I have below code in Bot framework app.
You can see in below code I have commented ValidateStartDate delegate, the reason behind it is that if I include delegate in formflow then after the delegate execution code jumps directly to BookingComplete delegate of "context.Call(Booking, BookingComplete);" i.e end of conversation.But ideally, it should execute rest of the fields from form builder.
Note that here StartDate is of type String, and I am manually validating date part.Also, no visible exception occurs during code execution
public static IForm<ConferenceBooking> BuildForm()
{
return new FormBuilder<ConferenceBooking>().Message("Tell me meeting details!")
.Field(nameof(title))
.Field(nameof(StartDate))//, validate: ValidateStartDate
.Field(nameof(EntryTime), validate:ValidateCallTime)
.Build();
}
Below is delegate part for StartDate
private static Task<ValidateResult> ValidateStartDate(ConferenceBooking state, object response)
{
var result = new ValidateResult();
DateTime startDt = Convert.ToDateTime(GetDate(Convert.ToString(response)));
if (startDt == null || startDt == DateTime.MinValue)
{
result.IsValid = false;
result.Feedback = "I could not understand this format.";
}
else if (startDt.Date < DateTime.Now.Date)
{
result.IsValid = false;
result.Feedback = "Sorry, back dated bookings are not allowed";
}
else
{
result.IsValid = true;
result.Value = startDt;
}
return Task.FromResult(result);
}
I have also noticed this behavior before and this was always due to an exception. The FormBuilder appears to catch all exceptions and exits the form in the catch block. This is why you don't see any exceptions popping up. Try going over your code step by step or executing it from outside the form.
Related
I’ve run into a performance obstacle and I’m uncertain of the cause, all of this is running under VS2022 & Net6.0. As this is my 1st time using this combination of a modal windows form, and progress bar, with the work running on a background thread and two Progress objects updating the UI, the progress bar, and a text label, I don’t know where to attack the problem. Prior to placing the workload on a background thread, everything was snappy, searching a thousand files with about 600 lines of text in each, in about a minute. Naturally, the windows form was frozen during this, which is why the workload was placed on a background thread.
After doing so, the workload will be 25-50% complete before the UI starts displaying the values from the Progress objects, and overall, the entire process now takes 10x as long to complete. Progress objects aren’t skipping over any values sent to them, the UI thread just seems slow in getting the information. Likewise, if I try to drag the modal form to a new spot on the desktop it’s unresponsive for 20—30 seconds before it finally moves. One more thing, I can step through the code on the background thread and see it calling the Progress updaters, but the UI thread is just very slow in responding to them.
I could use some suggestions on how to uncover the problem or if clearly evident, point out where the likely problem could be. Here are the essential controls and methods used.
public class SearchProgressForm : Form
{
private System.Windows.Forms.Button btnSearch = new Button();
private System.Windows.Forms.TextBox txtTextSearch = new TextBox();
private System.Windows.Forms.Label lblSearchFile = new Label();
private System.Windows.Forms.ProgressBar SearchProgressBar = new ProgressBar();
public event LogSearchEventHandler SearchSucceededEvent;
protected void OnSearchSucceeded(LogSearchEventArguments p_eventArguments)
{
LogSearchEventHandler handler = SearchSucceededEvent;
if (handler != null)
{
handler(this, p_eventArguments);
}
}
private void InitializeComponent()
{
this.btnSearch.Name = "btnSearch";
this.btnSearch.Text = "Search";
this.btnSearch.Click += new System.EventHandler(this.btnSearch_Click);
this.lblSearchFile.Text = "Searching File: ";
this.txtTextSearch.Text = "search string";
}
public SearchProgressForm() { }
private void btnSearch_Click(object sender, EventArgs e)
{
this.SearchByText(this.txtTextSearch.Text);
}
private void SearchByText(string p_searchParameter)
{
// Setup a progress report for thr ProgressBar
var _progressBarUpdate = new Progress<int>(value =>
{
this.SearchProgressBar.Value = value;
this.SearchProgressBar.Refresh();
});
var _progressFileNameUpdate = new Progress<string>(value =>
{
this.lblSearchFile.Text = "Searching File For : " + value;
this.lblSearchFile.Refresh();
});
// Start search on a backgroud thread and report progress as it occurs
Task.Run(async () => await this.SearchByStringAsync(p_searchParameter, _progressBarUpdate, _progressFileNameUpdate));
}
private async Task SearchByStringAsync(string p_searchParameter, IProgress<int> p_progressBar, IProgress<string> p_progressFileName)
{
await Task.Delay(1);
TextFileReader textFileReader = new TextFileReader();
LogSearchEventArguments logSearchEventArguments = null;
long _sessionloopCount = 0;
long _totalTextLinesCount = this.GetTotalSearchCount(p_searchParameter, SearchType.TextString);
// Get file names from SQL table
var _logFiles = DataOperations.LogFileSortableList(null);
foreach (var log in _logFiles)
{
// Format a file name to be read from the file system
string _fileName = log.Directory + "\\" + log.FileName;
p_progressFileName.Report(log.FileName);
// If we've raised an event for this file, then stop iterating over remaning text
if (logSearchEventArguments != null)
{
logSearchEventArguments = null;
break;
}
// Read in file contents from file system
List<string> _fileContents = textFileReader.ReadAndReturnStringList(_fileName);
long _fileTotalRecordCount = _fileContents.Count;
long _fileRecordCount = 0;
foreach (var _line in _fileContents)
{
if (_line.ToUpper().Contains(p_searchParameter.ToUpper()))
{
// Raise an event so search parameter and file name can be captured in another form
logSearchEventArguments =
new LogSearchEventArguments
(
"TextSearch", p_searchParameter, SearchType.TextString, true, log,
new DateTime(
Convert.ToInt32("20" + log.FileName.Substring(14, 2)),
Convert.ToInt32(log.FileName.Substring(16, 2)),
Convert.ToInt32(log.FileName.Substring(18, 2)))
);
// We found a match, so no further searching is needed in this log file,
// and it's been flagged in the DB, so raise the event to save search parameter and file name
// then break out of this loop to get the next file to search in.
this.OnSearchSucceeded(logSearchEventArguments);
break;
}
// These calcs are based on actual searches performed
_fileRecordCount++;
_sessionloopCount++;
p_progressBar.Report(Convert.ToInt32((_sessionloopCount * 100) / _totalTextLinesCount));
}
// Because we exit a search as soon as the 1st match is made, need to resynch all counts
// and update the progress bar accordingly
if (_fileRecordCount < _fileTotalRecordCount)
{
long _countDifference = _fileTotalRecordCount - _fileRecordCount;
// Add count difference to sessionLoopCount and update progress bar
_sessionloopCount += _countDifference;
p_progressBar.Report(Convert.ToInt32((_sessionloopCount * 100) / _totalTextLinesCount));
}
}
//Search is complete set Progress to 100% and report before exiting
p_progressBar.Report(100);
// Close the modal SearchForm and exit
this.Close();
}
}
I solved this problem but I'm still not certain of what caused it. I eliminated the method "private void SearchByText(string p_searchParameter)" and moved the code there into the btnSearch_Click event handler so I could call my background worker "SearchByStringAsync" directly from the button click event handler.
I also updated the EFCore NuGet Packages, which were version Net6.0 to version 6.0.4, because of single line of code in my Async background method, "var _logFiles = DataOperations.LogFileSortableList(null)".
That call returned a Sortable BindingList, using BindingList <T>. Between the NuGet updates and a minor change on a custom comparer method in my BindingList <T> class, the windows modal form now updates the ProgressBar and Label text as expected, and the form now responds immediately to user interaction.
I am trying to build a simple app using Microsoft Azure's Cognitive Services Speech To Text SDK in Unity3D. I've following this tutorial, and it worked quite well. The only problem with this tutorial is that the Speech-To-Text is activated by a button. When you press the button, it'll transcribe for the duration of a sentence, and you'll have to press the button again for it to transcribe again. My problem is I'd like it to start transcribing as soon as the program is run in Unity, rather than having to press a button each time I want to transcribe a sentence.
Here is the code.
public async void ButtonClick()
{
// Creates an instance of a speech config with specified subscription key and service region.
// Replace with your own subscription key and service region (e.g., "westus").
var config = SpeechConfig.FromSubscription("[My API Key]", "westus");
// Make sure to dispose the recognizer after use!
using (var recognizer = new SpeechRecognizer(config))
{
lock (threadLocker)
{
waitingForReco = true;
}
// Starts speech recognition, and returns after a single utterance is recognized. The end of a
// single utterance is determined by listening for silence at the end or until a maximum of 15
// seconds of audio is processed. The task returns the recognition text as result.
// Note: Since RecognizeOnceAsync() returns only a single utterance, it is suitable only for single
// shot recognition like command or query.
// For long-running multi-utterance recognition, use StartContinuousRecognitionAsync() instead.
var result = await recognizer.RecognizeOnceAsync().ConfigureAwait(false);
// Checks result.
string newMessage = string.Empty;
if (result.Reason == ResultReason.RecognizedSpeech)
{
newMessage = result.Text;
}
else if (result.Reason == ResultReason.NoMatch)
{
newMessage = "NOMATCH: Speech could not be recognized.";
}
else if (result.Reason == ResultReason.Canceled)
{
var cancellation = CancellationDetails.FromResult(result);
newMessage = $"CANCELED: Reason={cancellation.Reason} ErrorDetails={cancellation.ErrorDetails}";
}
lock (threadLocker)
{
message = newMessage;
waitingForReco = false;
}
}
}
void Start()
{
if (outputText == null)
{
UnityEngine.Debug.LogError("outputText property is null! Assign a UI Text element to it.");
}
else if (startRecoButton == null)
{
message = "startRecoButton property is null! Assign a UI Button to it.";
UnityEngine.Debug.LogError(message);
}
else
{
// Continue with normal initialization, Text and Button objects are present.
}
}
void Update()
{
lock (threadLocker)
{
if (startRecoButton != null)
{
startRecoButton.interactable = !waitingForReco && micPermissionGranted;
}
}
}
I've tried removing the Button object, but then the speech-to-text won't run.
Any tips or advice would be amazing. Thank you.
Per the comments in the script of the tutorial your referenced:
// Starts speech recognition, and returns after a single utterance is recognized. The end of a
// single utterance is determined by listening for silence at the end or until a maximum of 15
// seconds of audio is processed. The task returns the recognition text as result.
// Note: Since RecognizeOnceAsync() returns only a single utterance, it is suitable only for single
// shot recognition like command or query.
// For long-running multi-utterance recognition, use StartContinuousRecognitionAsync() instead.
But it's not as simple as replacing 'RecognizeOnceAsync' with 'StartContinuousRecognitionAsync', because the behaviours are different. RecognizeOnceAsync will basically turn on your mic for a maximum of 15 seconds, and then stop listening.
Instead, make the button into 'should I listen continuously or not?' using StartContinuousRecognitionAsync and StopContinuousRecognitionAsync, and then change your Start function to simply start up a new recognizer and have it waiting for the Speech Recognizer event to come through. Below is the script I used to enable this functionality:
using UnityEngine;
using UnityEngine.UI;
using Microsoft.CognitiveServices.Speech;
public class HelloWorld : MonoBehaviour
{
public Text outputText;
public Button startRecordButton;
// PULLED OUT OF BUTTON CLICK
SpeechRecognizer recognizer;
SpeechConfig config;
private object threadLocker = new object();
private bool speechStarted = false; //checking to see if you've started listening for speech
private string message;
private bool micPermissionGranted = false;
private void RecognizingHandler(object sender, SpeechRecognitionEventArgs e)
{
lock (threadLocker)
{
message = e.Result.Text;
}
}
public async void ButtonClick()
{
if (speechStarted)
{
await recognizer.StopContinuousRecognitionAsync().ConfigureAwait(false); // this stops the listening when you click the button, if it's already on
lock(threadLocker)
{
speechStarted = false;
}
}
else
{
await recognizer.StartContinuousRecognitionAsync().ConfigureAwait(false); // this will start the listening when you click the button, if it's already off
lock (threadLocker)
{
speechStarted = true;
}
}
}
void Start()
{
startRecordButton.onClick.AddListener(ButtonClick);
config = SpeechConfig.FromSubscription("KEY", "REGION");
recognizer = new SpeechRecognizer(config);
recognizer.Recognizing += RecognizingHandler;
}
void Update()
{
lock (threadLocker)
{
if (outputText != null)
{
outputText.text = message;
}
}
}
}
And below is a gif of me using this functionality. You'll not that I don't click the button at all (and it was only clicked once, prior to the gif being recorded)(also, sorry for the strange sentences, my coworkers kept interrupting asking who I was talking to)
I customized the ARPaymentEntry in which it creates a Journal Voucher Entry with created Credit Memo, it retrieves the Credit Memo applies the open invoice that is also applied in the current payment. when I create the instance to call the Credit Memo and add the Invoice in ARAdjust table, an error occurs when trying to insert it, giving a Reference Nbr cannot be found in the system, although when I'm trying to manually applying it I could see the open invoice.
public void ReleaseCreditMemo(string refNbr)
{
try
{
ARPaymentEntry docGraph = PXGraph.CreateInstance<ARPaymentEntry>();
List<ARRegister> list = new List<ARRegister>();
ARPayment payment;
ARRegister invoice = PXSelect<ARRegister, Where<ARRegister.docType, Equal<Required<ARRegister.docType>>, And<ARRegister.refNbr, Equal<Required<ARRegister.refNbr>>>>>.Select(docGraph, ARInvoiceType.CreditMemo, refNbr);
docGraph.Document.Current = PXSelect<ARPayment, Where<ARPayment.docType, Equal<Required<ARPayment.docType>>, And<ARPayment.refNbr, Equal<Required<ARPayment.refNbr>>>>>.Select(docGraph, ARInvoiceType.CreditMemo, refNbr);
payment = docGraph.Document.Current;
list.Add(payment);
foreach (ISARWhTax item in ARWhLine.Select())
{
decimal? _CuryAdjgAmt = payment.CuryOrigDocAmt > invoice.CuryDocBal ? invoice.CuryDocBal : payment.CuryOrigDocAmt;
decimal? _CuryAdjgDiscAmt = payment.CuryOrigDocAmt > invoice.CuryDocBal ? 0m : invoice.CuryDiscBal;
ARAdjust adj = new ARAdjust();
adj.AdjdBranchID = item.AdjdBranchID;
adj.AdjdDocType = ARInvoiceType.Invoice;
adj.AdjdRefNbr = item.AdjdRefNbr;
adj.AdjdCustomerID = item.CustomerID;
adj.AdjdDocDate = invoice.DocDate;
adj.CuryAdjgAmt = _CuryAdjgAmt;
adj.CuryAdjdDiscAmt = _CuryAdjgDiscAmt;
if (docGraph.Document.Current.CuryUnappliedBal == 0m && docGraph.Document.Current.CuryOrigDocAmt > 0m)
{
throw new PXLoadInvoiceException();
}
//This line code below OCCURS THE ERROR
docGraph.Adjustments.Insert(adj);
}
docGraph.Save.Press();
PXLongOperation.StartOperation(docGraph, delegate() { ARDocumentRelease.ReleaseDoc(list, false); });
}
catch (Exception ex)
{
throw new PXException(ex.Message);
}
}
I would look at the selector of the field causing the error ("Reference Nbr.") as having a selector on a field will validate the entered value to the selector's select statement (unless validatevalue=false for the selector). Maybe the selector will give you some pointers as to what is missing or causing the validation to fail.
I figured it out it that after the code below executes it does not immediately updates the View. So what I did is to execute my code at ARPayment_RowSelected event with a conditional statement if the document is released.
PXLongOperation.StartOperation(this.Base, delegate() { ARDocumentRelease.ReleaseDoc(list, false); });
I'm trying to filter some records using the RecordFilter interface. In my app I have a couple of interfaces similar to this one, on which the user can enter an ID or Name (he/she could enter both or neither of them too)
Here's what I've done so far:
The Customer filter.
Here if the user didn't enter an ID, I pass 0 as a default value, that's why I evaluate customerID!=0
public class CustomerFilter implements RecordFilter {
private String mName_Filter;
private int mID_Filter;
public CustomerFilter(String name_Filter, int id_Filter) {
this.mName_Filter = name_Filter.toLowerCase();
this.mID_Filter = id_Filter;
}
public boolean matches(byte[] candidate) {
try {
ByteArrayInputStream bis = new ByteArrayInputStream(candidate);
DataInputStream dis = new DataInputStream(bis);
int customerID = dis.readInt();
String customerName = dis.readUTF().toLowerCase();
if ((customerName != null && customerName.indexOf(mName_Filter) != -1) && (customerID != 0 && customerID == mID_Filter))
return true;
if (customerName != null && customerName.indexOf(mName_Filter) != -1 && customerID == 0)
return true;
if (customerName == null && (customerID != 0 && customerID == mID_Filter))
return true;
if (customerName == null && customerID == 0)
return true;
} catch (IOException ex) {
//What's the point in catching a exception here???
}
return false;
}
}
The search method:
Note: This method is in a class that I call "RMSCustomer", in which I deal with everything related to RMS access. The search method receives two parameters (id and name) and uses them to instantiate the filter.
public Customer[] search(int id, String name) throws RecordStoreException, IOException {
RecordStore rs = null;
RecordEnumeration recEnum = null;
Customer[] customerList = null;
try {
rs = RecordStore.openRecordStore(mRecordStoreName, true);
if (rs.getNumRecords() > 0) {
CustomerFilter filter = new CustomerFilter(name, id);
try {
recEnum = rs.enumerateRecords(filter, null, false);
if (recEnum.numRecords() > 0) {
customerList = new Customer[recEnum.numRecords()];
int counter = 0;
while (recEnum.hasNextElement()) {
Customer cust;
int idRecord = recEnum.nextRecordId();
byte[] filterRecord = rs.getRecord(idRecord);
cust = parseRecord(filterRecord);
cust.idRecord = idRecord;
customerList[counter] = cust;
counter++;
}
}
else{
customerList = new Customer[0];
//How to send a message to the midlet from here
//saying something like "No Record Exists.Please select another filter"
}
} finally {
recEnum.destroy();
}
}
else{
//How to send a message to the midlet from here
//saying something like "No Record Exists.Please Add record"
}
} finally {
rs.closeRecordStore();
}
return customerList;
}
Even though, the code shown above works I still have some questions/problems:
In the Filter :
1) How can I improve the code that evaluates the possible values of the filters (name,id)? What if I had more filters?? Will I have to test all the possible combinations??
2) If the user doesn’t enter neither a ID nor a name, should I display all the records or should I display a message "Please enter a name or ID"?? What would you do in this case?
3) Why do I have to put a try-catch in the filter when I can't do anything there?? I can't show any alert from there or can I?
In the search method:
1) How can I show a proper message to the user from that method? something like "No records" (see the "ELSE" parts in my code
Sorry If I asked too many questions, it's just that there's any complete example of filters.
Thanks in advance
How can I improve the code that evaluates the possible values of the
filters (name,id)?
The ID is the first field in the record and the fastest one to search for. If the Id matches, It doesn't really matter what the customer name is. Normally you'll be looking for the records where the ID matches OR the customer name matches, so once the ID matches you can return true. This is my proposal for the CustomerFilter class:
public class CustomerFilter implements RecordFilter {
private String mName_Filter;
//Use Integer instead of int.
//This way we can use null instead of zero if the user didn't type an ID.
//This allows us to store IDs with values like 0, -1, etc.
//It is a bit less memory efficient,
//but you are not creating hundreds of filters, are you? (If you are, don't).
private Integer mID_Filter;
public CustomerFilter(String name_Filter, Integer id_Filter) {
this.mName_Filter = normalizeString(mName_Filter);
this.mID_Filter = id_Filter;
}
//You should move this function to an StringUtils class and make it public.
//Other filters might need it in the future.
private static String normalizeString(final String s){
if(s != null){
//Warning: you might want to replace accentuated chars as well.
return s.toLowerCase();
}
return null;
}
public boolean matches(byte[] candidate) {
ByteArrayInputStream bis = new ByteArrayInputStream(candidate);
DataInputStream dis = new DataInputStream(bis);
try {
if(mID_Filter != null){
//If the ID is unique, and the search is ID OR other fields, this is fine
int customerID = dis.readInt();
if(mID_Filter.intValue == customerID){
return true;
} else {
return false;
}
}
if(mName_Filter != null){
String customerName = normalizeString(dis.readUTF());
if(customerName != null && customerName.indexOf(mName_Filter) != -1){
return true;
}
}
if(mID_Filter == null && mName_Filter == null){
return true; // No filtering, every record matches.
}
} catch (IOException ex) {
//Never swallow exceptions.
//Even if you are using an underlying ByteArrayInputStream, an exception
//can still be thrown when reading from DataInputStream if you try to read
//fields that do not exists.
//But even if no exceptions were ever thrown, never swallow exceptions :)
System.err.println(ex);
//Optional: throw ex;
} finally {
//Always close streams.
if(bis != null){
try {
bis.close();
} catch(IOException ioe){
System.err.println(ioe);
}
}
if(dis != null){
try {
dis.close();
} catch(IOException ioe){
System.err.println(ioe);
}
}
}
return false;
}
}
What if I had more filters?? Will I have to test all the possible
combinations??
It depends on your project. Usually the ID is unique and no two records exist with the same id. In this case you should explicitly design the screen so that the user understands that either he types an Id, or else he fills in the other fields. The condition would be like this:
idMatches OR (field1Matches AND field2Matches AND ... fieldNMatches)
If the user types nothing, then all records will be returned.
But then again this is more a UX issue, I don't know if it is valid for your requirements.
From the programming point of view, what is clear is that the more fields you add, the more messy your filter will became. To prevent this, you could use patterns like Decorator, Composite, and even Chain of responsibility. You'll probably have to trade good design for performance though.
If the user doesn’t enter neither a ID nor a name, should I display
all the records or should I display a message "Please enter a name or
ID"?? What would you do in this case?
It depends. Is there any other way to view all records? If so, then show the message.
Why do I have to put a try-catch in the filter when I can't do
anything there?? I can't show any alert from there or can I?
You shouldn't. This class is only responsible of filtering, not of interacting with the user. You can still log the error from the catch clause, and then throw the exception again. That will propagate the exception up to RMSCustomer.search, so whatever client code is calling that function will handle the exception in the same way you are handling the other ones thrown by that method. But keep the finally clause to close the streams.
How can I show a proper message to the user from that method?
something like "No records" (see the "ELSE" parts in my code)
You shouldn't do anything related to the GUI (like showing dialogs) from the RMSCustomer class. Even if you are not using the Model-View-Controller pattern, you still want to keep your class focused on a single responsibility (managing records). This is called the Single responsibility principle.
Keeping your class isolated from the GUI will allow you to test it and reuse it in environments without GUI.
The no records case should be handled by the screen when there are zero results. An array of lenght == 0 is fine here, and the screen will show the "No results" message. For other kinds of errors, you can extend the Exception class and throw your own custom exceptions, i.e: RecordParsingException, from the RMSCustomer.search method. The screen class will then map the different exceptions to the error message in the language of the user.
I wrote a method which changes backcolor of the rows before painting gridview in devexpress. It works fine but I realized that my code begins slowing down. Then I've found that the event firing continuously. It never stops. How can I handle this? Is there any way to stop firing event manually after gridview painted or should I try to solve this problem with an another event or another method???
Here is my event:
private void gvStep_CustomDrawCell(object sender, DevExpress.XtraGrid.Views.Base.RowCellCustomDrawEventArgs e)
{
try
{
DataRowView drw = (DataRowView)gvStep.GetRow(e.RowHandle);
byte actionTypeID = (byte)drw.Row["ActionType"];
//string colorCode = (new DivaDs()).GetBackColor(actionTypeID);
string colorCode = divaDs.GetBackColor(actionTypeID);
Color backColor = ColorTranslator.FromHtml(colorCode);
e.Appearance.BackColor = backColor;
}
catch (Exception ex)
{
XtraMessageBox.Show(ex.Message);
}
}
public string GetBackColor(byte actionTypeID)
{
string color = string.Empty;
using (SqlConnection conn = new SqlConnection(ConfigurationManager.ConnectionStrings[DivaSqlSiteConnString].ConnectionString))
{
using (SqlCommand cmd = new SqlCommand(#"Select BackColor from ActionTypes where ID = #actionTypeID"))
{
SqlParameter param = new SqlParameter("#actionTypeID", actionTypeID);
cmd.Parameters.Add(param);
cmd.Connection = conn;
conn.Open();
color = cmd.ExecuteScalar().ToString();
conn.Close();
}
}
return color;
}
My best guess is that some part of your code is just really slow.
The event only fires for each visible cell in the grid. If you attempt to debug the event, focus will shift to the debugger, and when you return to the application the cells need to be redrawn, causing the event to fire again, thus giving the impression that the event fires continuously. It does not, however.
Here are some pointers to improve performance:
You are constructing a new DivaDs every time the event fires
Instead, consider reusing the same instance of the class as a member variable
What happens in the constructor?
Take a closer look at the GetBackColor method or ColorTranslator.FromHtml and see if any modifications can be made to improve performance.
Update
It appears you are querying the database for each cell in the grid. This is a really bad idea.
A simple solution would be to preload all ActionTypes and their background colors (or at least the subset of ActionTypes that is displayed in the grid) before setting the grid's data source.
// member variable
private Dictionary<byte, Color> actionTypeColorDict;
void BuildActionTypeColorDictionary()
{
string connectionString = ConfigurationManager
.ConnectionStrings[DivaSqlSiteConnString].ConnectionString;
using (SqlConnection conn = new SqlConnection(connectionString))
using (SqlCommand cmd = conn.CreateCommand())
using (SqlDataAdapter adapter = new SqlDataAdapter(cmd))
{
// load all action type IDs and corresponding background color:
cmd.CommandText = #"SELECT ActionTypeID, BackColor FROM ActionTypes";
DataTable actionTypeTable = new DataTable();
adapter.Fill(actionTypeTable);
// build a dictionary consisting of action type IDs
// and their corresponding colors
actionTypeColorDict = actionTypeTable.AsEnumerable().ToDictionary(
r => r.Field<byte>("ActionTypeID"),
r => ColorTranslator.FromHtml(r.Field<string>("ColorCode")));
}
}
Call the BuildActionTypeColorDictionary method before setting the data source of the grid. In the RowStyle or CustomDrawCell events, use the new dictionary member to determine the background color. See the following modified version of your RowStyle code:
private void gvStep_RowStyle(object sender,DevExpress.XtraGrid.Views.Grid.RowStyleEventArgs e)
{
try
{
DataRow row = gvStep.GetDataRow(e.RowHandle);
if (row == null)
return;
byte actionTypeID = row.Field<byte>("ActionImage");
// look up color in the dictionary:
e.Appearance.BackColor = actionTypeColorDict[actionTypeID];
}
catch (Exception ex)
{
XtraMessageBox.Show(ex.Message);
}
}
How do you know it's firing continuously? Are you debbuging?
This code runs whenever the grid is redrawn, meaning whenever the form gets focus.
This event runs for each cell - so it will run quite a few times.
If you put a break-point in this event you'll never get out of it. It will break, you will debug, when it's done it will return focus to the form - causing the form to be redrawn using this event and the break-point is reached again.
And just a side note - Whenever I use that event I have to put e.Handled = true; in the code so that the cell isn't "drawn" by anyone but me :)
Finally, I found it. RowStyle event only fires same time with gridview's row count
private void gvStep_RowStyle(object sender, DevExpress.XtraGrid.Views.Grid.RowStyleEventArgs e)
{
try
{
DataRowView drw = (DataRowView)gridView1.GetRow(e.RowHandle);
if (drw == null)
return;
byte actionTypeID = (byte)drw.Row["ActionImage"];
string colorCode = divaDs.GetBackColor(actionTypeID);
Color backColor = ColorTranslator.FromHtml(colorCode);
e.Appearance.BackColor = backColor;
}
catch (Exception ex)
{
XtraMessageBox.Show(ex.Message);
}
}