The Big Programming Thread - Page 301

On May 14 2013 23:41 LightMind wrote:
Hmm need some help here. As a computer science student in my first year, I am planning to do some self-study during this summer break. I am hoping to go into game programming. I have some experience with Python, C and C++, and is hoping to go more advance with either C or C++.

Which language would you recommend to go in depth with for self-studying during this summer holiday? And are there any books to recommend?

Thanks!

On May 14 2013 23:41 LightMind wrote:
Hmm need some help here. As a computer science student in my first year, I am planning to do some self-study during this summer break. I am hoping to go into game programming. I have some experience with Python, C and C++, and is hoping to go more advance with either C or C++.

Which language would you recommend to go in depth with for self-studying during this summer holiday? And are there any books to recommend?

Thanks!

On May 15 2013 08:39 darkness wrote:
Does anyone here use JUnit & Ant? Thoughts? For those of you who don't know them, JUnit is a Java test framework, and Ant is a build tool.

On May 15 2013 12:35 obesechicken13 wrote:
I use selenium and testng with maven. Personally I feel like automated testing is a waste of valuable developer time that could be spent on more products but it's my line of work. I'm not going to bring it up to anyone at work that I think my work is a load of crock.

I'd rather have products with good features and functionality than having to constantly update test suites as products are developed. I often feel there should be one good test cases to test for past regressions that are likely to come back. More emphasis in my opinion should be spent on organizing good manual testing of software and employing customer feedback since there is just too much to do for automation testing.

On May 15 2013 13:11 tec27 wrote:

Tests are there to help you build new features and deal with requirements changes better, not to slow you down. Writing code without tests may seem faster in the short run, but in the long run you'll write code with more bugs and higher coupling that makes it harder to have time for doing new things. It sounds like from what you've said that you're mostly writing large scale integration tests, which is not really what people here are talking about. Such tests are necessary, but should not be the major focus of a proper testing strategy. They are hard to write, take a long time to run, and don't really give you all that much in the end. Unit tests on the other hand, as well as small-scale integration tests can be quite helpful and easy to write. I wouldn't throw the baby out with the bath water just because you find selenium to not be all that valuable.

I've been at a company that had next to no tests and a QA team larger than the actual dev team, and am now at a company with a large test suite and a QA team roughly 1/6th the size of the dev team. I can very easily say that the latter is vastly preferable

On May 15 2013 19:51 berated- wrote:


I'll second this. As a developer who works at a company with no test suites and a qa team of two, I can tell you that the lack of automated unit tests make our job pretty tough. Our main application code base is 10 yrs old -- we've built up a ton of technical debt that takes quite a bit of domain knowledge just to make the simplest of changes. This model worked really well when you have a team of 6 developers who have all been there since the beginning of the company, not so well when you have 10 developers who a majority of developers have been there less than 2 yrs. This model doesn't even pass the snicker test at bigger companies.

I've spent the last two years trying to push the benefits of testing to the other devs, but adoption is slow. Luckily, I got to experience the transition first hand of writing tests slows me down to writing tests makes my job easier/faster. It definitely takes time to become good at, though, so I wouldn't base any conclusions on how beneficial it is until you spend a couple months writing tests first and using mocks. I truly believe my code is better structured and modularized just because I started writing my tests first.

On May 15 2013 13:11 tec27 wrote:

Tests are there to help you build new features and deal with requirements changes better, not to slow you down. Writing code without tests may seem faster in the short run, but in the long run you'll write code with more bugs and higher coupling that makes it harder to have time for doing new things. It sounds like from what you've said that you're mostly writing large scale integration tests, which is not really what people here are talking about. Such tests are necessary, but should not be the major focus of a proper testing strategy. They are hard to write, take a long time to run, and don't really give you all that much in the end. Unit tests on the other hand, as well as small-scale integration tests can be quite helpful and easy to write. I wouldn't throw the baby out with the bath water just because you find selenium to not be all that valuable.

I've been at a company that had next to no tests and a QA team larger than the actual dev team, and am now at a company with a large test suite and a QA team roughly 1/6th the size of the dev team. I can very easily say that the latter is vastly preferable

On May 16 2013 01:27 obesechicken13 wrote:I'm not too familiar with unit testing but I feel like there would be many issues if you didn't test integration points or didn't test end user experience. A unit test can pass for some functionality but users might not be able to access your site at all.

On May 16 2013 10:21 phar wrote:

You do have to do integration tests, yes. Even if you have the most genius-designed system that's all modular and cool 'n shit, it's really easy to let some stupid stuff fall through the cracks if you don't have more end-to-end integration tests set up.

I think tek was just pointing out that you can't let integration tests be the majority of your emphasis, because, as someone pointed out above, integration tests can be a fucking pain in the ass to fix, and are as a whole a lot more brittle. And webdriver/selenium type stuff just adds another layer of absolute pain.

So integration tests are good, yes, but for a dev spending more time on smaller scale tests is a better use of time.

On May 16 2013 10:29 obesechicken13 wrote:

Ok. Well as I said, I know very little about unit testing. I think I tried it in school once.

On May 17 2013 06:38 Bionicrm wrote:
I'm sort of new to web design. I've learned the basics of Python, HTML, CSS, and am still learning JavaScript (which is easy because concepts are similar to Python) via Codecademy.com.

I'm making a website, and when I zoom-in in the browser window, my little menu bar gets moved over and under the banner, a form is falling out of the border, and other crazy stuff. Are there ways that people prevent this?

When I zoom-in to TL, things just get bigger and fall off the screen, which is what should happen. Instead, things feel free to move all over on my site.

Also, it is not visible on the web, as it's still in development (I view it from my computer), and I'm not sure if that has anything to do with it.

On May 17 2013 11:53 Birdie wrote:
If I'm dealing with a text file where the last line is Baggins Bilbo;05-89378 with no space or newline character at the end of the file, how do I detect the "last character" which doesn't exist, so that I can cut that number off and place it into a variable?

I can do it with other lines using code such as

if(input[i] == '\n') {
			temp[j] = '\0';
			j = -1;
			printf("Temp number: %s\n", temp);
			strcpy(tempPhone.number, temp);
		}

but that doesn't work when there IS no newline at the end of the number.

This is in C by the way.

I've tried using input[i] == NULL, which didn't work either (I would have thought it would because it's trying to get a character which doesn't exist). Perhaps I don't understand NULL properly yet.