The Big Programming Thread - Page 23

On November 14 2010 00:05 fabiano wrote:
I have a question, please dont laugh at me: how to make a new line in a text file in Java so that it works both in Windows and Linux?

I've this piece of code, where I append the \n char to make a new line in a string which will be saved to a text file eventually. It works on Linux, but on Windows I would have to use \r\n (or \n\r, dont really remember).

for(DBRow r : this.getRows()) {
    for(DBColumn c : r.getColumns()) {
        str += c.getValue() + "&";
    }

    StringBuffer buff = new StringBuffer(str);
    buff.setCharAt(str.lastIndexOf("&"), '\n');
                
    str = buff.toString();
}

 System.getProperty("line.separator"); 

On November 14 2010 00:20 MasterOfChaos wrote:
@KaiserJohan
There are several question on Stackoverflow about C# and audio. http://stackoverflow.com/questions/3924178/real-time-audio-playback-from-mic-c/ should contain what you want.

As codec you'll want to use Speex. There is CSpeex, a C# port of Speex (Afaik it supports both encoding and decoding) and probably there are some wrappers of the c library too.

On November 14 2010 01:58 RoTaNiMoD wrote:
Osmoses:

I recall reading in a game programming book a long time ago a solution for exactly the sort of problem you describe. Number the destination tile as 0. Then number any tiles adjacent to the destination tile as 1. Number any tiles that a 1 could move to -- other than 0 -- a 2. Repeat until you have numbered your Hero's current tile. Then as long as the hero moves to a smaller numbered tile than his current one every move, he is on a correct path. A benefit is that equal paths have equal likelihood instead of arbitrarily having to decide which path your algorithm will take.

Your example grid would look like this:




I think this could be what you are looking for... let us know.

On November 16 2010 21:50 KaiserJohan wrote:
Got a C#/general question:

Consider I have a custom class,

public class test {
private int a,b;
private string c;

<get / set functions, and also serialize/deserialize methods so I can stick it the class object into a byte[] and send it over a socket>
}

The problem is whenever I receive data on the other side, an event is triggered executes a receieve function which reads the data from a buffer and deserializes it back into my custom class test -- however, sometimes when I read from the buffer, number of bytes equals to 1½ or 2 classes might've been in the buffer already so when I deserialize, it deserializes into 1 complete class object, but throws the rest of the data away, so I dont get complete information.

The solution would be to know how big the class is, so I can only read from the buffer up to a certain index.... but does the size of the class object vary with the size of its variables?
For example, would string c = "kekeke" generate a smaller object than c= "kekekeekkeekeke"? In that case, how do I know how much to read form the buffer?

On November 16 2010 21:50 KaiserJohan wrote:
Got a C#/general question:

Consider I have a custom class,

public class test {
private int a,b;
private string c;

<get / set functions, and also serialize/deserialize methods so I can stick it the class object into a byte[] and send it over a socket>
}

The problem is whenever I receive data on the other side, an event is triggered executes a receieve function which reads the data from a buffer and deserializes it back into my custom class test -- however, sometimes when I read from the buffer, number of bytes equals to 1½ or 2 classes might've been in the buffer already so when I deserialize, it deserializes into 1 complete class object, but throws the rest of the data away, so I dont get complete information.

The solution would be to know how big the class is, so I can only read from the buffer up to a certain index.... but does the size of the class object vary with the size of its variables?
For example, would string c = "kekeke" generate a smaller object than c= "kekekeekkeekeke"? In that case, how do I know how much to read form the buffer?

On November 16 2010 22:46 mcc wrote:

Ahh, I didn't notice your last question
String c = "kekeke" generates object of the same size as c= "kekekeekkeekeke", but that is because the actual "string" is not stored within the object, rather that object contains pointer to the memory storing the "string".

But as I said in serialization/deserialization size of the object is not the only thing you need, you need also sizes of referenced objects and size of memory blocks referenced by those, and that is not fixed. So basically if you do not send size of deserialized object along with the data, you cannot know beforehand how big a chunk of the buffer you need. You would need to parse the buffer yourself to determine where one deserialized object ends and another begins.

Sorry if the formulation is not clear, if this is so, hopefully someone else will put it more clearly. This case is one reason I think all programmers should know C/C++ (or some other language with pointers). Because C#, Java, PHP, .... shelters you from pointers, but they are actually there all the time, and many things make so much more sense when you know that. /end of rant