Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Reelya]

Spoiler (click to show/hide)

I was having some weird trouble compiling a program using libtcod via console earlier (from a libtcod C++ roguelike tutorial) and I still can't get that code to work, however, after poking through some other tutorials and the like, I've been able to get another simple program to run both via console and codeblocks. I'm trying to figure out what my problem is with the original code.

This is what it looks like:

Code: [Select]

#include "libtcod.hpp"
int main() {
    TCODConsole::initRoot(80,50,"libtcod C++ tutorial",false);
    while ( !TCODConsole::isWindowClosed() ) {
        TCODSystem::checkForEvent(TCOD_EVENT_KEY_PRESS,NULL,NULL);
        TCODConsole::root->clear();
        TCODConsole::root->putChar(40,25,'@');
        TCODConsole::flush();
    }
    return 0;
}
As per the tutorial, I'm compiling it with this input:
g++ src/*.cpp -o tuto -Iinclude -Llib -ltcod-mingw -static-libgcc -static-libstdc++ -Wall        [Someone mind explaining what these commands actually do?]

I'll try ...

src/*.cpp is the files to compile, which in this case means "everything in src director which ends in .cpp". each cpp is compiled separately then they're combined into one thing during "linking". you used shared .h/.hpp file to allow communication between the .cpp files. The basis of good c/c++ program construction is to minimize what is in the .h files to a "need to know" basis and have all the implementation details in .cpp files. If you only edit one .cpp files then you only have to recompile that one cpp, but if you change a .h file then any cpp file which references it must be recompiled.

-o tuto specifies the output program name

-Iinclude -I tells the compiler which directory to look for to find the .h files - in this case the directory called "include"

-Llib -L tells the linker where to find pre-compiled ".lib" libraries (in a "lib" directory, but this is an arbitrary thing you could change) libs are basically pre-compiled cpp files that get added in during the linking phase just like you're own cpp files. you need to include a .h file which matches the .lib file and lets you access the commands it contains.

-ltcod-mingw -static-libgcc -static-libstdc++

Look like the names of libraries to use in the linking.

-Wall 

Should be read as W:all, W sets the "warning level" which spits out advice if stuff doesn't meet standards but would still work. Stamping out warnings is good practice.

I've set up the folders include and lib within the program's directory, containing the things they should contain. The code itself is stored as main.cpp within the src folder. Following some advice from the tutorial, I ran the code with a command line debugger (gdp, I think), which gives this output:

Code: [Select]

(gdb) r
Starting program: C:\Users\Soulusk\gameplace\tuto.exe
[New Thread 8900.0x14c8]
[New Thread 8900.0x888]
24 bits font.
key color : 0 0 0
character for ascii code 255 is colored
Using SDL renderer...

Program received signal SIGSEGV, Segmentation fault.
0x65e70d9f in TCODConsole::clear (this=0x1) at src/console.cpp:196
196     src/console.cpp: No such file or directory.
It seems, to my untrained eyes, that it can't find the file console.cpp, which is in libtcod's src folder (I checked). Anyone here actually know what my problem is and perhaps how I might go about solving it?

the program is looking for a file that's not there. Line 196 of src/console.cpp is the location in the code that triggered the problem, not the file that is missing. Compiling and linking went error-free. cpp files are only used in the compiling stage, they're never accessed by a running program.

It's pretty easy to narrow down what the problem probably is. What type of file is accessed during runtime by low-level library functions? DLLs are! Most likely you forgot to put the DLLs in the location you're running tuto.exe from.

[Skyrunner]

So ... does anyone here know if the University of Illinois at Urbana-Champaign's CS major offered by the College of Engineering is a strong course?

Also, is it extremely competitive?

It's become college app season now, and I'm all worried, heh.

[Kirbypowered]

the program is looking for a file that's not there. Line 196 of src/console.cpp is the location in the code that triggered the problem, not the file that is missing. Compiling and linking went error-free. cpp files are only used in the compiling stage, they're never accessed by a running program.

It's pretty easy to narrow down what the problem probably is. What type of file is accessed during runtime by low-level library functions? DLLs are! Most likely you forgot to put the DLLs in the location you're running tuto.exe from.

Ah, good things to remember. However, I have all the .dll files I can find in the directory containing tuto.exe...I even threw in the gui related ones just in case, to no avail. Have any other ideas?

[Valid_Dark]

Since starting my new job I haven't had time to program,  I work 8 hours a day, and when I get home I'm so exhausted I just want to sleep.
Working in a warehouse, moving heavy shit all day.

I really need to get back into programming. 

[Skyrunner]

So . Any of you guys heard of the following technologies? If yes, anyone brief me on why they're good, please?

Angularjs
Node.js
Mongodb
Grunt
Jade
Stylus


Also, how are logins handled? O.o stored as cookie or web storage woth some sort of authentication hash?

[Angle]

Odd, I happen to be engaged in trying to figure out how to use one of those, specifically the JavaScript Mongodb. I haven't figure it out yet though, so ~~... I did use the java version of Mongodb, or rather expanded upon code that used it, and it seemed to work just fine.

[Mephisto]

RE MongoDB: It's web scale.

[Skyrunner]

Relevant webscale

[Telgin]

Mongo is neat, but I don't have a whole lot of experience with it myself.  Node.js I can expand on a bit though.

It's a pretty different feel from developing traditional web applications, like say Perl or PHP on Apache.  The server keeps connections open and can keep persistent state, which leads to different and possibly more efficient design philosophies.  For example, our company's backend API used for many internal applications uses a MySQL database server.  Whenever requests are made against the API it has to load column definitions for the database so it can determine whether to perform adds or updates as required.  Right now it's coded in PHP and uses Apache, so it has to reload this data every time a request comes in.  Node.js can store that globally somewhere and never need to reload it again unless the database schema changes.  It also only needs one copy for all running connections, where the PHP version has to waste memory loading copies per connection.

Node.js is also nice because it lets you use web technologies like web sockets, where that's hard and clunky with something like PHP.  That might make it useful for computer games or similar applications that do lots of communication between clients.  Integration with stuff like Mongo might also be simpler with Node.js since it uses JavaScript as its native language.  JavaScript has its ups and downs, but coding server side software with it is kind of cool when you're used to PHP.  You'll want to get comfortable with callbacks and asynchronous function calls.

Node.js does have some downsides though.  First off, Apache and similar servers gives you a lot functionality that you have to either import as a library or code yourself.  A router is an example  So, with Apache it's trivial to get a document root set up and have it serving index.html, page2.html, some_directory/page3.html without any coding.  With Node.js you have to manually decipher incoming requests to determine what page the client's asking for, load it and process it, then return it with any needed headers.  Stuff like that, or access control, or redirections, or other stuff is going to need code from you or a library.  Another huge downside is that if any incoming request causes a fatal exception it kills the process.  There's only one process, so all of your client connections get dropped until you restart the server.  Oh, and since there's only one process, trying to do anything truly in parallel is going to hard.  I think there's some way to do that now, but I don't know how it works.

In short, Node.js is pretty cool, but I don't see it replacing traditional web servers for a while.

[MorleyDev]

I'd say node.js is more suited for small and focused web services. So it's a good choice for a ReST API that sits around a database or an OAuth server, small things that 'simply' process a request, make some db calls and give a response and for which a 'large' server like Apache+PHP or C#/Java are arguably overkill.

[Telgin]

Amusingly a REST API is precisely what we were and are considering using Node.js for.  We do a ton of business logic in that API though, and our current PHP implementation is over 50K lines if I recall.  So in our case at least, Apache and PHP isn't really overkill, since we do have the odd thing here or there that does some heavy lifting on the server side.  Forking processes, filesystem stuff and so on.  Node.js can still do everything we need, as far as I know, but it would take some effort to really get it to do what we need.

There's really just two things I'm still concerned about.  The first is the exceptions killing the server, which I know solutions exist for but I need to look into more.  The second is the fact that PHP's (arguably pointlessly) gigantic standard library goes away with Node.js's JavaScript implementation.  Off the top of my head I can't think of too terribly many things that we'd have to recode or find libraries for, but PHP's date and strtotime functions are things we use a lot...

[Kirbypowered]

Okay, aftering farting around and harassing the people over at the Temple of the Rogue forums for a little while, I've come to the conclusion that libtcod is just rude and I've decided not to bother with it any longer for the time being. I'm currently fiddling around with a little library someone over there recently released, called BearLibTerminal. It seems pretty nice and I've already got it working, so my impressions of it are good already.

Anyhow, thanks for putting up with my constant stream of dumb questions for the most part, and also sorry about that. I just recently noticed that there's a thread in the Creative Projects sub-forum specifically for "newb programming help"...so yeah. Sorry about that. XP

[MorleyDev]

Yeah. node.js is more suited towards a more distributed model than a 50K monolith.  It's more geared towards dropping pieces of work onto queues for other processes to pick up and work with, forming clusters of small processes working together. Which I'd think of as a cleaner design nowadays for that kind of work, but difficult to retrofit in.

As for exceptions, it's pretty easy inside a REST api to turn all exceptions into 500s. And if an exception does bring the program down, pm2 or forever are tools made to bring it back up (pm2 is newer but the recommended solution). It still does encourage one to write more unit tests for the code at least, but it's a valid concern that the libraries often don't handle this for you I guess. At least I know Restify doesn't, not used Express as extensively. PHP handles it because it's built with the notion of a web-page-as-a-process, node.js isn't built with that notion. Neither are C# and Java, obviously, but there the libraries typically handle all that for you.

Date functionality? Hmm, maybe some of the libraries recommended here could be of use?

[Gentlefish]

So uh. Quick question for y'all.

Where would be a good place to start if I want to develop a little program to make a sprite walk around? 2D or Overhead doesn't matter because I want to try both.

I know SDL is common but I can't seem to find a good tutorial on it. Any other recommendations? I'd be doing this on a Win 8 with code::blocks installed and set up.

[gigaraptor487]

Lazy Foo has an excellent tutorial for SDL