- updates large file upload task spec java examples

[baywet]

This pull request cleans up some of the language in the requirement, markdown formatting and updates the java samples.

Items left to discuss:

• Should we align the object model naming in the spec? (currently ChunkedUploadProvider is not really informative for Java)
• Should the task try to get the result (issuing a sub-query) after the upload completes on an empty response with a location header pointing to a resource on graph?
• Alternatively, should the task return a wrapper object instead of the direct type, so we could provide the type whenever available, or response information when not available?

We can have the discussion here before we merge that pull request. This way I can make subsequent updates before we merge this PR.

[baywet]

related work microsoftgraph/msgraph-sdk-javascript#392

[andrueastman]

With regards to the first query. I agree we should try to align the object modelling where we can.

[baywet]

@andrueastman @nikithauc What about aligning on LargeFileUploadXXX (Provider, result, etc...) ?
Another subsequent question I forgot to ask. What about guiding the implementers in terms of namespace naming. Currently in Java it lives in "concurrency" and without looking in the Microsoft Graph Docs or knowing this before hand, I'd never think it's in there.

[andrueastman]

With regards to namespace naming, I think it best lives in the same place as the PageIterator which is also a Task.

[baywet]

which we don't have in Java. So the guidance would be a tasks sub namespace under the root one then?

[baywet]

I just pushed another commit making the changes we've been talking about. Let me know what you think.

[nikithauc]

There is no create() function in LargeFileUpload task. It is only in OneDriveFileUpload. Please undo this for the time being. I will update this.

[baywet]

The goal of this document is to describe the future state of the solution, not the current one. I'd like this spec to be as final as possible so we each have clear guidelines on what changes need to be implemented in each SDK. The Java samples also do not map to current state here, this is expected.

[nikithauc]

LargeFileUploadTask.createUploadSession(client: Client, requestUrl, payload, headers) this is what is present.

[baywet]

which is the equivalent of graphClient.me().drive().item("blah.blob").createUploadSession() and is present as a static method because the JS SDK doesn't have a fluent API yet, correct?

What's the method to create an upload task from the session? a ctor like new LargeFileUploadTask(session, stream, options) ? And then task.upload()?

[nikithauc]

I came across this issue in JavaScript microsoftgraph/msgraph-sdk-serviceissues#39.
The non uniform casing had to be handled in JavaScript as JSON property name is case sensitive. I don't think C# has this issue though.

[andrueastman]

Yeah. The deserializer in C# is set to be case insensitive on properties.

[andrueastman]

This looks good to me now once we incorporate the comments from @nikithauc

[baywet]

@nikithauc Oh yeah! I remember running into the same issue when working on the Java one. I'm going to add a warning in the document to tell implementers to plan for that. Thanks for the reminder.

[baywet]

Thanks for the discussion! Merging. @nikithauc I haven't created issues in the JS repo to align on the spec because I know you already have multiple issues tracking work on that front. Let me know if I should create one like I did for java and dotnet.

[nikithauc]

Thanks for the discussion! Merging. @nikithauc I haven't created issues in the JS repo to align on the spec because I know you already have multiple issues tracking work on that front. Let me know if I should create one like I did for java and dotnet.

Thank you @baywet! The discussion on upload handling came up because of bugs reported in the JS repo. So we already have issues open for upload updates.