Are you implying they actually have developers? Heh...

You're never too much relaxed.
Ok, may be when your brain is flowing on the floor, here, it's too much.

Welcome back ! Hope you had good vacations :) (and it seems to be the case)

I just cleaned the cache last week when i had login issues during the sale... ugg...

edit: K cleared it and made no difference...

Of course they have.

Of course they have, but it seems management has them focus entirely on adding new features, rather than debugging existing ones.

I suspect they don't have enough people but have very ambitious goals (like Galaxy).

Indeed. And I would say that from their POV they are right. I don't suspect many people have left because of the bugs, but I seem to see hordes flocking in at the mere mention of a client (and a witcher).

*nod* That's why companies ahead of the curve trumpet "Never test by hand." It forces you to research suitably low-friction test systems (a one-time cost) and then you spend about the same amount of time but every test you ever do can be trivially re-run to catch regressions and, as a side-effect, you spend less time worrying about the correctness of code you've written and more time actually coding.

(It's also why companies like Google come up with solutions that let them code in a language with stricter compile-time guarantees like Java and then transpile to JavaScript. Programmer time is expensive and humans make bad auditors. CPU time is cheap and computers are pedantic by nature.)

With WebAssembly I expect it will be trivial to write in Rust and compile that into Web applications:
https://github.com/WebAssembly/design/blob/master/README.md
https://github.com/WebAssembly/design/blob/master/FAQ.md
https://www.w3.org/community/webassembly/

That's what I'm hoping for. Looking back on when I've burned out on personal projects, half of it has been trying to use unit test suites and manually-coded checks to attain Rust-level guarantees in another language.

As soon as I finish catching up the mess of backlogged stuff I currently have to do, I want to sit down and practice the hell out of Rust. (Initially, by doing a fairly straight port of my gif.py frame counter to satisfy my curiosity as to whether the Python runtime imparts statistically significant overhead on optimized, I/O-bound code.)

I'm going through the Rust book now gradually, and so far the language looks pretty neat. An interesting mixture of functional and object oriented approaches with a very strong bend on memory safety. I want to test it also on something practical.

Yeah. Before things really became a crusher, I read through the Rust book once and I've already started reimplementing gif.py's CLI using the clap argument parser. I just need time to get back to the effort and write the internals which are more heavily affected by the differences between Rust and Python.

So far, everything I've seen seems to indicate that the only bits I know I dislike are bits that are considered to be issues but were explicitly "postponed until after 1.0" and then postponed again because compilation speed and better Windows support were both low-hanging fruit (due to also being "postponed until after 1.0") and much more immediate concerns. (Stuff like the borrow checker being stricter than necessary)

From MaGog's logs for 5 July:

NOTE! CHANGED Broken Age: The Complete Adventure, changelog (http://www.an-ovel.com/downloads/chg/chg_1207662943.htm)
NOTE! RMV --- Changelog for version 2.2 (added 13.05.2015)
NOTE! ADD +++ Patch 2.2 (13.05.2015)
NOTE! RMV --- Changelog for version 2.1 (added 13.05.2015)
NOTE! ADD +++ Patch 2.1 (13.05.2015)
NOTE! RMV --- Changelog for version 2.0 (added 28.04.2015)
NOTE! ADD +++ Patch 2.0 (28.04.2015)

[Nice to see someone taking care of the changelogs]