tags:

views:

104

answers:

1

We are using ExpertPDF to take URLs and turn them into PDFs. Everything we do is through memory, so we build up the request and then read the stream into ExpertPDF and then write the bits to file. All the files we have been requesting so far are just plain HTML documents. Our designers update CSS files or change the HTML and rerequest the documents as PDFs, but often times, things are getting cached. Take, for example, if I rename the only CSS file and view the HTML page through a web browser, the page looks broke because the CSS doesn't exist. But if I request that page through the PDF Generator, it still looks ok, which means somewhere the CSS is cached. Here's the relevant PDF creation code:

// Create a request
HttpWebRequest request = (HttpWebRequest)HttpWebRequest.Create(url);
request.UserAgent = "IE 8.0";
request.ContentType = "application/x-www-form-urlencoded";
request.Method = "GET";

// Send the request
HttpWebResponse resp = (HttpWebResponse)request.GetResponse();
if (resp.IsFromCache) {
    System.Web.HttpContext.Current.Trace.Write("FROM THE CACHE!!!");
} else {
    System.Web.HttpContext.Current.Trace.Write("not from cache");
}

// Read the response
pdf.SavePdfFromHtmlStream(resp.GetResponseStream(), System.Text.Encoding.UTF8, "Output.pdf");

When I check the trace file, nothing is being loaded from cache. I checked the IIS log file and found a 200 response coming from the request, even after a file had been updated (I would expect a 302). We've tried putting the No-Cache attribute on all HTML pages, but still no luck. I even turned off all caching at the IIS level. Is there anything in ExpertPDF that might be caching somewhere or something I can do to the request object to do a hard refresh of all resources?

UPDATE

I put ?foo at the end of my style href links and this updates the CSS everytime. Is there a setting someplace that can prevent stylesheets from being cached so I don't have to do this inelegant solution?

+1  A: 

Actually this is a perfectly normal solution, though I would recommend using something like the current date and time attached to the PDF link/file name (like you did for the css sheet

HttpWebRequest request = (HttpWebRequest)HttpWebRequest.Create(url + DateTime.Now.ToString().Replace(":", "").Replace("-", "").Replace(" ", ""));)

rather than foo on your stlye sheet. As the date and time will ALWAYS change, you will force the download each time.

I would venture to guess that the caching is not the CSS style sheet, but rather the PDF is being cached by the client. Adding the URL variable to your stylesheet is preventing it from being cached. (I think you fixed the problem, but probably not, in my opinion, the best way) Try the above tip, and you should not have any file caching problems.

PS. I know you can use DateTime.Now.ToString(formathere) but I am too lazy to look it up right now ;)

andrewWinn
The PDF is never sent to the client. It is stored on the server, which is what has me confused as there wouldn't be any client to cache things.
Josh