TheFeedingHand
Member
That's what we said. People just prefer to use the term librería to refer to Programming languages' libraries (at least here)
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Programming |OT| C is better than C++! No, C++ is better than C
- Thread starter Kikarian
- Start date
That's what we said. People just prefer to use the term librería to refer to Programming languages' libraries (at least here)
Osiris
I permanently banned my 6 year old daughter from using the PS4 for mistakenly sending grief reports as it's too hard to watch or talk to her
My first professional language was COBOL.
"I Eat Diced Pineapple!"
Kill me now, I remember even slight changes to old code requiring a full rewrite, damned verbose language that it was.
Yeah, that's what we were saying. The correct term is biblioteca, but you'll find way more people say libreria anyway in the industry.
Yes it's the exact same in French with bibliothèque vs. librairie. I know biblioteca is the correct translation, but I was wondering which one was most used by hispanophone developers.
dragonelite
Member
I hate unit testing when i was forced to use it.
But most of the stuff i write aren't really unit test friendly or i have to spend 30% of the time thinking about the test. When i could just put some break points verify it with my on paper done calculation then verify it with my eyes. At least that is how i like to test my code so far that i have written. Can be that when i need to go into a other field after graduating maybe Unit test are the better tool. When visually checking isn't that easy.
They're significantly less useful when you're the only person working on a program, and you're going to write it once and never touch it again.
They're much more useful when you're writing a component of a larger, complex application that will have to live for years, be changed lots of times by other people, and depends on other parts of the application that will change too.
lastflowers
Banned
Yep. When run on a regular basis, they help prevent major regressions in functionality.
dragonelite
Member
Maybe the teacher choose the wrong project to introduce unit testing with imo.
But then again most programmer i know hate typing documentation and writing unit tests and we all know they are useful later on.
BomberMouse
Member
Went to some interviews today, totally bombed whiteboard coding. Had to commute for 2.5 hours to get there so I was pretty tired when I arrived. By the end of the day I was a complete zombie mumbling nonsense stuff (specially since english is not my first language).
TheFeedingHand
Member
Sounds like Amazon's interviews
MaximusNeo701
Member
Carfax had technical interviews as well, none on a whiteboard but a 3 part process.
2 teams do technical questions then the 3rd is a sit in programming with a team member. They used the XP agile approach with 2 people per pc, each with a keyboard monitor and mouse and the screens were set to duplicate with the 3rd middle one set separate for requirements.
You actually you were dropped in and in XP pair, one person would write a unit test and you would write code to make the test true, starting with the most basic of test. It became a contest where the 1st test would be return value is a string; so you would just write return "string". Then he would write it has to give back a certain kind of string and you could just return the actual string, then the opponent would say okay it has to be one of the strings in the list of acceptable returns. So again just have it pick a random one of them. Eventually they would force you to use the input you were given to get the expected output, but the unit testing was very very thorough since we treated it like a contest, and the code was also much better because of it.
Actually I did well enough they let me spend the rest of the afternoon there pairing in with a team to see how well I did.
I would say white boarding is harder, on another interview they just had me answer very generic white board questions and I was really stunned because of how open ended they were, which I think the interviewer did a pretty bad job of asking the questions since they were all over the place and back and forth and not specific. He actually asked me how I would make decision's in code and I was going to feel dumb if I said "If statements or switch cases" which happened to be the answer.... Then they would ask some crazy hard question about using agents in microsofts business intelligence tools. Then back to some question that was way to easy again like how do you declare a variable, or how you come up with a naming scheme. So don't feel bad about not doing so well on one, it happens to all of us.
toddhunter
Member
Yep and by regularly that will be on every check in through a continuous integration server. I'm a big fan of unit tests and indeed test driven development in general. As usual, it depends what you are doing but it is a tool to embrace if you can.
As for interviews, my best failure was when I was asked to write out the xml for the main controller in spring. I'd been working with it for about 3 years and still didn't have a clue. Who doesn't just copy paste that shit? Seriously.
still salty...
ThoseDeafMutes
Member
I've been teaching introductory level programming as a tutor for my university for about 6 weeks now. Really enjoying it overall. Not sure if other countries have the same name for this position; I guide them through their practical sessions since there are too many classes for the lecturers to take.
lastflowers
Banned
lorebringer
Member
Yeah where I am it's a demonstrator (if helping in lab assignments) or a tutor (if running actual classes).
Some code really doesn't lend itself to unit testing, and I hate it. Working at the moment with a lot of code that's not very test friendly (because of tight but heavy frameworks it has to work in and other coding restrictions) and it's frustrating. And most unit tests wouldn't have caught the issues we've been having lately, as they only kick in in complex timing-related situations. And timing-related unit tests can be a nightmare - there was a vile one at a previous job which failed if a particular task took more than a specified time, which meant that anything running in the background could cause random test failures (on a test suite that took tens of minutes to run and was required as part of any commit).
Wow. I'd be salty about that too. Such a meaningless thing to test.
TheFeedingHand
Member
It's called a Teaching Assistant in the States and Canada.
Anyone know why my value isn't changing in this constructor?
My class file is
The "analyze" function is from another assignment I did previously and it calls that and decodes a block of text. What I'm assuming is if I call that anaylze function, it'll change the values in the class. I guess this is wrong though? But when I cout << the values, it says the proper values that should be changed.
And if that isn't the proper way of doing it, can someone tell me the proper way? Thanks!
Code:
Seti::Seti(const char rawData[ ], int freq, int a, int b){
cout << "in specific Seti constructor!" << endl;
if((freq >= 0 || freq <= 9) && ((a >= 0 || a <= 255) && (b >= 0 || b <= 255))){
result = analyze(rawData, signal, freq, a, b);
}
else{
freq = 0;
a = 0;
b = 0;
}
}My class file is
Code:
class Seti{
private:
char rawData[1500];
char signal[121];
int freq;
int a;
int b;
long result;
public:
Seti(const char rawData[ ], int freq, int a, int b);
};The "analyze" function is from another assignment I did previously and it calls that and decodes a block of text. What I'm assuming is if I call that anaylze function, it'll change the values in the class. I guess this is wrong though? But when I cout << the values, it says the proper values that should be changed.
And if that isn't the proper way of doing it, can someone tell me the proper way? Thanks!
What value doesn't change? result?
lorebringer
Member
If I'm reading you correctly you mean that the class variables for freq etc. aren't being changed? This is because you never assign them. If you print them out inside the constructor, you aren't printing out the freq, a or b that belong to that object, you are printing out the values you passed in. Since you have local variables hiding the names of the class variables, you need to use the this pointer to assign to the class variables.
this->freq is a variable belonging to the object under construction.
freq is a variable that you have passed into the constructor, modifying it will have no effect on the object.
Does that help? Notice the this-> in the else clause as well.
this->freq is a variable belonging to the object under construction.
freq is a variable that you have passed into the constructor, modifying it will have no effect on the object.
Code:
Seti::Seti(const char rawData[ ], int freq, int a, int b){
cout << "in specific Seti constructor!" << endl;
if((freq >= 0 || freq <= 9) && ((a >= 0 || a <= 255) && (b >= 0 || b <= 255))) {
// if this is where you want to assign them
this->freq = freq;
this->a = a;
this->b = b;
result = analyze(rawData, signal, freq, a, b);
}
else{
this->freq = 0;
this->a = 0;
this->b = 0;
}
}Does that help? Notice the this-> in the else clause as well.
a, b, freq. I'm pretty sure it's because analyze doesn't return their values but I need to set the variables in the class to them and I don't know how to target those specific values to set it from that one function
Thanks, I tried this but it crashes after the test. I'll try to figure out what's wrong before asking again though. Thanks!
IMO, all timing dependencies in tests must be faked out. Even if a bad timing results in that test failing only 1 in 10000 times, the problem appears when you have 10000 tests. It may then be unusual for any particular test to fail, but normal for some test to fail. If I've learned anything in my time doing unit testing, it is that if your test is not deterministic, it is not worth having.
False positives are the bane of a useful test suite. Crying wolf will render it useless in short order.
Orbis Tabula
Member
So is this a good thread to ask a programming question? I've been learning Java in University for almost all my programming classes, but the last few semesters were really general stuff. Now for the last class I need, I'm back in Java and I've forgotten a ton.
My assignment is to make a world of nodes that are randomly connected. There have to be 200+ nodes with any number of connections assigned randomly, followed by some calculations I need to do with them. I generally know how I'm going to do everything regarding the work, but I'm having a little trouble in the basic implementation, which is embarrassing.
I've set it up with a Node class, which just has an ArrayList of Nodes to keep track of all the connections. There's also a Network class which will have contain the ArrayList of all nodes and where the main method and everything will happen.
My question is, how do I populate that ArrayList with 200+ unique nodes without doing "Node node1 = new Node();" over and over and over and over? I know there's a way to do it. After I've got them all, I just need to make the connections, which should be easy enough.
I feel confident I can finish off the whole project but it's just this stupid basic problem I'm having. Thanks for the help, GAF Wizards.
My assignment is to make a world of nodes that are randomly connected. There have to be 200+ nodes with any number of connections assigned randomly, followed by some calculations I need to do with them. I generally know how I'm going to do everything regarding the work, but I'm having a little trouble in the basic implementation, which is embarrassing.
I've set it up with a Node class, which just has an ArrayList of Nodes to keep track of all the connections. There's also a Network class which will have contain the ArrayList of all nodes and where the main method and everything will happen.
My question is, how do I populate that ArrayList with 200+ unique nodes without doing "Node node1 = new Node();" over and over and over and over? I know there's a way to do it. After I've got them all, I just need to make the connections, which should be easy enough.
I feel confident I can finish off the whole project but it's just this stupid basic problem I'm having. Thanks for the help, GAF Wizards.
If you want 200 nodes, do something like this:
Code:
for(int i = 0; i < 200; i++)
{
Node node = new Node();
network.Add(node); //or whatever method you use to add a Node to your Network.
}jeremy70583
Member
Why not use a list of node's, then something like
Code:
List<Node> list = new ArrayList<Node>();
for(i=1;i<=200;i++)
{
list.add(new Node());
}Orbis Tabula
Member
Wow, now I feel stupid. I didn't think it would be that simple. I thought there'd be some problem with the fact that all the nodes have the name "node" but that definitely works. I thought I'd have to have a node1, node2, node3, etc., etc. This works, though. Thanks a ton to both of you!
chickdigger802
Banned
4 weeks in work... finally get my first defect to look at.
tip... look at the log. Problem is I barely have a clue what shit suppose to look like when stuff is suppose to work right lol. Think I'm gonna annoy the heck out of co-workers (not my fault that there is pretty much no documentation). At least walk me through something
tip... look at the log. Problem is I barely have a clue what shit suppose to look like when stuff is suppose to work right lol. Think I'm gonna annoy the heck out of co-workers (not my fault that there is pretty much no documentation). At least walk me through something
dragonelite
Member
Jezus after 1.5 day reading on homogeneous clipping for a software renderer and a shit load of on paper calculation to understand the steps the pioneers took almost 30 years ago. Finally found a good paper that is so clear im facepalming myself on why i didn't see most of it myself. Now work the algorithm out on paper and gonna implement it tomorrow in code its almost 1am here right now.
/Math kicked me in the balls hard.
/Math kicked me in the balls hard.
WPF looks so pretty... I just finished my first "true" web application and its going live tomorrow. Pretty excited to be honest. Hope the "launch" goes smooth though we are in a testing period for 2 weeks for it. I've already been assigned a second project and I am seriously thinking of going with WPF just to test it out. Any takes on it?
Post your analyze function.
My group has run into an interesting problem on this software engineer project. Only one person has been coding most of the stuff throughout the past week so we haven't noticed it until now. Now that there are more than two people trying to work on the programming, we're having to wait until one person finishes just to make sure two people aren't editing the same source files at the same time.
How do you guys usually handle this stuff?
How do you guys usually handle this stuff?
Though it's complex to grasp at first, Git + github (or other source control: SVN, Murcurial etc).
Version control systems can be pretty smart about merging changes. They know what version of the file you based your changes off of, and what version of the file he based his changes off of. So long as the changes are not directly conflicting, it can usually resolve them sensibly. Wikipedia talks about that here.
However, if you make changes not only to the same file, but to the same parts of the same file, you will probably have to manually choose whose changes to take (or rewrite that section entirely if neither is satisfactory).
There's a lot of benefits to having a version control system, but if you just want to the fast and easy solution you could use a stand-alone merge tool instead. Take a look at WinMerge.
StrangeADT
Banned
Well you just need to realize that identifiers (names) are attached to references.
For example, if you go:
Node myNode = new Node();
myNode = new Node();
myNode will now be pointing at the 2nd node you created. Since Java is garbage collected, if you have no more identifiers referencing the first object created, it's actually earmarked for deletion whenever the garbage collector runs.
In a loop, like the ones posted above, if you didn't add the node you create to the list, it's basically the same as what I just stated but on a larger scale (cause it will happen each time the loop executes). The slight difference is that the identifier is declared every time around but that's a minor variation in terms of what you need to be concerned with. Just know that the identifier only points the the most recent object that you have associated with it, and that unless you need to keep what was previously pointed to around, there's no problem with it at all. Adding it to the lists is what you want.
Your project sounds quite basic so I'm sure you won't have any more problems.
Considering the time of year and the scope of their assignment, it might be best to stick with manually managing it for now. Getting 3 or more people to learn GIT and effectively use it to merge branches is probably not a good idea considering they likely have less than a week to finish the project itself at this point.
Oh god, our version control has been so terrible throughout (probably my fault), but it worked fine up till now where our only "version control" was for documentation. Since there was only one person working on most of the programming, we never considered it for the programming bit.
Thanks everyone, I'll check this stuff out.
Thanks everyone, I'll check this stuff out.
Yeah, too late at this point. Future reference and all though, so it's all good!Considering the time of year and the scope of their assignment, it might be best to stick with manually managing it for now. Getting 3 or more people to learn GIT and effectively use it to merge branches is probably not a good idea considering they likely have less than a week to finish the project itself at this point.
Lasthope106
Member
Check out bitbucket. It supports both Mercurial and Git, and unlimited private repositories, and best of all it's free!
mugurumakensei
Member
4 weeks in work... finally get my first defect to look at.
tip... look at the log. Problem is I barely have a clue what shit suppose to look like when stuff is suppose to work right lol. Think I'm gonna annoy the heck out of co-workers (not my fault that there is pretty much no documentation). At least walk me through something
I've been working at my current company for almost 3 years now, and there's still times I wonder how such illogical code managed to work at all. It's especially difficult because the older programs are pure C CGI programs. Newer stuff is a combination of web services in C, clearsilver (for web templating) + C cgi program, plain html pages, and dojo for rich client controls.
Luckily, I mainly work on client side things these days. However, given the old stuff is plain C code with a bunch of printfs, I have to deal with pains of maintaining such messes.
Still beats data conversion and having to bang my head against the wall figuring out what key and algorithm was used for encrypting a credit card data so I could migrate credit cards from clients old business management system to our system.
It's really long. I divided it up to several sub functions. Why? No idea lol.
I'll post it in a sec hold on...
EDIT - Here: http://pastebin.com/6dZL9Cqk
I have to delete it soon because I'm not sure I'm allowed to show it but yeah.
SlipperySlope
Banned
I'm at a crossroads. What language do you guys use for Windows-based tools development?
I'm creating a level editor. My game is written in C++. I would like the editor to use the same graphics engine that my game runs - DirectX. I already built an easy-to-use engine around DirectX, so why not use it here?
So I'm trying to decide what language to code the tool in. I've contemplated C++ with the Win32 API, MFC, .NET...
.NET is kinda tricky since I'd need a wrapper for DirectX, and couldn't use my graphics library without another wrapper... or I can say screw DirectX and create some graphics routines through .NET...
What do you guys code simple level editors in?
I'm creating a level editor. My game is written in C++. I would like the editor to use the same graphics engine that my game runs - DirectX. I already built an easy-to-use engine around DirectX, so why not use it here?
So I'm trying to decide what language to code the tool in. I've contemplated C++ with the Win32 API, MFC, .NET...
.NET is kinda tricky since I'd need a wrapper for DirectX, and couldn't use my graphics library without another wrapper... or I can say screw DirectX and create some graphics routines through .NET...
What do you guys code simple level editors in?
dragonelite
Member
I dont know but i heard C# is almost as good as industry standard for tools building.
If it is a level editor then performance is not of a issue. A friend of mine did something with dll and wrapped those in C# but i have no experience doing that in c++.
Do you have Game Coding Complete 4th edition?
In chapter 22 of that book they talk about it. It is also the chapter my friend used to make his editor.
dutchguyjack
Member
I'll second that - C# is used for tool development in most companies.
DragonElite - Would you recommend Game Coding Complete? I've been looking for a good book, but there's a lot of them and a lot of poor ones out there.
dragonelite
Member
From what read of it 30% its really good written and a lot of useful tips so far. All revisions have high scores on amazon which is what i use a lot for seeing how the rest thinks about a book.
These are tech books so big changes are not a lot of casual readers will dive into them.
Its a more written in a personal way joking aside and stuff instead of really professional like most other tech books. I have way too many books and way too little time.
But a cheap deal is a cheap deal and can't pass a good book up on ebay.
Anyone ever developed a plugin/plug-and-play architecture with a webapp?
At work we have a sophisticated web app, and one of the requested features is a new section that allows members from other teams to extend the app with scripts or other tools, not requiring a new build, just something they can add to the app and use.
We're using spring + java but if anyone has done this to any degree, I'd love to hear about it.
At work we have a sophisticated web app, and one of the requested features is a new section that allows members from other teams to extend the app with scripts or other tools, not requiring a new build, just something they can add to the app and use.
We're using spring + java but if anyone has done this to any degree, I'd love to hear about it.
Lasthope106
Member
If you are coding a windows application in 2013 with C++ and MFC, you are probably doing it wrong.
C# and .NET are the way to go. I've seen some DX wrappers, but can't tell you which one is the best one.
Need some help with a linked list in C. I have a client-server program with communication over tcp and udp. When clients connect to the server they are added to a linked list and a communication thread is created for that connection.
Problem is that the thread that adds the client to the list loops endlessly when i try to connect a second client. It doesn't do that when i remove the add_llist() function, so I'm sure there's something wrong with that function. It's probably the most obvious thing ever, but I just cant figure out what causes it.
This is my code: http://pastebin.com/tCxey1DA
Problem is that the thread that adds the client to the list loops endlessly when i try to connect a second client. It doesn't do that when i remove the add_llist() function, so I'm sure there's something wrong with that function. It's probably the most obvious thing ever, but I just cant figure out what causes it.
This is my code: http://pastebin.com/tCxey1DA