Integrated code formatter

[Adam D Ruppe]

On Tuesday, 18 January 2022 at 19:39:11 UTC, Walter Bright wrote:
> Please don't make us guess what you're referring to. If you 
> want to bring up a specific issue here, please provide a link.

I gave titles there, the search ought to find the link instantly.

But, ok, here's a few more:

https://issues.dlang.org/show_bug.cgi?id=3818

*12 years old*

https://issues.dlang.org/show_bug.cgi?id=8102

*10 years old*

Both of those are examples of where the compiler just vomits 
irrelevant messages. If you know the compiler already, the first 
message is actually useful enough - you know to look at the first 
one and the line it gives or the one immediately before.

But it still could just do something better.

And here's another big one:

https://issues.dlang.org/show_bug.cgi?id=5096

*11 years old*

And hey, look, a comment I left there two years ago - often, when 
I go to put it in bugzilla and find something already there, I'll 
comment instead of making a duplicate.

Those first ones are fairly minor, I've gotten used to them and 
they only take a few seconds. But that brace mismatch one 
actually eats a *significant* amount of time every time it 
happens.

At least the overload and template constraint messages have 
gotten better recently though they're still pretty bad. And those 
opDispatch messages are so unbelievably bad it sometimes doesn't 
even issue them and just generates invalid code instead! 
<https://issues.dlang.org/show_bug.cgi?id=20808> (2 years old - 
and it worked before that, hence why it is a regression) But even 
when it does issue an error, it is just "no such property" which 
is well-known.

Or the new attributes issue errors at the wrong place 
<https://issues.dlang.org/show_bug.cgi?id=17374> (5 years old!) 
making using them with inferred templates much harder than it 
needs to be.

And it isn't just me getting these - error messages come up a lot 
in new user support.