WebServer: rename & expose internal 'contentLength' variable for POST…

[WebDust21]

The purpose of this PR is to expose the internal "contentHeader" of the WebServer, so user code can determine the approximate size of an incoming HTTP POST upload (file or otherwise).

Description of Change

Currently, there is no way to determine the size of a file upload to the ESP32 until after all of the data has been read. The WebServer.Upload structure does a good job providing all of the available HTTP information to the user software, but for whatever HTTP reason, the total size is not provided in the HTTP form header (see picture below):

It turns out that the initial HTTP POST request header does provide a general idea of how much data is incoming via the "content-length" header: Yes, I am fully aware that this is not the exact size of the uploaded file (as the size of the HTTP headers are included in this count--the file I uploaded here is only 84,103 bytes)...but for larger files, this can be very useful to the user software for determining whether to process or ignore the incoming data. If someone's trying to upload their prized photo that's a 6MB JPEG into an ESP32 with only 498KB of available space, it would make sense to be able to reject the upload on the spot.

The WebServer implementation does not store the HTTP headers for the user software, but it does parse them. It turned out that the "content-length" header was parsed and internally used--but not exposed to the user.

This PR changes the scope of the internal "content-length" variable, and exposes a read-only access point in the WebServer structure.

Tests scenarios

Tested on the latest "master" (2.0.4) with an ESP32-WROOM32 on a custom PCB. Works beautifully ;-)

Example code snippit:

void uploadHandler()
{
  // This is called 
  //    Status = 0 -> when the request starts
  //    Status = 1 -> every time there is a new "chunk" of data
  //    Status = 2 -> when the last chunk of data comes in
  HTTPUpload& upload = server.upload();
  if (upload.status == 0) {
    Serial.print("Packet Length: "); Serial.println(server.clientContentLength());  // new function in this PR
  }
  
  Serial.print("ulStatus: "); Serial.println(upload.status);
  Serial.print("ulFilename: "); Serial.println(upload.filename);
  Serial.print("ulName: "); Serial.println(upload.name);
  Serial.print("ulType: "); Serial.println(upload.type);
  Serial.print("ulLen: "); Serial.println(upload.totalSize);
  Serial.print("ulBufLen: "); Serial.println(upload.currentSize);
  Serial.write(upload.buf, 50);  // just print the beginning of the buffer
  Serial.println();
}

Result when uploading a file to the ESP32:

AFAIK the PR should not break anything. Changes are minimal. I am not sure what the best name for the exposed function is. To the best of my knowledge, it is only usable for incoming HTTP POST requests to the ESP32...and is not to be confused with the existing "setContentLength" function (used for setting the length of an outgoing HTTP request)

(also includes a quick code optimization, consolidating 2 IF statements on the same boolean variable into an IF-ELSE. Also couldn't quite help myself with some spaces vs tabs & lining comments up...)

[CLAassistant]

All committers have signed the CLA.

[vortigont]

Nice feature, the same is already present in AsyncWebServer. But maybe it should be better accessible as HTTPUpload struct member than HTTPServer, i.e. for GET requests it makes no sense even if reported as zero. Seems more logical to me.

[WebDust21]

Nice feature, the same is already present in AsyncWebServer. But maybe it should be better accessible as HTTPUpload struct member than HTTPServer, i.e. for GET requests it makes no sense even if reported as zero. Seems more logical to me.

AsyncWebServer has everything, I just don't want to have to rewrite my entire project to switch over ;-). Plus then I'd have the old WebServer code consuming space in the binary...PLUS the AsyncWebServer.

Whatever way makes the most sense... ...my reservations with putting it in the HTTPUpload struct is that it isn't specifically limited to HTTPUpload alone. And if HTTPUpload can parse multiple file uploads in one POST stream (which I haven't tested), the "content-length" property will include the total of all files, not just the one currently being handled by said struct.

[vortigont]

my reservations with putting it in the HTTPUpload struct is that it isn't specifically limited to HTTPUpload alone

you are right, this makes sense actually. POST could be without file-body also, i.e. large json for REST. Anyway being able to get body size length is handy. Like it!

[SuGlider]

@WebDust21 - Thanks for the contribution. I just added a comment in the review. PTAL