Well done Mackinnon, thanks for hosting over the last couple of years. It has been good watching become more confident with time. All the best in your new role.
Upgrading to .NET 9 completely breaks the Hot Reload experience for a Blazor developer. It went from being usable to almost completely unusable and interfering with productivity. The only change to our codebase was updating the TFM version and upgrading the NuGet packages. We are considering the idea of rolling back to .NET 8 if this hot reload functionality isn't repaired soon.
This is funny. Because I see hot reload fixed finally in .NET9. It is not ideal but it works. Watch window also works mich better without annoying "cannot evaluate the variable". Maybe try to reinstall the whole VS? I did that. I still have problems with auto complete on razor pages. It is much better now but VS still cannot find controls by namespace. I type
Hot Reload has never worked reliably enough to rely on in .Net. 90% of the time it works correctly, 5% of the time it doesn't work and tells you it doesn't, but 5% of the time it tells you it works but doesn't. Before that last 5% is converted into either "Fails and tells you" or ideally "Works" then I can't trust it (Why does this change not fix my issue; it really should") and I'll just work as I always have.
A suggestion for Blazor team: please consider borrowing htmx concepts for Blazor SSR. Until then I will try to create some components (e g. HxAnchor, HxInput, HxButton that wrap the enhanced forms). Keep up the great job with Blazor Wasm & SSR !
Actually HTMX can easily replace enhanced forms, and I do prefer that way as I don't need to include the Blazor script at all. The reason I say so is if you really want to use Blazor and the HTMX approach, you will not want to have any Blazor interactivity active (at least in my opinion) As I described in another comment, the real issue here is SSR has serious limitation with forms that are not bound to simple POCO objects (only primitives), that is a big problem they should be tackling
Htmx is what HTML really should have been. You'll need no Blazor for that, as your API can directly answer in plain HTML. And yes, Razor isn't really handy here, too. I happen to find VB's XML literals being much more powerful. To make XML literals generate and answer in HTML is an extension method six lines long. The result is parsable and rewritable markup, and that should be #1 on the list.
@McZsh well, blazor could enable htmx like behavior and blazor SSR components are ideal for html templating. I'd love this to be part of Blazor's SSR approach.
Hiring more cooks doesn't make a better soup. Pouring a lot of resources into a project doesn't always lead to better quality. You can put the brightest people into some work but it takes time for them to get into it. So it is tradeoff between quantity and quality within a specific timeframe when that is expected to be delivered. Which determines the scope. I know developers usually don't want to think about that. The management side, in particular resource management and constraints. And yes. It is a difference between startup and enterprise also.
“Involving the community” seems like a euphemism for exploiting unpaid labor, where Microsoft benefits from others’ expertise without fair compensation-essentially taking ideas and work for free under the guise of engagement.
Please be just honest and tell everybody blazor has no future. I would like they would somehow replace nodejs so everybody can use every javascript framework with .NET. And I mean not just work as a proxy
The big advantage of blazor is blazor server. All the wasm issues and feature requests to me are massively slowing down progress. Can’t they just use React?
I would strongly affirm the contrary. Wasm should be top priority. It's the future fabric of web. People are smart. All community projects that were showcased were done in Webassembly!
Well done Mackinnon, thanks for hosting over the last couple of years. It has been good watching become more confident with time. All the best in your new role.
Upgrading to .NET 9 completely breaks the Hot Reload experience for a Blazor developer. It went from being usable to almost completely unusable and interfering with productivity. The only change to our codebase was updating the TFM version and upgrading the NuGet packages. We are considering the idea of rolling back to .NET 8 if this hot reload functionality isn't repaired soon.
This is funny.
Because I see hot reload fixed finally in .NET9. It is not ideal but it works.
Watch window also works mich better without annoying "cannot evaluate the variable". Maybe try to reinstall the whole VS? I did that.
I still have problems with auto complete on razor pages. It is much better now but VS still cannot find controls by namespace. I type
It's breack too with descktop apps
Hot Reload has never worked reliably enough to rely on in .Net.
90% of the time it works correctly, 5% of the time it doesn't work and tells you it doesn't, but 5% of the time it tells you it works but doesn't.
Before that last 5% is converted into either "Fails and tells you" or ideally "Works" then I can't trust it (Why does this change not fix my issue; it really should") and I'll just work as I always have.
Good session. Loads of interesting questions answered. Will be looking for the next one.
Interesting to see the internal decisionaoing process. Thanks for this
I think for me we need a more advanced demo maybe not as complex as e shop but a mnmal project kinda like how win ui gallery desktop works
A suggestion for Blazor team: please consider borrowing htmx concepts for Blazor SSR.
Until then I will try to create some components (e g. HxAnchor, HxInput, HxButton that wrap the enhanced forms).
Keep up the great job with Blazor Wasm & SSR !
Actually HTMX can easily replace enhanced forms, and I do prefer that way as I don't need to include the Blazor script at all. The reason I say so is if you really want to use Blazor and the HTMX approach, you will not want to have any Blazor interactivity active (at least in my opinion) As I described in another comment, the real issue here is SSR has serious limitation with forms that are not bound to simple POCO objects (only primitives), that is a big problem they should be tackling
Htmx is what HTML really should have been. You'll need no Blazor for that, as your API can directly answer in plain HTML. And yes, Razor isn't really handy here, too.
I happen to find VB's XML literals being much more powerful. To make XML literals generate and answer in HTML is an extension method six lines long. The result is parsable and rewritable markup, and that should be #1 on the list.
@McZsh well, blazor could enable htmx like behavior and blazor SSR components are ideal for html templating.
I'd love this to be part of Blazor's SSR approach.
Hiring more cooks doesn't make a better soup. Pouring a lot of resources into a project doesn't always lead to better quality. You can put the brightest people into some work but it takes time for them to get into it. So it is tradeoff between quantity and quality within a specific timeframe when that is expected to be delivered. Which determines the scope. I know developers usually don't want to think about that. The management side, in particular resource management and constraints. And yes. It is a difference between startup and enterprise also.
They’ll never get hot reload working..
@danilel roth @ 12:12 say it that nickchapsas nailed you recently
“Involving the community” seems like a euphemism for exploiting unpaid labor, where Microsoft benefits from others’ expertise without fair compensation-essentially taking ideas and work for free under the guise of engagement.
Yeah, considering just how much money MS has, they do seem to run things a bit thin. Maybe that's why they have lots of money!
Please be just honest and tell everybody blazor has no future.
I would like they would somehow replace nodejs so everybody can use every javascript framework with .NET. And I mean not just work as a proxy
The big advantage of blazor is blazor server. All the wasm issues and feature requests to me are massively slowing down progress. Can’t they just use React?
I would strongly affirm the contrary. Wasm should be top priority. It's the future fabric of web.
People are smart. All community projects that were showcased were done in Webassembly!
Yeah, I agree. Blazor Server is the big advantage. Use cases for WASM are actually quite few.