Based on #18, will rebase after that's merged; until then just look at latest commit.
This still needs docs, some tests, and some better failsafes, but I've tested it manually and it seems to work great. A bit WIP, but the core of it, including approach for dealing with non-json-serializable values and deeply nested tables, is done; open to comments on implementation strategy/conf options/etc.
To deal with non-json-serializable values, I made a json_prep_value helper which takes a lua value and makes it safe to pass to cjson.encode by either tostringing or dropping pieces as necessary to comply with cjson's specified requirements, mainly:
JSON object keys are always strings. Hence cjson.encode only supports table keys which are type number or string. All other types will generate an error.
and
The remaining Lua types will generate an error:
function
lightuserdata (non-NULL values)
thread
userdata
There are still a couple cases missing that I need to add, mainly NaN/Infinity, and I also want to have a better failsafe for any general unexpected case where cjson.encode fails.
Based on #18, will rebase after that's merged; until then just look at latest commit.
This still needs docs, some tests, and some better failsafes, but I've tested it manually and it seems to work great. A bit WIP, but the core of it, including approach for dealing with non-json-serializable values and deeply nested tables, is done; open to comments on implementation strategy/conf options/etc.
To deal with non-json-serializable values, I made a
json_prep_valuehelper which takes a lua value and makes it safe to pass tocjson.encodeby either tostringing or dropping pieces as necessary to comply with cjson's specified requirements, mainly:and
There are still a couple cases missing that I need to add, mainly NaN/Infinity, and I also want to have a better failsafe for any general unexpected case where
cjson.encodefails.