Ross A. Baker
Ross A. Baker
This relates to #407. For arbitrary `A`, we don't necessarily have a way to concatenate `A` in a reasonable way. I'd say it should depend on a `Monoid[A]`, but that...
Catching up on backlog, and I stand by my previous comment.
I agree that you're right that sending multiple is bad. I also think it's annoying, because multiple Vary headers are plausible and your semigroup instance is reasonable _semantically_. We could...
It's a [limitation of the JDK8 client]( https://github.com/openjdk/jdk/blob/70e5173968cdc245b328b3f3cf3e91786392c5ef/jdk/src/share/classes/sun/net/www/protocol/http/HttpURLConnection.java#L194), which [appears to be fixed](https://bugs.openjdk.java.net/browse/JDK-8212692) in later JDKs.
Yeah. It's a shame, because this kind of thing is exactly what that JDK client should be good for. But Ember will remain in the core repo, so that's the...
Okay, great, I'm glad the bug is exposed. A similar problem has come up recently, where people want to peek into a stream before committing to a status code. Here's...
In Spring, they have [two ways](https://docs.spring.io/spring-security/site/docs/3.2.0.CI-SNAPSHOT/reference/html/csrf.html#csrf-multipart). It has been many years since I've done Spring, so I'm not sure precisely what they're doing with that `MultipartFilter`. The second option, putting...
Just a note that I haven't forgotten this but it's been buried behind more urgent looking things and I'm very sorry about it.
The main value proposition is that there valid headers we can't parse to a `Long`-backed model, and we need repeated logic to rule out invalid headers that `Long` can model....
The Literally library makes those macros easy as well. The distinction is perhaps whether we want refined types (i.e., `NonNegLong`) vs semantic types (i.e., `Length`).