Unable to retrieve data from request GET typed with getBody()

[leomlr]

Hi,
Thanks for implementing this library to manage GET/POST requests.
I encountered a problem when trying to retrieve data from an HTML page after sending a GET request.

It seems that the http.getBody() method is unable to retrieve the string content due to low memory. Considering the size of the data in an HTML page, I suggest using std::move() which steals the memory location of http.body. So I modified the getBody() function accordingly.

In R4HttpClient.h, I changed for:

String R4HttpClient::getBody()
{
  return std::move(this->body);
}

In this sense, it is possible to reassign a variable of type String in the main program and thus use it.
What do you think about this?

[piscodev]

Hello! 👍

May I know the byte size of the HTML page after sending a GET request? I would like to reproduce it on my side. Also, thanks for the suggestion, I would very much consider it! As I was testing few months ago, I also encountered the peak memory capacity right after having few requests. So, I revised quite few things then I hit dangling, revised things again, and then now, as for your suggestion... it may be a good way to optimize allocated space. I did read some articles and reviewed few examples, and I believe your approach could work.

The current:

String R4HttpClient::getBody() const
{
  // this makes a COPY of this->body
  return this->body;
}

So when I call:
String response = http.getBody();
Arduino will allocate new memory for String 'response' and copy all characters from this->body into it. Doubling the buffer.

For example: If 'this->body' contains 10KB of HTML, it will temporarily use 10KB extra RAM for the copy. Now, As for your suggestion...

String a = http.getBody();  // with std::move
// now `http.body` is cleared, `a` owns the data

String b = http.getBody();  // with copy
// `http.body` is unchanged, but `b` is a duplicate

Based on that, I’m planning to revise getBody() to return with std::move you suggested so that the buffer isn’t duplicated anymore, thus optimizing it.

Thanks again for pointing me in this direction. I’ve appreciated your thoughts on whether this adjustment is the right approach.