Hi Colleen,

I think what might be happening with the files you're trying to download
is that the server is setting certain http header values to indicate to
your web browser that it should download the file rather than display
the file. Every link that you click on in a web browser simply downloads
the file that is linked to. The browser decides whether or not to
display the file or to ask the user to save it somewhere. Browsers know
how to display html, text, and images, which is why they never ask about
those. When the server wants to force the client browser to save the
file rather than display it, because the browser will try to display an
xsl file, it will include a header like, 'Content-Disposition:
attachment; filename="somefile.xsl";'. When the client browser receives
a file with a header like this, it will normally display a "Save As"
dialog where the user can choose to save the file.

So, there should be no difference for trying to save the file received
in this manner with curl versus any other URL. I don't know what your
code looks like, but I would try something like the following:

<?php

  // Set parameters.
  $url = "https://www.linkpointcentral.com/lpc/servlet/LPCReport";
  $certificateFile = 'C:/php/curl-ca-bundle.crt';
  $cookiesFile = 'C:/windows/temp/cookies.txt';
  $outputFilename = 'C:/output_file.xsl';
 
  // Set up the post data.
  $postData = array(
    'pageno' => '1',
    'expflag' => '1',
    'rptaction' => 'qryordertime',
    // ... etc.
  );
  $postFields = implode('&', $postData);
 
  // Create the curl handle with the data to send.
  $ch = curl_init();
  curl_setopt($ch, CURLOPT_POST, true);
  curl_setopt($ch, CURLOPT_POSTFIELDS, $postFields);
  curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
  curl_setopt($ch, CURLOPT_URL, $url);

  // Set other options.
  curl_setopt($ch, CURLOPT_VERBOSE, true);
  curl_setopt($ch, CURLOPT_CAINFO, $certificateFile);
  curl_setopt($ch, CURLOPT_CAPATH, $certificateFile);
  curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 0);
  curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
  curl_setopt($ch, CURLOPT_COOKIEJAR, $cookiesFile);
  curl_setopt($ch, CURLOPT_COOKIEFILE, $cookiesFile);
 
  // Get the result and output.
  $result = curl_exec($ch);

  // Output the result to a file.
  if (file_put_contents($outputFilename, $result) === false) {
    print('Failed to write file contents to '.$outputFilename."\n");
  }

  // Close the handle.
  curl_close($ch);
 
?>

Let us know if that helps,
Morgan Galpin.

Colleen R. Dick wrote:
> I have managed to log into a site and make it do a query and have my
> custom data appear on a dynamic web page (if I choose to print out the
> results of curl_exec.)
> There is a BUTTON on that page that allows you to download a spreadsheet
> of that data. You can't just fetch the file with wget or curl because
> I'm sure it doesn't live anywhere under document root so it wouldn't
> have a URL. you have to submit the form to ask the server for it and it
> downloads it as some sort of attachment with a dialog when you do it
> manually. Obviously I want to set it up so it just falls into some
> file handle. I checked what the button does, it just calls a
> javascript to set up some other post variables before submitting the
> form.
>
> Here is the submit button:
>
> <input class="button" type="button" value="Export All Data"
> onclick="exportFile(); return(false);
>
> Here is the javascript exportFile:
>
> function exportFile()
> {
> document.RptForm.pageno.value = 1;
> document.RptForm.expflag.value = 1;
> document.RptForm.rptaction.value = "qryordertime";
>
> document.RptForm.pagefile.value = "temp1196982492309.1040408.1";
>
> exptype = document.RptForm.exptype.options[document.RptForm.exptype.selectedIndex].value;
> if (exptype=="csv") document.RptForm.expfile.value="exp_ordercsv.xsl";
> else if(exptype=="xml") {
> document.RptForm.target = "_blank";
> document.RptForm.expfile.value="exp_orderxml.xsl";
> }
> document.RptForm.submit();
> document.RptForm.target = "_self";
> }
>
> So what I did was hard code all the post variables like the javascript does them
> and submits the form.
> It doesn't throw any errors, I don't think. I made it print out curl_getinfo
> Array
> (
> [url] => https://www.linkpointcentral.com/lpc/servlet/LPCReport
> [content_type] => text/html; charset=ISO-2259-1
> [http_code] => 200
> [header_size] => 153
> [request_size] => 665
> [filetime] => -1
> [ssl_verify_result] => 0
> [redirect_count] => 0
> [total_time] => 2.981608
> [namelookup_time] => 0.000101
> [connect_time] => 0.113942
> [pretransfer_time] => 0.529531
> [size_upload] => 0
> [size_download] => 0
> [speed_download] => 0
> [speed_upload] => 0
> [download_content_length] => 0
> [upload_content_length] => 0
> [starttransfer_time] => 2.981528
> [redirect_time] => 0
> )
>
> It names the document rpt*******.csv but I would like it to download into a file name of my choice.
> To that end I have set up and opened a filehandle but it always ends up of size zero, which makes sense if
> I don't know how to tell CURL to use it. Where is the document going if anywhere?
> It is not in tmp or in my web tree. I'm sorry if I am stupid but I have really looked very hard for the answer.
> All I see is little examples to use Curl to download a file where it has a URL. I know how to do that.
>
> _______________________________________________
> http://cool.haxx.se/cgi-bin/mailman/listinfo/curl-and-php
>
_______________________________________________
http://cool.haxx.se/cgi-bin/mailman/listinfo/curl-and-php
Received on 2007-12-07