Spark - Discussions Enabled

[westonwalker]

Hi all,

Thanks for checking out the Blazor Spark project.

This is a place to discuss ideas, the future of Spark, and potential issues you are having.

Cheers! - Wes

[sturlath]

Love it! I have enabled notification on this repo so Ill get to follow everything! 😎

[mariobot]

Sounds very interesting the project. I´ll be following the changes to continue updated.

[westonwalker]

Thanks Mario. I'll try to keep the updates coming.

[AndrewJesse]

Hey saw your post on reddit today. I have been looking for a project like this!

[westonwalker]

Thanks for the support!

[AndrewJesse]

my pleasure. I am new to open sourcing. I forked the repo and make a pull request on the readme. Let me know if I am doing anything wrong

[okelot]

Hey @westonwalker

Thanks for creating this wonderful project. I always wanted to have something similar to Django admin in dotnet ecosystem as well. Spark can potentially be that thing. Do you have product roadmap in mind? I will try to contribute as much as I can.

[westonwalker]

Thanks for the kind words

I have somewhat of a roadmap on the docs site - https://blazorspark.com/docs/introduction/

But it's not complete. Small things like user password resets and a user settings page baked into the scaffolding aren't on the roadmap yet but I plan on adding them (or accepting contributions).

I'm not very familiar with Django. What does the admin part of it do?

[kevingolding2001]

I really like where this is going. And I like the way you have structured your code with the service layers etc.

There are a couple of little things I notice need fixing. How can I contribute these?

[kevingolding2001]

It looks like you are accepting pull requests. I'll try creating one.

[kevingolding2001]

Support for VS Code? (or at least get rid of the error message in Linux)

I prefer doing development on Linux, which means using VS Code instead of Visual Studio. However the 'spark open' command doesn't work very well/at all, depending on the OS.

On Windows, setting the association to use VS Code for .sln files results in the spark open command opening the solution file as a text file, but does not actually load the rest of the solution. This is because VS Code works a bit differently. With VS Code, you open the folder that the solution file is in, and then it scans for things that look like projects/solutions/etc.

On Linux, currently it just crashes with this error message: gio: file:///home/kg0/SparkApp%5CSparkApp.sln: Error when getting information for file “/home/kg0/SparkApp\SparkApp.sln”: No such file or directory (because of the back slash).

One option is to add support for VS Code, in which case the 'open' command would run the 'code' program with the folder name as a parameter. But presumably this would require another configuration parameter to specify the choice of editor. I'm not really sure if it's worth adding yet another thing to configure just to support what might be a fairly limited use case.

I'm happy to try and implement the change, but I just wanted to get some feedback as to whether it's worth it.

But in any case, the backslash should be changed to use Path.DirectorySeparatorChar just to stop the command from crashing on Linux.

[westonwalker]

Yeah, the spark open command has only been used and tested on Windows where the default application for a .sln is Visual Studio.

I would accept a pull request if it helped support other Operating Systems or code editors without breaking current functionality.

I know that's hard when it depends on owning multiple machines with different operating systems.

[kevingolding2001]

I run Linux in a VM on Windows on my work PC, so would be testing on both Windows and Linux. For completeness I could probably also try it out on my MacBook.

Do you have an preference for what the configuration option should be called?

[westonwalker]

Hmmm now that you mention it, I'm not sure about vscode. Like your saying, their is already the code command in the terminal to open a folder in vscode. The spark open command was really just a convenient way for me to open projects in Visual Studio since there was no equivalent to the code command for vscode.

I think I'd rather leave spark open the way it is now. Maybe updating the docs to explain how to open it in different editors would be the easier and better route.

[kevingolding2001]

Yeah, I was thinking the same thing when trying to decide if it was really worth it.

Can you at least merge in my pull request to use Path.DirectorySeparatorChar to stop it from crashing.

[kevingolding2001]

How is the Storage folder populated when creating a new App?

I had an issue today where an app I created, then pushed to BitBucket and pulled down onto the dev server crashed when I tried to run it.

The problem was that the Storage/Mail folder did not exist. This is probably because the top level .gitignore is set to ignore all .log files, so when I pushed the project to BitBucket it did not include the Storage/Mail folder because all the contents were ignored (*.log).

The Storage/Logging folder exists in my repo because of the existence of a .gitignore file, which basically says "ignore everything except me".

Obviously the way to fix this problem is to add the same .gitignore file to Storage/Mail.

But I can't seem to see where these folders are created in the first place.

[westonwalker]

The Storage/Mail folder is a part of the template project that generates a new Spark project.

I see why it's so confusing now.

On my local machine, I build the Nuget for the Spark Template project, which has the Storage/Mail folder and log file. But since it's missing that .gitignore file in the directory, it's not getting pushed to the repo and you can't find it. I'll go ahead and make a quick fix for this today.

[kevingolding2001]

What is ApplicationDbContextModelSnapshot.cs in the Migrations folder?

If I want to get rid of all the migrations and start again with a clean slate, should I also delete this file?

[westonwalker]

It's used by entity framework to keep track of the current state of your models when running new migrations.

If you wanted to start from a clean slate, you would want to delete the snapshot file and all of the migration files in the Migrations folder.

If you wanted to change your database provider, you would also have to delete all of those files and re-run your migrations since the migration code can change depending on the database provider you are using.

[kevingolding2001]

Hi there. thanks for the reply.
I deleted the snapshot file along with all the rest, did a 'spark make migration reset' and 'spark migrate' and it seemed to work okay.
However I'm still not sold on the EF migrations idea just because it allowed me to get into an impossible migration path to begin with.
But I've bought a book about EF Core, so will reserve judgement till I've finished reading the book.
Thanks
Kevin

BTW. Did you see my pull request in the docs?

[westonwalker]

Thanks for giving me a heads up on the pull request. Approved.

[westonwalker]

As mentioned in another discussion thread, BlazorSpark is being renamed to Spark.NET. This is mainly because I want to support other ASP.NET project types. Particularly MVC, Razor Pages, MVC API and Blazor SSR.

These changes will be additive and won't take away from using Blazor if you choose.

Over the next couple weeks I'm going to work on adding new templates for MVC and Razor pages.

Because you can't rename Nuget packages, I had to deprecate the old Nuget packages and upload new ones. You can find the 3 Nuget packages here:
https://www.nuget.org/packages/Spark.CLI/
https://www.nuget.org/packages/Spark.Library/
https://www.nuget.org/packages/Spark.Templates/

The documentation website url has also been updated:
https://spark-framework.net/

On the docs site, there is an upgrade guide to move from BlazorSpark to Spark.NET. It's super simple:
https://spark-framework.net/docs/upgrade-guide/#upgrading-blazorspark-to-sparknet

[JAzevedoDev]

Hello, thank you for the "Spark" project. I have just one question: does it make any sense to use MudBlazor with Spark? Is there any limitation or obstacle?

[westonwalker]

I haven't tried it but I don't see a reason why it wouldn't work. As far as making sense, sure if that is what you like to use then yes.

[retinqdon]

Hi

I am following your blog video. But i don't use the pico css, i go with the default, tailwind. But it seems tailwind is not working. In my create form, the form is messed up.
Can you help me with this.
Thanks

[JAzevedoDev]

MudBlazor with Spark
If someone is interested in using Spark with MudBlazor:

https://github.com/JAzevedoDev/SparkMudBlazor

[westonwalker]

Well done!

[d31tcjg]

Hey there! I started building my clubs website using this framework and its awesome!

Question for authentication how can I extend it using oauth or enabling an email verification?

[JAzevedoDev]

Did you run it in IIS (not in VS)?

[westonwalker]

Cool to hear Cesar.

EMAIL VERIFICATION

This is something I want to add to Spark out of the box but haven't got there.

Spark does send a welcome email to new users out of the box however. It does this via a broadcast listener which is called in the register code. You can find this listener at /Applicaiton/Events/Listeners/EmailNewUser.cs

To turn this into email verification, a couple things would need to happen.

1. The user model needs a new property to store a random GUID as string and also a boolean column called IsVerified.
2. Add a page to your app called verify-user or something along those lines.
3. In the welcome email, add a link to the verify-user page with the GUID as a query string parameter.
4. Add logic to the verify-user page to pull the GUID and mark the User as IsVerified.

There is probably some more devil in the details but at a high level that is what needs done.

OAUTH

Out of the box, Spark uses cookie/session auth. If you wanted to change this to some oauth implementation, you would have to alter the login and register logic. You would also need to alter the CookieService which is what verifies the user is logged in whenever they try to access a authorized route.

Depending on what project type you created, you can find the login/register logic in different places.
For blazor: /Pages/Auth/Login.cshtml + /Pages/Auth/Register.cshtml
For MVC: /Applicatoin/Controllers/AuthController.cs

The cookie service logic is always in the same place regardless of project type:
/Application/Services/Auth/CookieService.cs

[d31tcjg]

@JAzevedoDev yes, I am running it on my own IIS server where I self host a couple of sites. I have put a test version of the blazor spark app I am working on and have been successful in running it.

@westonwalker thank you for the guidance! I think this is the direction I needed, I will first be working on getting some form of email verification working, then purse OAuth with Discord.

On a side note, I started developing my site with the default sqlite server and now have migrated to an instance of MSSQL server I am running and im facing some issues with queries hanging when calling the db.

[westonwalker]

If you switch databases, make sure to recreate your migrations as they will change depending on the db type you are using.

So delete the /Application/Database/Migrations folder.

Then recreate + run your migrations with

• spark make migration MigrationName
• spark migrate

You may have did this but just wanted to mention it. I've seen others get tripped up by that.

[westonwalker]

FYI, I'm going to work on setting up a Discord server this week. Stay tuned :)

[westonwalker]

Spark now has an official Discord! https://discord.gg/avRzNqtY

[chadrickm]

So in my command prompt I'm able to create the solution and get it to open. When I try to make the migration it is failing with the following error message:

C:\PoCs\SparkFramework\BlazorPicoApp>spark make migration Initial

Creating a new migration

Could not execute because the specified command or file was not found.
Possible reasons for this include:

• You misspelled a built-in dotnet command.
• You intended to execute a .NET program, but dotnet-ef does not exist.
• You intended to run a global tool, but a dotnet-prefixed executable with this name could not be found on the PATH.

---

Blazor Spark migration Initial created

Now before I run the make migration command I see the Migrations folder under Database is marked in vs as a folder that is removed from the project.

Is there a step I need to do before I run this make migration command?

[chadrickm]

Wait, I didn't see the comment about making sure I had the EntityFrameworkCore.Tools installed. Trying again.

Yep, that was it. I don't recall seeing that in the docs, but I do see it on the Readme here in github.

Update: No the docs are correct, I'm just blind and skipped right over the top of it.

[westonwalker]

Glad it worked for you. Let me know if you have anymore questions.

[keyse]

I FINALLY had to chance to create my first Spark Blazor with Tailwind and it worked awesome :)

Looks like tailwind styling is working fine however, One question I have is how is Spark.NET is applying the tailwind utilities since I can't see the tailwind.config file in the project?

Is Spark.NET calling the tailwind cli behind the scenes or should I install tailwind cli per their site's documentation to get rid of VS underlining the utility class??

[westonwalker]

Out of the box it uses the Tailwind CDN.

Your free to install the CLI if you'd like.

I'd like to switch the framework over to the CLI eventually. I'm still working on a setup that would include Vite for bundling frontend assets (like tailwind) on dotnet build. If anyone has experience with this, let me know!