Analysis of the response parsing process of SAP UI5 batch request

Reply to batch request:

Parse with the appropriate handler:

The handler accept field: multipart/mixed

First, according to the response field, get the content-type:

content-type field, separated by ; :

media type：multipart/mixed

Returns a JSON object with mediaType and boundary properties:

The handler for batch requests is batchParser :

Execute the next line of the readBatch method:

An example of boundary: batchresponse_16aba97e-4311-41a4-8c64-302727d1d02e
The specific processing logic of readBatch is based on string matching, that is, using indexOf to find the boundary string:

The implementation of readline is to read a single line:

The header field of the first part of the batch is read:

The media type must match the hardcoded value: application/http

The data of the first request count included in the batch has been successfully read:

It has been read successfully, but also dispatch:

The media-type of count is text/plain:

return false, the body of the response has a value, and the data is undefined:

There are 4 available handlers:

text/plain is responsible for reading response.body.

The logic of textParser is the simplest, it directly returns the body field to response.data:

Read the second part:

The content-type of this part is json, so use the json handler:

if (handlerAccepts(handler, cType)) {
            var readContext = createReadWriteContext(cType, version, context, handler);
            readContext.response = response;
            response.data = parseCallback(handler, body, readContext);
            return response.data !== undefined;
        }

The semantics of the above code is that only after a handler has passed the filtering of function handlerAccepts , can parseCallback be called, and the parsed result will be passed to response.data .

The json handler uses the JSON.parse native API to serialize json data.