C# Printing and Threading

Question

I have a need for users to be able to scan a series of items and for each item print off x number of labels. I am currently trying to use a background worker to accomplish this but I have run into a problem where they are scanning items so fast and there are so many labels to print for each item the background worker chokes. This is how I am generating the background worker thread for each scan because contention was occurring when there was a large number of labels be printed.

 private void RunPrintWorker()
    {
        if (printWorker.IsBusy)
        {
            printWorker = new BackgroundWorker();
            printWorker.DoWork += new DoWorkEventHandler(printWorker_DoWork);
            printWorker.RunWorkerAsync();
        }
        else
            printWorker.RunWorkerAsync();
    }

I don't get any exceptions from the background worker it just seems to not be creating threads fast enough. I am newer to using multiple threads so can anyone point me in the direction of what I am doing wrong?

Thanks.

EDIT: Thanks everyone for the suggestions and reading material this should really help. The order the labels are being printed doesn't really matter since they are scanning pretty fast and the labels are only being printed to one printer too. I will mark an answer after I get the implementation up and running.

EDIT: Austin, below is how I have my printing method setup. Before I was just calling LabelPrinter.PrintLabels in my RunPrintWorker method. Now that I am redoing this I can't figure out what to pass into the SizeQueue method. Should I be passing the newly created print document into it?

 public class LabelPrinter
{
    private int CurrentCount = 0;

    private List<int> _selectedRows = new List<int>();
    public List<int> SelectedRows
    {
        get { return _selectedRows; }
        set { _selectedRows = value; }
    }

    private string _selectedTemplate;
    public string SelectedTemplate
    {
        get { return _selectedTemplate; }
        set { _selectedTemplate = value; }
    }

    private string _templateDirectory = string.Empty;
    public string TemplateDirectory
    {
        get { return _templateDirectory; }
        set { _templateDirectory = value; }
    }

    public void PrintLabels(PrintDocument printDoc, PageSettings pgSettings, PrinterSettings printerSettings, List<int> selectedRows, string selectedTemplate, string templateDir)
    {
        this._selectedRows = selectedRows;
        this._selectedTemplate = selectedTemplate;
        this._templateDirectory = templateDir;

        printDoc.DefaultPageSettings = pgSettings;
        printDoc.PrinterSettings = printerSettings;

        printDoc.PrinterSettings.MaximumPage = selectedRows.Count();
        printDoc.DefaultPageSettings.PrinterSettings.ToPage = selectedRows.Count();
        printDoc.PrinterSettings.FromPage = 1;

        printDoc.PrintPage += new PrintPageEventHandler(printDoc_PrintPage);

        printDoc.Print();
    }

    private void printDoc_PrintPage(object sender, PrintPageEventArgs e)
    {
        CurrentCount = DrawLabel.DrawLabelsForPrinting(e, SelectedTemplate, SelectedRows, CurrentCount, TemplateDirectory);
    }
}

Answer 1

Try adding the items to a queue (for example, Queue<Item>) and have the BackgroundWorker process the queue.

EDIT: Adding some simple, untested code that may work for you. I would encapsulate the print queue with its processor and just send it jobs.

class SimpleLabelPrinter
{
    public bool KeepProcessing { get; set; }
    public IPrinter Printer { get; set; }

    public SimpleLabelPrinter(IPrinter printer)
    {
        Printer = printer;
    }


    /* For thread-safety use the SizeQueue from Marc Gravell (SO #5030228) */        
    SizeQueue<string> printQueue = new SizeQueue<string>();

    public void AddPrintItem(string item)
    {
        printQueue.Enqueue(item);
    }

    public void ProcessQueue()
    {
        KeepProcessing = true;

        while (KeepProcessing)
        {
            while (printQueue.Count > 0)
            {
                Printer.Print(printQueue.Dequeue());
            }

            Thread.CurrentThread.Join(2 * 1000); //2 secs
        }
    }
}

class Program
{
    static void Main(string[] args)
    {
        SimpleLabelPrinter printer1 = new SimpleLabelPrinter(...);
        SimpleLabelPrinter printer2 = new SimpleLabelPrinter(...);

        Thread printer1Thread = new Thread(printer1.ProcessQueue);
        Thread printer2Thread = new Thread(printer2.ProcessQueue);

        //...

        printer1.KeepProcessing = false;  //let the thread run its course...
        printer2.KeepProcessing = false;  //let the thread run its course...
    }
}

SizeQueue implementation

EDIT 2: Addressing the updated code in the question

First, I would define a PrintJob class that contains the number of copies to print and either the complete label text or enough data to derive it (like IDs for a DB query). This would lead you to replace SizeQueue<string> in my code above to SizeQueue<PrintJob> as well as AddPrintItem(string item) to AddPrintJob(PrintJob job).

Second, I would keep your LabelPrinter code separated (perhaps create that IPrinter interface) and pass that into the constructor of my SimpleLabelPrinter (which may not be the best name at this point but I'll let you handle that).

Next, create your LabelPrinter and SimpleLabelPrinter (say printer1 for this example) wherever it's appropriate for your app (in your apps Closing or "cleanup" method, be sure to set KeepProcessing to false so its thread ends).

Now when you scan an item you send it to the SimpleLabelPrinter as:

printer1.AddPrintJob(new PrintJob(labelText, numberOfCopies));

Answer 2

Basically what you are doing is saying if the printer is busy, overwrite your printWorker object with another worker and start that one. You then have no reference to your old worker object and don't do any cleanup.

There are all kinds of problems with this.

• Background workers must be disposed of when finished to prevent leaks.
• Creating more background works doesn't mean the printer will run any faster, it just means you have more workers trying to get to the printer at the same time.

What you need to do is think about queuing. I would say you have two main options.

First - Use ThreadPool.QueueUserWorkItem(...)

ThreadPool.QueueUserWorkItem(new WaitCallback(o=>{printWorker_DoWork();}));

This will use the .net threadpool to queue up your tasks and process them on the pool. (The pool will auto size to an appropriate number of threads to handle your queued requests). This requires no disposal or clean up, you just fire and forget, although you need to be aware that the work items are not guaranteed to be processed in the same order you add them in (it's likely some will be processed simultaneously), so if you care about order, this is no good.

Second - Implement the Producer-consumer thread queue pattern. This requires a synchronised queue to which one thread, or threads (the producer) adds items, and other threads (the consumers) remove and process those items. If you are new to threading this can be quite tricky to get right as you have to make sure your reading/writing to the shared queue is properly synchronised so the order is maintained and nothing is lost.

Be careful of the Queue<T> type, although it could be used, it is not automatically thread safe on it's own. Make sure if you do use it you add some of your own locking and synchronisation.

From the MSDN page:

Any instance members are not guaranteed to be thread safe.

A Queue<T>) can support multiple readers concurrently, as long as the collection is not modified. Even so, enumerating through a collection is intrinsically not a thread-safe procedure. To guarantee thread safety during enumeration, you can lock the collection during the entire enumeration. To allow the collection to be accessed by multiple threads for reading and writing, you must implement your own synchronization.

I would recommend the more simple first option if you don't care about order, but if you find you need the more fine grained control over the queuing and processing, or order is important, that is when you could move to the more complex pattern - if you follow the links at the bottom of the article, you will find examples and source code, alternatively you can just use the Queue class and make sure you get your locking right.