daboross

Not at all, that would be great! I'm not sure if we want any extra information sent alongside the record to mpsc::Sender - it would be simplest just to send...

It's all good, connecting things can't be too bad. I'll be open to a future PR! If any more information is needed on `fern`'s structure or other bits, I'm definitely...

Support for `std::sync::mpsc::Sender` is now added as of version 0.4.3. I'm still aiming to add direct threading support into `fern`, but it may have to wait until after more changes...

Thanks! I hadn't realized that the test framework's hook was per-thread rather than global. Makes sense now that I think about it, though. Adding that as an example sounds like...

Huh, weird. I believe it should exist, but if you're looking at the default docs.rs documentation it will be built for linux, and exclude the function. Here's what I've looked...

There's a workaround built in for this, the `meta-logging-in-format` feature. Enable it (like with `fern = { version = "...", features = ["meta-logging-in-format"]`), and fern will use an additional buffer...

As an example solution, you might do something like the following: ```rust let colors_line = ColoredLevelConfig::new() .error(Color::Red) .warn(Color::Yellow) .trace(Color::BrightBlack); let colors_level = colors_line.clone().info(Color::Green); fern::Dispatch::new() .level(log::LevelFilter::Warn) .chain( fern::Dispatch::new() .format(move |out, message,...

I think fixing this through detection of output would require significantly rearchitecting fern's logging pipeline. There might be a workaround we could implement, but it really isn't designed to have...

While this might be a good API, I think we'd need to either change it quite a bit or change the way we work internally in order to make this...

My apologies for not replying to this earlier! This is not possible with the current fern, as you have found, yes.