The Big Programming Thread - Page 898

On July 07 2017 09:08 berated- wrote:


Can you verbalize all the reasons you disagree with their implementation? How much code are we talking about? If it's under a couple hours work, I generally try to rework it in the way that I think that it should be done differently without pushing any code off hours. This has a couple benefits, if you figure out along the way that you've missed some assumptions that are actual trade offs that they already considered, you did so without putting your foot in your mouth first. It also provides the benefit that if it works out as you expected then you would now have a talking point that you can approach, I saw your code but was curious did you consider x or y?

Either way, hopefully you work in an environment where you can freely and openly question why things are done. Seek to understand why they did it their way instead of imposing your own way. If you really think you know better then you should be able to ask them questions that lead them to your own conclusion without making any statements making them defensive. I personally also like to have these meetings 1 on 1 to make people feel less vulnerable and less likely to get defensive.

void function() {
    
    string a;
    if (condition) {

// BEGIN NEW CODE
        if (condition) {
            doSomething
            doSomethingElse
            return
        }
// END NEW CODE

        a = some a
    } else {
        a = other a
    }

    string b;
    if (condition) {
        b = some b
    }

    call(a, b, c, ...)
}

On July 12 2017 01:00 spinesheath wrote:
So if someone accepts it that overrides the explicit disapproval of 2 other people? Or did you just not approve it as in not say whether it was ok or not?

On July 11 2017 15:34 Blisse wrote:
We do have a code review policy. It's one approval. 2 people who left feedback (me + coworker) didn't approve it. A third reviewer approved it but didn't leave feedback. I believe the root problem is the one coworker who approves things but never leaves any feedback and I asked my manager about it.

Also, the code adds unnecessary nesting and an abrupt change in control flow. Just generally lowers the quality of the codebase. I happily push back against that. I disagree with the idea of moving on instead of fixing the root problems, though I can understand if it sounds like I'm nitpicking on small details.

On July 12 2017 07:55 Blisse wrote:
Regarding the code, I just haven't heard a satisfactory argument re: how it's not objectively worse. If there was worry about speed (which there is not) we could just early return at the top. The existing function has a flat, logical structure to each block of code. The new code adds nesting complexity and breaks the intended logic in each code block.

void function()
{
    string a = calculateA();
    string b = calculateB();
    call(a, b, ...);
}

On July 14 2017 23:14 sabas123 wrote:

Even if you use tools for computing cyclomatic complexity, you still run into a lot of murky areas for instance I don't think its possible convincing someone that a cyclomatic complexity of 10 is a generally a bad thing without resorting to opinions.

On July 14 2017 21:06 netherh wrote:


I think part of the problem is that the existing code is substandard too (or at least not defensively written). It's declaring a variable outside a scope, then doing a bunch of logic to populate the string (or not?), and then repeat. This "pattern" should actually be enforced in code by splitting it into separate functions. That way it's easier to understand, and a lot harder to change it "incorrectly". e.g.

void function()
{
    string a = calculateA();
    string b = calculateB();
    call(a, b, ...);
}

Now you can't mess with the control flow of "function()" while calculating "a", unless you do something drastic.

On July 15 2017 08:29 BluzMan wrote:


Funnily enough, I find this code pretty bad because it makes me think:
- Is a used anywhere else? Is b? Why are they variables? Am I missing something?
Almost unavoidable in a messy language like Java, but there's no excuse for that in C or C++.

On July 15 2017 08:44 Blisse wrote:



And what is your solution for that abstraction in C/C++ that isn't possible in Java... ? Other than exposing methods unnecessarily it looks.. fine?

void callSomeMethods()
{
    call(calculateA(), calculateB(), ...);
}

On July 18 2017 00:18 MrBrokenPenis wrote:
I hope this is the right place to post this.
Is there any library or API to get your opponents name at the start of a SCII match?
Would prefer python, if possible