Notices by Wolf480pl (wolf480pl@niu.moe), page 23

@reyk it's not a thing anymore. Now it's a conscious being!

@BluRaf looks tough

Be me.
Update a small internal-use django app from Django 1.4 to 1.11
Get the gerrit change merged without realizing there's an auto-deploy script.
And that the app is deployed using a system-wide python installation.
On an old server running PLD Linux.
With a broken /usr/bin/virtualenv

@tek what happens when you make an average solution to a complex problem and then put a facade on top of it so that it looks simple?

@deavmi @paco @rysiek @HerraBRE Remote Code Execution vulnerabilities.

@rhiaro @phryk online forums had a page called "rules" for as long as I remember, and I don't recall anyone wantinf a "rules-free zone".

@phryk @rhiaro IMO the fact that it's called "CoC" is enough to make people not want it. The first time I've ever heard the word "CoC" was closely related to some "all men should die" drama. And even if your CoC is good and you intend do enforce it impartially, when people hear "CoC" they think about far-left people in community-management roles who hate white cis males.

@cjd in Java, the JIT can act on the declared types, and then watch the code run for a while, infer/guess some additional information, and recompile the method to be faster.

Declaring a type doesn't disallow the JIT from guessing, but it narrows the spectrum of types that the JIT has to guess from, so it decreases the chance of a wrong guess.

Also, dual question: have you ever resurrected someone's half-done side project?

Did anyone ever resurrect your half-done 3 years old side project?

I don't remember it happening to me, and I have a feeling that if I was to be the one resurrecting, I'd rather start from scratch.

@mwlucas @joedicastro @mulander yeah, if I got some issue reports or PRs on a dead project, I'd mark it as such.
OTOH, if people use something "because it works", then maybe it doesn't matter if it's dead... it works afterall...

@mwlucas @joedicastro
@mulander how about looking at date of the last commit, and the number of issues?

If I make a side project, and then don't have time / get tired of working on it, I don't know if or when I'm gonna go back to it, so there's no discrete "it's dead from now" moment.

@cjd if anything, it would allow JIT to do more

@cjd AFAIK every theorem can be represented as a type, so static assertions are a form of static typing :P

Anyway, I think the default level of static checking should be somewhere between Java and Haskell. For JS, the default is zero.

Also, I don't see how static types could force JIT to do anything. JIT may treat it as a hint, but having more information shouldn't be limiting what it can do.