Interactively execute code while debugging in Scala - debugging

Is there a method for inserting code into a Scala application while debugging? For example, could I have something like this,
var c = 1.0
0.until(10).foreach{ i =>
if (i == 5) {
startDebuggingMagicHere()
}
}
where I could then inspect and interact with c, i, and any other variable in scope via methods and assignment.

In the Scala-IDE plugin for Eclipse, debugging is supported. Set a breakpoint in the Scala code and you will be able to do limited things. Current support is not as good as that for Java. I believe it is planned to improve by the Summer.
What you can do is:
use the "variables" view to see the current values of variables and modify values,
tell the debugger to "drop to frame" so that it starts the current method call again, using the original values from the stack,
modify code and save the file which causes the debugger to insert the new code and drop to frame, i.e. restart the method call (this does not always work - depending how much you change the code you may need to restart the app, e.g. if you change the class definition)
What you can't do is:
inspect variables from the editor,
write code in the "display" view and execute it. Both of these work with Java, but not Scala.

Related

What is the "=>" operator?

In this link from MDN, it explains how to write a unit test for developing an addon for firefox. However, there are several segments that I don't understand and didn't find any useful result after searching google.
The first one, the following is a paragraph quoted from above link:
In a web page, you can perform Base64 encoding and decoding using the
btoa() and atob() functions. Unfortunately these functions are
attached to the window object: since this object is not available in
your main add-on code, atob() and btoa() aren't available either. So
we'll create a base64 module to expose these functions from the
platform (see Creating Reusable Modules).
What does "the platform" in the above paragraph mean? the "Services.jsm"?
Also in the following code:
const { atob, btoa } = require("resource://gre/modules/Services.jsm");
this makes atob and btoa as one of the Services that is available for other class? or make (constant variables)atob and btoa both reference to the Services.jsm?
The Second one:
what are these two lines of code do?
exports.atob = a => atob(a);
exports.btoa = b => btoa(b);
I understand the part
exports.atob
which enables atob function to be available from other classes outside the "base64.js".
but what does the following mean?
= a => atob(a);
I didn't find that javascript has "=>" operator!
From my understaning, the reason we create a base64.js is because atob and btoa can only be called if we have the window object. Therefore we can use Services.jsm to get a reference of the window object?
For "platform" in the paragraph you quoted they mean the set of functions which is not strictly Firefox browser code but rather implements basic, share functionalities. This usually lives in toolkit/modules in the mozilla-central repository. Services.jsm lives there as well so yes, that's part of the platform. Moreover, atob and btoa are both imported from Services.jsm.
The arrow => in exports.atob = a => atob(a); defines what's called an arrow function: it's a new, shorter syntax to define functions in JavaScript. This SO answer has many useful informations about it.
From my understanding, the reason we create a base64.js is because atob and btoa can only be called if we have the window object. Therefore we can use Services.jsm to get a reference of the window object?
That's almost correct: you need to export the function from the underlying platform as you don't have a window object there. If you had a window object, you would have just done window.atob or something like that. That call would have still called the same function you imported from Services.jsm.
So you're not using Services.jsm to get a reference to the window object, but rather directly importing the needed functions so that you don't need to have a window object.

Language Service: ParseReason.Check never called after migrating to VS2010

I just migrated my language service from VS2008 to VS2010. Everything works fine except for one important thing: I no longer get LanguageService.ParseSource invoked for ParseReason.Check. It do get a single invoke after opening a file. But after editing code, it no longer gets invoked.
Any ideas what could be causing that?
I also migrated a language service from 2008 to 2010. Can you check if you've fallowed all of these steps?
http://msdn.microsoft.com/en-us/library/dd885475.aspx
I didn't have to do anything else, which I verified by diffing the important files in our depot before and after the change.
I don't know if you ever figured your question out, but have you tried making sure that your Source class' LastParseTime is set to 0 when creating it? I seem to recall some issues with Check not happening unless you manually set LastParseTime to 0 when creating your Source object.
Protip: If you use .NET Reflector, you can disassemble all of the base classes for the LanguageService framework and get a pretty good understanding of how it all works under the hood. The classes you'd be interested in live in Microsoft.VisualStudio.Package.LanguageService.10.0.dll, which should be installed in the GAC. I've found this to be unimaginably helpful when trying to figure out why things weren't working in my own Language Service, and being able to step through the source code in the debugger mitigates almost all the pain of working with these frameworks!
When your Source object is initialized, it starts off with a LastParseTime of Int32.MaxValue. The code that causes fires off a ParseRequest with ParseReason.Check checks the LastParseTime value to see if the time since the last change to the text is less than the time it takes to run a parse (or the CodeSenseDelay setting, whichever is greater).
The code that handles the response from ParseSource is supposed to set the LastParseTime, but as far as I can tell, it only does that if the ParseReason is Check.
You can get around this issue by setting Source.LastParseTime = 0 when you initialize your Source. This has the side-effect of setting CompletedFirstParse to true, even if the first parse hasn't finished yet.
Another way to fix this issue is to override Source.OnIdle to fire off the first call to BeginParse() This is the way I would recommend.
public override void OnIdle(bool periodic)
{
// Once first "Check" parse completes, revert to base implementation
if (this.CompletedFirstParse)
{
base.OnIdle(periodic);
}
// Same as base implementation, except we don't check lastParseTime
else if (!periodic || this.LanguageService == null || this.LanguageService.LastActiveTextView == null || (this.IsCompletorActive) || (!this.IsDirty || this.LanguageService.IsParsing))
{
this.BeginParse();
}
}

Copy object values in Visual Studio debug mode

In Visual Studio debug mode it's possible to hover over variables to show their value and then right-click to "Copy", "Copy Expression" or "Copy Value".
In case the variable is an object and not just a basic type, there's a + sign to expand and explore the object. It there a way to copy all that into the clipboard?
In the immediate window, type
?name_of_variable
This will print out everything, and you can manually copy that anywhere you want, or use the immediate window's logging features to automatically write it to a file.
UPDATE: I assume you were asking how to copy/paste the nested structure of the values so that you could either search it textually, or so that you can save it on the side and then later compare the object's state to it. If I'm right, you might want to check out the commercial extension to Visual Studio that I created, called OzCode, which lets you do these thing much more easily through the "Search" and "Compare" features.
UPDATE 2 To answer #ppumkin's question, our new EAP has a new Export feature allows users to Export the variable values to Json, XML, Excel, or C# code.
Full disclosure: I'm the co-creator of the tool I described here.
You can run below code in immediate window and it will export to an xml file the serialized XML representation of an object:
(new System.Xml.Serialization.XmlSerializer(obj.GetType())).Serialize(new System.IO.StreamWriter(#"c:\temp\text.xml"), obj)
Source: Visual Studio how to serialize object from debugger
Most popular answer from https://stackoverflow.com/a/23362097/2680660:
With any luck you have Json.Net in you appdomain already. In which
case pop this into your Immediate window:
Newtonsoft.Json.JsonConvert.SerializeObject(someVariable)
Edit: With .NET Core 3.0, the following works too:
System.Text.Json.JsonSerializer.Serialize(someVariable)
There is a extension called Object Exporter that does this conveniently.
http://www.omarelabd.net/exporting-objects-from-the-visual-studio-debugger/
Extension: https://visualstudiogallery.msdn.microsoft.com/c6a21c68-f815-4895-999f-cd0885d8774f
You can add a watch for that object, and in the watch window, expand and select everything you want to copy and then copy it.
By using attributes to decorate your classes and methods you can have a specific value from your object display during debugging with the DebuggerDisplay attribute e.g.
[DebuggerDisplay("Person - {Name} is {Age} years old")]
public class Person
{
public string Name { get; set; }
public int Age { get; set; }
}
I always use:
string myJsonString = JsonConvert.SerializeObject(<some object>);
Then I copy the string value which unfortunately also copies the back slashes.
To remove the backlashes go here:
https://www.w3schools.com/jsref/tryit.asp?filename=tryjsref_replace
Then within the <p id="demo">Visit Microsoft!</p> element replace the text with the text you copied.
then replace the var res = str.replace("Microsoft", "W3Schools"); line with
var res = str.replace(/\\/g, '')
Run these new changes but don't forget to click the "try it" button on the right.
Now you should have all the text of the object in json format that you can drop in a json formatter like http://jsonformatter.org or to create a POCO you can now use http://json2csharp.com/
ObjectDumper.NET
This is an awesome way!
You probably need this data for a unit test, so create a Sandbox.cs temporary test or you can create a Console App.
Make sure to get NuGet package, ObjectDumper.NET, not ObjectDumper.
Run this test (or console app)
View test output or text file to get the C# initializer code!
Code:
[TestClass]
public class Sandbox
{
[TestMethod]
public void GetInitializerCode()
{
var db = TestServices.GetDbContext();
var list = db.MyObjects.ToList();
var literal = ObjectDumper.Dump(list, new DumpOptions
{
DumpStyle = DumpStyle.CSharp,
IndentSize = 4
});
Console.WriteLine(literal); // Some test runners will truncate this, so use the file in that case.
File.WriteAllText(#"C:\temp\dump.txt", literal);
}
}
I used to use Object Exporter, but it is 5 years old and no longer supported in Visual Studio. It seems like Visual Studio Extensions come and go, but let's hope this NuGet package is here to stay! (Also it is actively maintained as of this writing.)
Google led me to this 8-year-old question and I ended up using ObjectDumper to achieve something very similar to copy-pasting debugger data. It was a breeze.
I know the question asked specifically about information from the debugger, but ObjectDumper gives information that is basically the same. I'm assuming those who google this question are like me and just need the data for debugging purposes and don't care whether it technically comes from the debugger or not.
I know I'm a bit late to the party, but I wrote a JSON implementation for serializing an object, if you prefer to have JSON output. Uses Newtonsoft.Json reference.
private static void WriteDebugJSON (dynamic obj, string filePath)
{
using (StreamWriter d = new StreamWriter(filePath))
{
d.Write(JsonConvert.SerializeObject(obj));
}
}
I've just right clicked on the variable and selected AddWatch, that's bring up watch window that consists of all the values. I selected all and paste it in a text a text editor, that's all.
Object Dumper is a free and open source extension for Visual Studio and Visual Studio Code.
"Dump as" commands are available via context menu in the Code and Immediate windows.
It's exporting objects to:
C# object initialization code,
JSON,
Visual Basic object initialization code,
XML,
YAML.
I believe that combined with the Diff tool it can be helpful.
I'm the author of this tool.
if you have a list and you want to find a specific variable:
In the immediate window, type
myList.Any(s => s.ID == 5062);
if this returns true
var myDebugVar = myList.FirstOrDefault(s => s.ID == 5062);
?myDebugVar
useful tips here, I'll add my preference for when i next end up here asking this question again in the future.
if you don't mind adding an extension that doesn't require output files or such there's the Hex Visualizer extension for visual studio, by mladen mihajlovic, he's done versions since 2015.
provides a nice display of the array via the usual magnifine glass view object from the local variables window.
https://marketplace.visualstudio.com/items?itemName=Mika76.HexVisualizer2019 is the 2019 version.
If you're in debug mode, you can copy any variable by writing copy() in the debug terminal.
This works with nested objects and also removes truncation and copies the complete value.
Tip: you can right click a variable, and click Copy as Expression and then paste that in the copy-function.
System.IO.File.WriteAllText("b.json", page.DebugInfo().ToJson())
Works great to avoid to deal with string debug format " for quote.
As #OmerRaviv says, you can go to Debug → Windows → Immediate where you can type:
myVariable
(as #bombek pointed out in the comments you don't need the question mark) although as some have found this limits to 100 lines.
I found a better way was to right click the variable → Add Watch, then press the + for anything I wanted to expand, then used #GeneWhitaker's solution, which is Ctrl+A, then copy Ctrl+C and paste into a text editor Ctrl+V.

VB6.0 : initialize method of a User Control called when loading a VB project

Whenever we load a VB project it will call Initialize event of a User Control ( if there is any in the project). My problem is that is that I have some code in UserControl_Initialize that will try to create instances of other COM objects. On my build machine those controls are not registered. One option is to move the code to some other method other than Initialize but I want to know if there is a better solution? Somewhere I found that we may have a check to see if the calling application is VB editor then skip the initialization code...
You can use:
If Not Me.DesignMode Then
...
End If
An other solution we used was a little function which can be used globally:
Public Function IsRuntime() as Boolean
On Error Goto NotRuntime
Debug.Print(1 / 0)
IsRuntime = True
Exit Function
NotRuntime:
IsRuntime = False
End If
Don't know if it is syntactically well formed, but the idea should be clear:
Only in the IDE the debug statement is called.
This only happens if your project was saved with the form designer open: this means that at startup the form is displayed (maybe in the background) and consequently, all controls on it need to be initialized. Hence, your user control initializer is called if this control is used on the form.
To prevent this, simply save the project with the form designer closed.

Programatically moving through a ListView in Qt (Ruby)

I'm making a small file-browser for my own use, in Ruby, and using Qt for the view. The idea is that it'll end up on my TV, where I can use the remote to move up and down and launch files.
Everything works fine, until I'm going to move the selection using the remote. I managed to set up a D-Bus service, so I'll just call the methods using LIRC.
The code I'm using for setting up the view looks like this:
#dm = Qt::DirModel.new
#sm = Qt::ItemSelectionModel.new(#dm)
#lv = Qt::ListView.new
#lv.model = #dm
#lv.selectionModel = #sm
cwd = #dm.index(#dir)
#lv.rootIndex = cwd
And then I'm unsure how to change the selection. Think I must have tried about every setIndex, setSelection and every method sounding similar, on the DirModel, ItemSelectionModel and ListView, without any luck. I've been googling and reading through the API without finding anything.
Ideally, I would have something like "moveSelectionDown" and "moveSelectionUp" that takes care of it, and making sure it wraps around correctly. But I can't seem to find anything.
Managed to fix it through the ItemSelectionModel every view apparently has.
moving up:
curIndex = #lv.currentIndex
#lv.selectionModel.setCurrentIndex(curIndex.sibling(curIndex.row-1, 0), Qt::ItemSelectionModel::ClearAndSelect)
or adding one to move down
I think you're forgetting that you have to create the ModelIndex through your model:
#dm.index(3, 0, None)
I'd try this method (Though I'm not really sure, this deselects the other cells.)
#lv.setCurrentIndex(#dm.index(3, 0, None))
I haven't used Ruby for ages, so I'm not exactly sure there's None.

Resources