beginChunkedResponse missing request object

[EmileSpecialProducts]

I have an issue regarding the beginChunkedResponse
I am missing the AsyncWebServerRequest *request object.
As then I can use the File _tempFile; or void *_tempObject; object to keep track of what file I need to read or generate if it
If I use a global object this will be overwritten is more then one Request is made.

Or is there another way to do this? example ?

The “ request->send(SD, request->url(), String(), false);” will not work for me as I am using the https://github.com/adafruit/SdFat

This Code below will work for one Response at a time,

static file_t douwnloadFile; // Global
douwnloadFile=sd.open(request->url());
           AsyncWebServerResponse *response = request->beginChunkedResponse(dataType, [](uint8_t *buffer, size_t maxLen, size_t index)-> size_t
           {
             int len =0;
             len=douwnloadFile.read(buffer,maxLen);
             if( len == 0)  douwnloadFile.close(); 
             return len;
           });
           request->send(response);

related issue
related issue2

You can find my code portable-Async-disk-driveEx this is under construction

Or for SD card <= 32GB portable-Async-disk-drive and this is working

And for the LittleFS this version will work ESP-LittleFS-Async-Web-Server

[EmileSpecialProducts]

I have found a way to improve the handling of more then one request.
It is not 100% perfect but is an improvement of the one file request.
You can find the code in the Github page. portable-Async-disk-driveEx

server.onNotFound([](AsyncWebServerRequest *request)
    { 
      struct DownLoadFile {file_t File; size_t index;unsigned long time; };
      static DownLoadFile DownLoadFiles[5];       
      DBG_OUTPUT_PORT.printf("url NotFound %s , Method =%s\n", request->url().c_str(), request->methodToString());
      if (request->method() == HTTP_GET)
      {
          if (sd.exists(request->url())) 
          {
            String dataType = "text/plain";
            String url= request->url();
            url.toLowerCase();
                 if (url.endsWith(".htm"))   dataType = "text/html";
            else if (url.endsWith(".html"))  dataType = "text/html";
            else if (url.endsWith(".css"))   dataType = "text/css";
            else if (url.endsWith(".js"))    dataType = "application/javascript";
            else if (url.endsWith(".png"))   dataType = "image/png";
            else if (url.endsWith(".gif"))   dataType = "image/gif";
            else if (url.endsWith(".jpg"))   dataType = "image/jpeg";
            else if (url.endsWith(".bmp"))   dataType = "image/bmp";
            else if (url.endsWith(".ico"))   dataType = "image/x-icon";
            else if (url.endsWith(".xml"))   dataType = "text/xml";
            else if (url.endsWith(".pdf"))   dataType = "application/pdf";
            else if (url.endsWith(".zip"))   dataType = "application/zip";

            // clean up is time out
            for( int f=0;f<(int )(sizeof(DownLoadFiles)/sizeof(DownLoadFiles[0]));f++)
            {
                if( DownLoadFiles[f].File && DownLoadFiles[f].File.isOpen() && DownLoadFiles[f].time < millis() )
                {   DownLoadFiles[f].File.close(); 
                    DownLoadFiles[f].index=0;
                    //DBG_OUTPUT_PORT.printf("clean up %d timeout\n",f);
                }  
            }

            for( int f=0;f<(int )(sizeof(DownLoadFiles)/sizeof(DownLoadFiles[0]));f++)
              {  
                if(!DownLoadFiles[f].File.isOpen())
                {
                  DownLoadFiles[f].File=sd.open(request->url());
                  DownLoadFiles[f].index=0;
                  //DBG_OUTPUT_PORT.printf("fileindex %d Created\n",f);
                  break;
                }
              }
        
            AsyncWebServerResponse *response = request->beginChunkedResponse(dataType, [](uint8_t *buffer, size_t maxLen, size_t index)-> size_t
            {
              int len =0;
              int fileindex=-1;
              for( int f=0;f<(int )(sizeof(DownLoadFiles)/sizeof(DownLoadFiles[0]));f++)
              {
                if(DownLoadFiles[f].File && DownLoadFiles[f].File.isOpen() && DownLoadFiles[f].index==index )
                {
                  fileindex=f;
                  if(index==0) maxLen-=fileindex;
                  //DBG_OUTPUT_PORT.printf("fileindex %d found\n",f);
                  break;
                }  
              }
              if( fileindex ==-1)
              { 
                //DBG_OUTPUT_PORT.printf("fileindex not found index = %d \n",index);
              } 
              else
              { 
              
              len=DownLoadFiles[fileindex].File.read(buffer,maxLen);
              DownLoadFiles[fileindex].time= millis()+3000; // add 3000 ms for the time out of the connection.
              DownLoadFiles[fileindex].index=index+len;
              if( len == 0)  {
                              DownLoadFiles[fileindex].File.close();
                              DownLoadFiles[fileindex].index=0;
                              //DBG_OUTPUT_PORT.printf("fileindex %d Closed\n",fileindex);
                              } 
              }
              return len;
            });
            request->send(response);
            }
            else
            {
            if (request->url() == "/index.html")
                    reply(request, 200, "text/html", Index_html, sizeof(Index_html) - 1);
            else if (request->url() == "/error404.html")
                    reply(request, 404, "text/html", error404_html, sizeof(error404_html) - 1);
            else 
                request->redirect("/error404.html");
            }
        }
        else if (request->method() == MY_HTTP_POST)
        {
            reply(request, 404, "text/html", error404_html, sizeof(error404_html) - 1);
        }
    });

[mathieucarbou]

Thanks for sharing.

Yes this is pretty much the idea! Controlling the chunks and keeping track of the indexes.

Just one comment:

request->url().endsWith() => you could first extract the extension from the url path, (if it exists), then convert it to lowercase, then compare it with the extension values instead of doing a endsWith() check each time.

[EmileSpecialProducts]

Thanks for the feedback.
I have updated the code portable-Async-disk-driveEx
That will now also support UpperCase file extention.