See more at the announce forum.
I already wrote about most the changes here: http://dpldocs.info/this-week-in-d/Blog.Posted_2020_11_16.html#what-adam-is-working-on but now it is rolled up and tagged on dub.
The new thing I didn't write before is http2.d's behavior on connection failure and timeout is different. Previously, it would throw an exception and cancel all running requests as a result. Now, it just marks those individual requests as having failed (by setting a code > 0 && code < 100 in the response - codes >= 100 are HTTP responses, ones under 100 are internal messages from the lib.)
This is a breaking change - if you depended on that exception you should change it. But you really should be checking the response.code anyway since it can fail server-side too... but now certain connection errors can be encoded in there too and that is different.
I also started logging breaking changes in the README.md with this. I will not do a detailed changelog there though - you can still look for the "History" comments or sometimes the git log though.
I consider the idea of "google it" to be harmful. Clearly, it should be "bing it" lol j/k.
I somewhat understand it. It can be annoying to answer questions sometimes, but you're better off saying nothing than saying "google it" (and note that posting a link to a search result page is just as bad). Why? Well, it is, at best, of zero help, and it is frequently of negative value.
For one, consider that searchers tend to find results of other people asking similar questions on public websites. If your response is "google it", even if it is justified to that one individual, you must consider that you're condemning future people performing that search to hit your dead end. Indeed, once I searched something, found a hit of someone asking the same thing, but the only reply was a link to a search results page... that led right back here. A loop of pain.
In a way, I get it, being able to find your own answers is a good skill for people to learn. But just telling them to search doesn't help teach this skill (and indeed, the right search term can be hard to know), and I'm concerned about the individual-focused culture this promotes.
See, in a professional environment, in my experience anyway, people tend to waste a lot of time searching for things that their co-workers may be able to help with a lot faster and it seems to me this is the result of this search culture combined with some myths about interruption.
People just seem to be really reluctant to ask, and that might come from fear of looking like a n00b. I think we should change this both with ourselves and with up and coming new programmers. Yes, knowing how to search is good, but so is just asking people too so you can carry on.
With the interruption thing, let's say it does throw the person off. Well, look at the big picture: is it better for you to waste hours searching in circles or throw someone else half an hour off then you both get back to work?
Productivity of course isn't all there is to life, but if it is your goal... the gains of helping someone is bigger than the loss. And besides, answering their questions help you to understand what problems are actually happening.
So I think we need to stop telling people to search. My feeling is we'll make better search results and a lot more work done if we encourage asking and answering instead.
arsd 9.0 out with some arsd.http2 changes and the ttf OperatingSystemFont thing in arsd.simpledisplay, among other things. I also write about how searching is a good skill, but it shouldn't discourage us from asking colleagues.