Add omitempty to HTTPMethod in APIGatewayProxyRequest

[LegNeato]

This can be empty for websocket events, see calavera/aws-lambda-events#33

Issue #, if available:

Description of changes:

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[codecov-io]

Codecov Report

Merging #350 (6b77cd7) into master (b3dd246) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #350   +/-   ##
=======================================
  Coverage   72.22%   72.22%           
=======================================
  Files          19       19           
  Lines         738      738           
=======================================
  Hits          533      533           
  Misses        138      138           
  Partials       67       67           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b3dd246...6b77cd7. Read the comment docs.

[harrisonhjones]

I'm just curious: why do you want this? Are you marshalling this struct to JSON? Why?

[LegNeato]

I generate rust code from the go code: calavera/aws-lambda-events#33

[bmoffatt]

did you mean to modify APIGatewayWebsocketProxyRequest instead?

Since it's unused, I think we could deprecate the field, and remove the json tag. Would that play well with your code generator?

[LegNeato]

I'm pretty sure I modified the correct struct. If you look at the corresponding C# class (https://github.com/aws/aws-lambda-dotnet/blob/master/Libraries/src/Amazon.Lambda.APIGatewayEvents/APIGatewayProxyRequest.cs) they point out that different fields can be null for this event depending on if the source is http or websockets.

I just look at the structs, so if there is additional marshalling logic elsewhere that chooses between the two structs please let me know and I'll see what I can do on my side.

[bmoffatt]

There's no marshaling logic, it's up the the author of the function to select the correct struct for the event source. The java events project also has a separate type for web socket events

[calavera]

@LegNeato it might make more sense to modify the Websocket structure like @bmoffatt suggests. Leaving up to the user is kind of unfortunate because for them both requests come from origin. It will probably be reasonable to supplement with specific examples of usage for each one of the structs.

[bmoffatt]

superseded by #355