Actix Web – Project Future
github.com
github.com
Figuring out whether an author's trust is earned or if they just used unsafe to stop the borrow checker from nagging is non-trivial for precisely that reason.
So you question is very likely impossible to answer even by the author himself. If you are really eager to know, dig yourself.
There was absolutely no drama involved. As long as you credit the original authors and isn't passive aggressive about forking, all is good.
In fact, one of the authors of a library I forked noticed the attention my fork was getting and wanted me to merge back. Everyone won in the end.
Seems like I'd take a few days off one week and a twitter storm would brew and then a doxing campaign would fire up and there would be lynchmobs and angry pitchforks outside the building where I work. Obligatory comparison to Hitler or such and then I'd shrug and say, "its a tetris clone, people, calm down!"
I'm hoping that is all an exaggeration. But from what I have seen recently its not that far off in terms of online hate that can appear.
Is there a guide for maintainers of such things so deal with the generic overhead that must come from having a "community" around some popular tech software?
For whatever reason I sort of assume that there was a barrier on github where users who bother to read the code also code and would have some level of understanding that coding is hard, not every bit of code everyone writes is the way they want it ... and that would add to some level of understanding and civility.
It seems like I was wrong.
I don't agree with the mean things. Yet I can understand the frustration: here is a new patch that works and fixes something that could lead to a security issue? No one wants to fork the entire project and start their own version.
Honestly, the maintainer seemed kinda arrogant to boot. It might have just been a language barrier too, so who knows. I think if you took enough time to really look at all of it, it's a much more complex situation that what everyone is simplifying it too.
The mistake seems to have been that no effective countermeasures were taken before things boiled over, but at least now the whole mess has ended on a positive note.
However, many people saw that this was a caviller attitude to take, especially as it wasn't signposted, and gave the author grief for this.
The problem was, though, that the cumulative effect of these "unsafe is bad, therefore you are bad" comments (which shouldn't have been made) was that the author was dismissive towards reports of actual unsafe behaviour in their library which could have caused security issues.
The author got frustrated and moved the repos to his own personal account and left a patronising message in its place.
Bare in mind that Atix had been built up as one of (if not the) best Rust libraries for HTTP servers, partly off the back of the benchmarks (for which safety had been sacrificed) and partly from the fact it ran on stable (non-nightly) Rust.
This weekend, someone found an unsound use of unsafe code, reported a bug, and eventually a patch was supplied that switches out some optimized Actix types for somewhat less optimized standard library types. The author dismissed the patch as "boring," the internet got mad, and the author got mad.
I had a similar problem with Chrono [1], the most popular Rust date & time library at the moment (though now rivalled by time-rs [2] ;-), where I was struggling with personal issues and yet (thus?) unable to sufficiently materialize my vision. Eventually I gave up and passed a torch to other interested people, one of which is now the primary maintainer, but no major API design and stabilization that would lead to the 1.0 is not happening as far as I concern.
There is a lot of criticism in the Software Engineering community in general but the Actix story made me feel way more welcome and open. I'm very happy to see that!
you'll find all the references there