Detailed explanation of alternative solutions to Html to Pdf

Background

The project requires generating a pdf document from an HTML page (payment result). The page has pictures and tables, but it seems that the open source iTextSharp cannot handle it.

After some searching, I found wkhtmltopdf, a command line open source conversion tool that supports specifying url or local html The path of file worked well after the trial. I also specially wrote a tool using wkhtmltopdf to back up the pdf of blog posts to the local. I will share this tool when I have time later.

But, send After two days of testing for customers, the running results were not ideal, and some unknown errors occurred. What's strange is that there were no problems in the test environment, but frequent errors occurred in the official environment. Finally, the customer gave up on this solution
Attached is WkhtmlToXSharp C# wrapper wrapper (using P/Invoke) for the excelent Html to PDF conversion library wkhtmltopdf library.
***
OK, come to the topic, an alternative solution: Hook

Call the IE printing function, use the XPS printer, first generate the HTML file into an xps document, and then generate a pdf

New WinForm project, drag into the WebBrowser control, the code specifies the Url to the local html file path, wait for the document to be loaded, WebBrowser.Print(); OK, run, the dialog box for selecting the printer will pop up, as shown in Figure 1. After clicking print, a save as dialog box will pop up. Enter the xps path and save it (Figure 2), and you will get an xps document.

Picture 1: Select the printer

Picture 2: Enter the xps path

As you can see from the above, The printing here requires interaction with the UI, manually click print, and enter the xps path to save.
Next, I searched on the Internet: How to directly print and generate xps files without displaying the dialog box. I read a lot on stackoverflow and codeproject, but I couldn't find a way. Later, I accidentally read the article of Yuanzi's predecessor, using hook method and UI Automation to complete the printing and saving actions. I think this solution is feasible

Let's code it next

    //调用WebBrowser.Print的代码就忽略了，直接看钩子
    IntPtr hwndDialog;
    string pathFile;
    EnumBrowserFileSaveType saveType;

    // Imports of the User32 DLL. 
    [DllImport("user32.dll", CharSet = CharSet.Auto)]
    public static extern IntPtr SendMessage(IntPtr hWnd, int msg, int wParam, int lParam);

    [DllImport("user32.dll", CharSet = CharSet.Auto)]
    public static extern IntPtr GetDlgItem(IntPtr hWnd, int nIDDlgItem);

    [DllImport("user32.dll", CharSet = CharSet.Auto)]
    static extern private bool SetWindowText(IntPtr hWnd, string lpString);
    [DllImport("user32.dll")]
    [return: MarshalAs(UnmanagedType.Bool)]
    static extern bool IsWindowVisible(IntPtr hWnd);

    //Win32 Api定义
    [DllImport("user32.dll")]
    static extern IntPtr FindWindow(string lpClassName, string lpWindowName);

    [DllImport("user32.dll")]
    static extern IntPtr FindWindowEx(IntPtr hwndParent, IntPtr hwndChildAfeter, string lpszClass, string lpszWindow);

    [DllImport("user32.dll")]
    static extern IntPtr SendMessage(IntPtr hWnd, uint Msg, IntPtr wParam, String lParam);

    [DllImport("user32.dll")]
    static extern bool PostMessage(IntPtr hWnd, uint Msg, IntPtr wParam, IntPtr lParam);


    //Win32消息定义
    const uint WM_SETTEXT = 0x000c;
    const uint WM_IME_KEYDOWN = 0x0290;
    const uint WM_LBUTTONDOWN = 0x0201;
    const uint WM_LBUTTONUP = 0x0202;

    // The thread procedure performs the message loop and place the data
    public void ThreadProc()
    {
        int maxRetry = 10;
        int retry = 0;
        IntPtr hWndPrint = FindWindow("#32770", "打印");
        IntPtr hWnd = FindWindow("#32770", "文件另存为");
        if (hWnd != IntPtr.Zero)
        {
            log.InfoFormat("got saveas dialog handle. Printer Dialog skipped.");
        }
        else
        {
            Thread.Sleep(200);
            hWndPrint = FindWindow("#32770", "打印");

            //这里有时候获取不到window,所以加了Sleep，多试几次
            while (hWndPrint == IntPtr.Zero && retry < maxRetry)
            {
                Thread.Sleep(200);
                log.InfoFormat("retry get Print dialog handle.retry:{0}", retry);
                hWndPrint = FindWindow("#32770", "打印");
                retry++;
            }
            if (hWndPrint == IntPtr.Zero)
            {
                //wait 1 second,retry again
                Thread.Sleep(1000);
                hWndPrint = FindWindow("#32770", "打印");
            }
            if (hWndPrint == IntPtr.Zero)
            {
                log.InfoFormat("Did not get Print dialog handle.retry:{0}", retry);
                return;
            }
            log.InfoFormat("got Print dialog handle.retry:{0}", retry);
            //select printer dialog
            IntPtr hChildP;
            hChildP = IntPtr.Zero;
            hChildP = FindWindowEx(hWndPrint, IntPtr.Zero, "Button", "打印(&P)");
            // 向保存按钮发送2个消息，以模拟click消息，借此来按下保存按钮
            PostMessage(hChildP, WM_LBUTTONDOWN, IntPtr.Zero, IntPtr.Zero);
            PostMessage(hChildP, WM_LBUTTONUP, IntPtr.Zero, IntPtr.Zero);
            Application.DoEvents();
        }

        //hWnd = FindWindow("#32770", null);
        hWnd = FindWindow("#32770", "文件另存为");
        //To avoid race condition, we are forcing this thread to wait until Saveas dialog is displayed.
        retry = 0;
        while ((!IsWindowVisible(hWnd) || hWnd == IntPtr.Zero) && retry < maxRetry)
        {
            Thread.Sleep(200);
            log.InfoFormat("retry get saveas dialog handle.retry:{0}", retry);
            hWnd = FindWindow("#32770", null);
            retry++;
            Application.DoEvents();
        }
        log.InfoFormat("got saveas dialog handle.retry:{0}", retry);
        if (hWnd == IntPtr.Zero)
        {
            //wait 1 second,retry again
            Thread.Sleep(1000);
            hWnd = FindWindow("#32770", "文件另存为");
        }
        if (hWnd == IntPtr.Zero)
        {
            return;
        }
        Application.DoEvents();

        IntPtr hChild;
        // 由于输入框被多个控件嵌套，因此需要一级一级的往控件内找到输入框
        hChild = FindWindowEx(hWnd, IntPtr.Zero, "DUIViewWndClassName", String.Empty);
        hChild = FindWindowEx(hChild, IntPtr.Zero, "DirectUIHWND", String.Empty);
        hChild = FindWindowEx(hChild, IntPtr.Zero, "FloatNotifySink", String.Empty);
        hChild = FindWindowEx(hChild, IntPtr.Zero, "ComboBox", String.Empty);
        hChild = FindWindowEx(hChild, IntPtr.Zero, "Edit", String.Empty); // File name edit control
        // 向输入框发送消息，填充目标xps文件名
        SendMessage(hChild, WM_SETTEXT, IntPtr.Zero, pathFile);
        // 等待1秒钟
        System.Threading.Thread.Sleep(1000);
        // 找到对话框内的保存按钮
        hChild = IntPtr.Zero;
        hChild = FindWindowEx(hWnd, IntPtr.Zero, "Button", "保存(&S)");
        // 向保存按钮发送2个消息，以模拟click消息，借此来按下保存按钮
        PostMessage(hChild, WM_LBUTTONDOWN, IntPtr.Zero, IntPtr.Zero);
        PostMessage(hChild, WM_LBUTTONUP, IntPtr.Zero, IntPtr.Zero);

        // Clean up GUI - we have clicked save button.
        //GC is going to do that cleanup job, so we are OK
        Application.DoEvents();
        //Terminate the thread.
        return;
    }

The next part is about xps To convert pdf, I used Spire.Pdf. There is an official demo. I will not explain it here.

There are pictures and the truth.

About automatic selection of XPS Document I haven't completed the Writer hook code yet, please give me some advice!

The above is the detailed content of Detailed explanation of alternative solutions to Html to Pdf. For more information, please follow other related articles on the PHP Chinese website!