PDF Generation fails when the application is run with compilation debug = false

Question

I am using Nreco PDF to convert html pages to PDF . Recently Bundling and Minification was implemented and hence the compilation debug = false was set in the web config .

Since then the PDF Generation is failing and chrome shows up this message saying "Failed to Load PDF Document".

When i turn on debug mode , everything works as expected.

Here is the code snippet:

public ActionResult ABC()
{
    var htmlContent =System.IO.File.ReadAllText(Server.MapPath("~/Views/abc/abc.html"));
    createpdf(htmlContent);
}

public void createpdf(string htmlContent)
{
    var htmlToPdf = new NReco.PdfGenerator.HtmlToPdfConverter();
    var pdfBytes = htmlToPdf.GeneratePdf(htmlContent);

    Response.ContentType = "application/pdf";
    Response.ContentEncoding = System.Text.Encoding.UTF8;
    Response.AddHeader("Content-Disposition", "Inline; filename=TEST.pdf");
    Response.BinaryWrite(pdfBytes);
    Response.Flush();
    Response.End();
}

I would like to know whats causing this code to fail when run with debug =false.

Have you used DEBUG flag anywhere in this code ? ALso is the html same in Debug and normal mode ? — Vandita, Apr 28 '17 at 10:25
No i have not .. this exact code generates pdf when run debug = true , and fails when debug = false (web config) — REDEVI_, Apr 28 '17 at 10:26
Can you see `NReco.PdfGenerator.dll` in your `bin\Debug\` Folder ? — Vandita, Apr 28 '17 at 10:29
its present in bin , not sure if its there in debug folder at the test env. — REDEVI_, Apr 28 '17 at 10:33
@Redevi_ When you say it fails... can you tell us an error message or what? — Milney, Apr 28 '17 at 10:46
@Milney Since the output is a response , i just get the message from chrome saying Failed to load PDF Docuement , there are no exceptions either in console or in code (used try catch ) — REDEVI_, Apr 28 '17 at 10:51
@Milney i just got this more extra info , net::ERR_CONTENT_DECODING_FAILED — REDEVI_, Apr 28 '17 at 10:53
@Redevi_ Ahh it may be the GZip encoding... try checking the headers about compression and either fix them or turn that off.. look here: http://stackoverflow.com/questions/14039804/error-330-neterr-content-decoding-failed — Milney, Apr 28 '17 at 10:53
I have created MVC template trying to reproduce the issue and I could not. Code works in both debug true and false attributes. Could you create a simple project on GitHub where you can reproduce this behavior? — Andrii Litvinov, May 02 '17 at 06:01
Do you host it on IIS or IIS Express? Can you download static PDF from your server? — Andrii Litvinov, May 02 '17 at 06:02
the pdf is dynamically generated from code , and shown inline to the user .. it runs on iisexpress — REDEVI_, May 02 '17 at 06:29
Can you save the pdfBytes to a local file before write to response, if the local pdf file can open normally, maybe there is a problem at Response.End(), otherwise Neco has problem in release mode(eg: invoke the x64 htmltopdf.exe ). — aspark, May 02 '17 at 06:52
I did save the pdf , but it was broken , adobe dint open it saying it could not decode ,.. — REDEVI_, May 02 '17 at 06:54
How long time does it take?, could be a timeout issue when generating pdf. in debugmode MVC calls don't timeout — Daniel Stackenland, May 02 '17 at 07:55
Doesnt seem to be a timeout issue , i just returned one line of html hardcoded to it and it still fails .. — REDEVI_, May 02 '17 at 07:58
Are you using latest version of NReco? Try create new project and do the same. Will it fail? — Andrii Litvinov, May 02 '17 at 08:02
Cool, try also updated to latest 1.1.15 version, because they do behave differently on my PC, though don't fail. — Andrii Litvinov, May 02 '17 at 08:13

score 2 · Accepted Answer · edited Jun 20 '20 at 09:12

Not necessary an answer:

Decorate code with a try -> catch, debug an see if any errors in catch, most likely htmlToPdf.GeneratePdf will fail, if not continue with next debugging steps

Make sure your PDF generator is working correctly, meaning you will have a valid pdf file, before returning the bytes store pdf on you App_Data folder in solution

   var appDataPath = Server.MapPath("~/App_Data");
   var filename = string.Format(@"{0}.pdf", DateTime.Now.Ticks);
   var path = Path.Combine(appDataPath, filename);
   System.IO.File.WriteAllBytes(path, pdfBytes.ToArray());

Check creating and closing response

Response.Clear();
Response.ClearContent();
Response.ClearHeaders();
Response.ContentType = "application/pdf";
Response.ContentEncoding = System.Text.Encoding.UTF8;
Response.AddHeader("Content-Disposition", "Inline; filename=TEST.pdf");
Response.BinaryWrite(pdfBytes);
Response.Flush();
Response.Close()
Response.End()

Edit: After long debugging sessions, what was discovered: When using WebMarkupMin.Mvc for compress content or minimization Controller Action result is compressed incorrect (is correct, but not with your way of return the pdf).

webMarkupMin default settings are as following:

enableMinification="true"

disableMinificationInDebugMode="false"

enableCompression="true"

disableCompressionInDebugMode="false"

This is the reason why in debug mode was running correct when

compilation debug="true"

To get same error when debug="true" in web.config set:

<webMarkupMin xmlns="http://tempuri.org/WebMarkupMin.Configuration.xsd">
  <webExtensions enableMinification="true" disableMinificationInDebugMode="false" enableCompression="true" disableCompressionInDebugMode="false" />
 <!-- rest of shebang -->
</webMarkupMin>

Where is the issue you might ask, well is simple in your implmentation Response stream is compressed incorrect.

Overcoming the problem:

 //you can use FileResult, same outcome
    public ActionResult ABC()
            {
                var htmlContent = System.IO.File.ReadAllText(Server.MapPath("~/Views/abc/abc.html"));
                var htmlToPdf = new NReco.PdfGenerator.HtmlToPdfConverter();
                var pdfBytes = htmlToPdf.GeneratePdf(htmlContent);
                //return File(pdfBytes, "application/pdf", "TEST.pdf"); 
                //will add filename to your response, downside is that browser downloads the response

                //if your really need Content Disposition in response headers uncomment below line
                //Response.AddHeader("Content-Disposition", "Inline; filename=TEST.pdf");
                return File(pdfBytes, "application/pdf");
                
            }

The issue is when i run in debug mode , the file is generated properly and served . I did add a try catch block but no exceptions are occurring when run in debug mode = true , I have this issue occurring only when debug mode = false as i have stated in the question — REDEVI_, May 02 '17 at 07:06
also note i m reading from a html file and preparing a pdf out of it — REDEVI_, May 02 '17 at 07:14
Run with debug mode = false and check is file is saved if yes, debug in this way=> try, catch and finally, in finally set the only break point, check if fails in browser. — SilentTremor, May 02 '17 at 07:15
unfortunately it doesn't work , can u help me in identifying which part of code may have a diff behavior when debug is off ? — REDEVI_, May 02 '17 at 07:28
What .net framework are you using? to try and replicate the thingy? — SilentTremor, May 02 '17 at 07:31
Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/143156/discussion-between-redevi-and-silenttremor). — REDEVI_, May 02 '17 at 08:24
Great job with investigation! I have tried `File` from start when trying to reproduce, but thought both approaches should work same way. — Andrii Litvinov, May 02 '17 at 18:49

score 0 · Answer 2 · answered May 02 '17 at 12:00

change to use File() to return the bytes, like : return File(pdfBytes, "application/pdf"); do not use the Response.BinaryWrite() directly return the bytes, this will cause the minifer can not read the pdf stream, so return a empty response to client

PDF Generation fails when the application is run with compilation debug = false

2 Answers2