@Luc I do not particularly view questions with oddly specific requirements to be a good opportunity to twist into a more general case. When you do that, a reader might get the impression that using bash script to generate passwords that you can remember is indeed what one should be doing, when it is far more practical to the general case to simply use a password manager to generate passwords. Again, in which case, memorability is not at issue.
@Luc So, I think you're trying to make this question something it isn't, and fit a square peg into a round hole, to speak. It is my opinion that an answer that caters to the specific requirements of the question without adding additional arbitrary requirements is best.
@Xander Actually I think the command line is a perfectly fine place to generate passwords, but that might be just me (after Firefox the terminal is my most-used program, or so Whatpulse tells me)
@Xander Yeah I am bending the question into something useful, but it doesn't differ so much from the OP's question that I'd say it's fitting a square peg into a round hole
@DavidFreitag I try to put checks in my code so that if it only works over a certain range of input, it will at least stop before it tries to do something with input outside of that range -- with a decent exception message to boot. So then, when it breaks and I look up the exception code, I have a note giving me a hint of what I could do :-)
@Tinned_Tuna Ah yeah that was a simplified example. Most of the time there are no huge glaring problems like an exception because usually the code is in C on a microcontroller.
Also I should note that these are for personal projects not work stuff..
@DavidFreitag ah, I work on big production servers. When something goes wrong, we notice and we get logging output. I think I have a comparatively easy life.
@Tinned_Tuna Yeah thats more in line with my work environment. My boss goes through each of my commits so I don't really have the luxury of being lazy like I do in my personal life.
@Tinned_Tuna Not really, he just goes through it. When a bug comes up and I find the issue and report it to him I usually get an, "Ah yeah I noticed that in your commit"
We're hoping to do better than that. We're going through improving our development process, code style, static analysis, code review at a minimum. We need to develop some guidelines for how to code review and what makes "good" code...
It's going to take us some time, but we want to be in the situation that when we build/compile our system, we catch as many issues as reasonable before we go live. Also, we're expanding the team, so having a bit of control to ensure that everyone codes roughly the same and >1 person has at least seen the code and can have a crack at debugging it is fairly i
@Tinned_Tuna unfortunately all our stuff is .Net and in order for Jenkins to play nice it needs to be running on windows, and we have no spare licences
The first step is to get the things like static analysis, style enforcement into the build, then module by module, clean up the system and make it fail if it's violated for each module. Repeat until we cover the whole codebase.
@DavidFreitag ah that sucks :-( We're a Java shop for the most part, though we do some research prototyping in Haskell because it's relatively quick to do so, and easier to model things in it.
@Tinned_Tuna Yeah Jenkins is a perfect fit for Java :]
@Tinned_Tuna Well we don't need much horsepower... But I'm working to get the hardware updated to the AM335X CPU line (the same as the BeagleBone black)
Thankfully that will run embedded linux and all the drivers are written already
I did get given a heap of hardware stuff, some stuff I don't even recognise. But I've been so busy it's mostly been gathering dust except for the mbed and the raspberry Pis
woo yea! Heys tutorial SPN with modified SBox is implemented. And the tests pass. Life is gooood. Now if only I could hammer out my correlation attack :-(