Peter Atashian

It seems to work better for me if I drag it slowly. If I drag it too fast it snaps back.

> Also this RFC proposes that it's not an error to use names that don't exist for the current target (they just evaluate to false). I think that would need...

I am very strongly in favor of removing those panic messages from the binary because they waste space, cannot be stripped, and it is just a poor man's version of...

Last I checked, I did not see an API in dbghelp to get the column number. That said, the only time lack of column numbers would actually be a *regression*...

@est31 Multiple panics on the same line in that case would only apply when the `Add` or `Index` operator is built in. It wouldn't apply to `Vec` indexing, nor would...

After doing some digging, it turns out that dbghelp on Windows *does* allow getting the column. Both `SymGetLineFromAddr64` and `SymGetLineFromInlineContext` have a `pdwDisplacement` parameter which `std` and the `backtrace` crate...

Yay, finally some progress on this front! I haven't dug into the details of this RFC, but based on a quick skim I am generally in favor.

I'm really not sure how to handle the MinGW layout being slightly different than the Windows API layout. That's such an unfortunate situation. If we're going to make any changes...

> #[repr(C)]: The current and existing behavior. packed and aligned fields are banned. > Except that they are (accidentally) allowed when using generics, so this ban makes no sense -...

It's not that we're normalizing paths, it's that we're canonicalizing paths, which results in getting root local device paths back. If all we're doing is printing out paths for the...