Description

@Hejle has pointed out that our current support for dumping Newtonsoft.Json and System.Text.Json objects are lacking.

Examples/reproduction:

Suggested solution:

We can either use DumpConfig.Default.AddCustomTypeHandler(typeof(JValue), (obj, _, _, _) => obj.ToString()); or we can create a custom type renderer for Json objects.

I think we should compile a list of objects of both Newtonsoft.Json and System.Text.Json that we would have to custom render and decide how to render each one of them.
We may start with a simple solution, which is using AddCustomTypeHandler until we implement a proper custom renderer.

In the same way LINQPad will put it's label at the beginning of Dump output.

¦ Heading
¦ ╭───────────────╮
¦ │ "Output Data" │
¦ ╰───────────────╯

When using EF Core Lazy Loading Dump() will load the whole tree.
Using MaxDepth = 1 helps, but then you will see a lot of [Exceeded max depth 1] in the console.
Is there a way to limit the library to ignore virtual members.

I am running .NET Framework 4.8. No choice due to VSTO not being upgraded by Microsoft.

The package fails on nuget due to 'You are trying to install this package into a project that targets '.NETFramework,Version=v4.8', but the package does not contain any assembly references or content files that are compatible'. Not sure but this should compile to .NET Standard 2.0.

I can download and compile myself. Thanks for introducing me to Spectre.Console. Did not know about that.

Dumping a char doesn't display correctly, as a guess it looks like it is supposed to show a table with the character and then it's ascii value but on only the table headers are shown.

Dumpify currently uses MSTest for unit testing. However, MSTest is basically an unmaintained unit testing framework. It aggregates with no actual value compared to more up-to-date and robust unit testing frameworks such as xUnit.

It'd be a strategic move that will provide long-term benefits for the project. And will ensure that it stays on top of best practices regarding to unit testing.

I'm "re-using" Dumpify as a way to do a quick'n'dirty representation of a 2D array of char values for a BattleShip concept. It's just a temporary usage for a concept presentation. But I've noticed a very rare and unpredictable messing up of the border characters every once in a while when I run my code (even without code changes). Here's what I'm seeing.

When it happens, it's not always those incorrect characters - sometimes it's different ones. Again, this is quite rare and hard to reproduce. The code usage is quite simple (.NET 8 console app, running in PowerShell, on Windows 11):

void Concept()
{
    char[,] grid = new char[10,10];
    // Initialize
    for(int row = 0; row < 10; row++)
        for(int col = 0; col < 10; col++)
            grid[row,col] = Random.Shared.Next(10) switch
            {
                < 7 => ' ',
                >= 7 and < 9 => '-',
                _   => Random.Shared.Next(5) switch
                {
                    < 4 => '-',
                    _ => 'H'
                }
            };

    grid.Dump("Player Won");
}

I was trying to dump a NewtonSoft Json object, when I ran into the following error:
Unable to cast object of type 'd__64' to type 'System.Collections.IEnumerable'.

I tried the same with System.Text.Json and got a familiar exception:
Unable to cast object of type 'd__18[System.Text.Json.Nodes.JsonNode]' to type 'System.Collections.IEnumerable'.

[Failed to Render Newtonsoft.Json.Linq.JObject - {
  "key1": "value1",
  "key2": "value2"
}]. Unable to cast object of type '<GetEnumerator>d__64' to type 'System.Collections.IEnumerable'.
[Failed to Render System.Text.Json.Nodes.JsonObject - {
  "key1": "value1",
  "key2": "value2"
}]. Unable to cast object of type '<GetEnumerator>d__18[System.Text.Json.Nodes.JsonNode]' to type 'System.Collections.IEnumerable'.

I have created a fix for this which I will create a pullrequest for in a bit.

Note:
Even with the fix, I wouldn't say that Json is Supported.
The output after my Pull Request would look like this:

To get NewtonSofts output to look nice, it is possible to add a CustomRenderer like this, which will format the output nicely:

DumpConfig.Default.AddCustomTypeHandler(typeof(JValue), (obj, _, _, _) => obj.ToString());

       JObject
┌────────┬──────────┐
│ Key    │ Value    │
├────────┼──────────┤
│ "key1" │ "value1" │
│ "key2" │ "value2" │
└────────┴──────────┘

In case there is much to dump on a single line I saw values being continued in next line. It would be good to be able to allow that all stuff is rendered to one line only. Below image shows the issue

Markdown is very popular format and support by GitHub. I propose to add the possibility of output to it.

if you create a public class with public fields like
public class MyClass { public uint MyNumber; }
and call Dump() on it you just get an empty table with the header
but if you change the fields to properties it will work, i.e
public class MyClass { public uint MyNumber { get; set; } }

Greetings,

Is there any hard rule to keep .NET dependencies >= .NETStandard,Version=2.1.
Can this support Version=2.0 too?

Dumpify is incredibly useful for dumping out messages for console applications. A great addition would be support for HTML output. The generated markup can be a table with simple classes so it can be easily styled.

Having just tested the MembersConfig options, i ended up with
myObject.Dump(null, null, null, null, null, null, membersConfig);
Would Dump benefit by a DumpifySettings class similar to XmlSettings and others used in .NET?

Yes I know there is the global setting =)

Some sort of configuration option to eliminate double-quotes for simple string output.

"This value".Dump()

to look like

This value

instead of

"This value"

It would be very nice if this lib have an extensions method for enumerable that display a dynamic progressbar.

var list = new List<int> {1,2,3,4};
foreach(var i un list.WithProgess("Caption"))
{
    // Do slow stuff
}

Dumpify Assemblies should be signed.

One of the handy things with LinqPad .Dump() is when you use it on a collection, you get a table with columns for each property. e.g. in LinqPad the code:

void Main()
{
	var people = new[] 
	{
		new Person("Marie", "Curie"),   
		new Person("Albert", "Einstein"),
		new Person("Ada", "Lovelace"),
		new Person("Cleopatra", "Philopator"),
		new Person("Nikola", "Tesla")
	};

	people.Dump();
}

record Person(string FirstName, string LastName);

gives:

but with Dumpify, it shows:

So I'd like to be able to have the more concise column-for-property format with collections.

Dump() does print out DirectoryInfo object but the recursive nature of the references (JsonSerializer just fails on it so that's good!) makes a few references up or down the directory tree make the panels start appearing very squashed.
I don't know if it is the aim of Dumpify to remain generic for all objects of if special consideration be given for certain types?
If so I think there are many improvements that could be made for DirInfo, like not printing all info for recursive reference and just the dirname, maybe put it in a tree instead of nested panels etc..

error: NU1108: Cycle detected.
error: dumpify -> Dumpify (>= 0.6.5).
info : Package 'Dumpify' is compatible with all the specified frameworks in project 'C:\Users\USER\source\repos\dotnet\dumpify\dumpify.csproj'.
error: Value cannot be null. (Parameter 'path1')

I'm getting this error while trying to install dumpify both in vs code and in visual studio.