-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[WIP] Updating to On Demand #1
Conversation
Also, I am not sure on how to run the tests. |
Ping @lemire |
@NicolasJiaxin I had to approve the tests, it seems, but now they run (with failures). |
Ok thank you! At least now I can see what I have to fix |
I think that a reasonable default is just to parse everything as 'double' which is what you would expect from a JSON output. |
But that is maybe something we can smooth over. |
Though for now, I suggest you just treat all numbers as 'double'. |
I will try that. |
Interesting. It appears that I have to approve each run. Hmmm... :-) |
This will probably not work. It need more cleaning.. |
I will push here when I think I have a good version. |
It is fine. I do not mind approving the tests... I just do not understand why it does that. |
@NicolasJiaxin I gave you access to this repo., so maybe the tests will run automatically? |
I think I am going to start over from a fresh branch. I think I have deleted too much stuff that it is causing me problem.. Should not take too long now that I have a clear idea of what to do. |
Great! |
Closing. |
I have begun to update RcppSimdJson to On Demand. The main complication that I have seen is that On Demand does not have specific JSON type for numbers. It labels double, signed integers and unsigned integers as
json_type::number
. To work around this, I have created a enum classcomplete_json_type
that can identify the type of ajson_type::number
. This might not be an optimal solution, but it can serve as a prototype. I have not yet worked on matrix and data frame deserialization, only on scalar and vector deserialization, but there might still be typos and broken things in those files.