains/slash": { "test": 1 }` is not. As suggested in sbcgua#169, I replaced the slash with a `tab`. Parsing now works well.
For methods that require a path as input, we could either leave it as is and remember to use a tab there, too, or we pre-process the input.
```
{ "contains/slash": { "test": 1 } }
r->get( '/contains/slash/test' ). " never worked
r->get( |/contains\tslash/test| ). " works with this PR
r->get( '/contains\/slash/test' ). " could work if the parameter is preprocessed i.e. replace \/ with \t
```
I would go for the middle case and not worry about the parameters.
Closes sbcgua#169.
I'm trying to parse https://raw.githubusercontent.com/abapedia/steampunk-2305-api/main/src/_status.json with AJSON.
{ "clas,/dmo/cx_rap_generator": { "status": "RELEASED", "successors": [] }, "clas,/iwbep/cl_cp_factory_remote": { "status": "RELEASED", "successors": [] }, "clas,/iwbep/cl_cp_factory_unit_tst": { "status": "RELEASED", "successors": [] },There are no errors but the result is incorrect (the path column):
Clearly the
/in the name confuses AJSON and should be escaped. Not sure how you want to approach a fix here.The text was updated successfully, but these errors were encountered: