Print

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

[MagmaMcFry]

For a logarithmic spiral, just keep the angle of the bearing to the spiral's center constant. For an Archimedean spiral, set the bearing to be at right angles to the center, then turn the ship by cot^-1(c * the distance to the center).

[Mego]

But a ship can't magically change its position. It has variables of bearing, acceleration, and speed which it can change to a degree.

Vectors and calculus. Assuming constant acceleration, it all boils down to a few equations:

r(t) = r₀ + v₀*t + 1/2*a*t²
(r is the position vector, r₀ is the initial position vector, v₀ is the initial velocity vector, and a is the acceleration vector)

x(t) = ||r(t)||*cos(A)
y(t) = ||r(t)||*sin(A)
(||r(t)|| is the magnitude of the position at time t, A is the angle made with the positive x axis (east), otherwise known as bearing)

For varied acceleration, it's probably easier in this case to treat acceleration like a piecewise function and reevaluate the formula at small time intervals with the acceleration and velocity for that time.

[Telgin]

I really hate working with Fortran.  I was handed a 10,000 line Fortran program and tasked with doing some painful things with it.  I only have a pretty basic understanding of Fortran, and am currently very stumped on why the program is crashing when reading from an input file.

The input file looks like this:

Code: [Select]

       3
       6       26        8
       6       26        8
       4       26        8
 0.00000000000e+00  1.01010096073e+00  2.02020192146e+00  3.03030300140e+00
 4.04040384293e+00  5.05050516129e+00  0.00000000000e+00  1.01010096073e+00
 2.02020192146e+00  3.03030300140e+00  4.04040384293e+00  5.05050516129e+00
...etc...

The code that opens and begins reading it looks like this:

Code: [Select]

      open (2,file=gridFileName, iostat=openStatus)
      if (openStatus>0) stop "**Open input file failed**"
 
      read (2,*) BloNo
      print *, 'Block No.: ',BloNo
      iTotal(1)=0

It opens the file successfully.  It also reports that the file's size is correct if I use INQUIRE on it.  Using Intel's compiler works.  If I use PGI's compiler, which I must, then it crashes at the read line claiming that it's reading past the end of the line.

I have no clue why this is happening.  By everything I'm reading it should be perfectly fine to do this, and it works using ifort.  It should read the 3 into BloNo as per list-directed input.   Why does pgfortran produce a binary that crashes here!?

[olemars]

Try status='OLD' in the open statement, just to make sure it's not creating a blank file somewhere randomly. As a general precaution I'd also recommend action='READ' if the file is precious to you somehow. The latter might actually imply status=old, I'm a bit rusty after a few years of little exposure to scientific code. The terror of FORMAT statements linger though (though apart from IO, I actually like fortran).

edit: somehow failed to read the very clear sentence where you point out you get the right file size. Then I don't know.

[Telgin]

Yeah, I tried the status=old thing too thinking the same thing about it creating a file and that didn't help.  I haven't tried adding action='read', which can't hurt anything at least.  I did try converting the code to formatted input instead to see if that helped, and it didn't.  I think I might have gotten the format specification wrong though.

Thanks anyway.  I foresee a few more painful hours in my near future...

[MrWillsauce]

[C++]
Would it be possible for me to make my program generate maps with the name of the thing they contain based on strings in a .dat file? For instance, if there are a set of values that I would categorize as "structures" I would like to have the text file set up like this:

Spoiler (click to show/hide)

STRUCTURES
Towns: 0
Farms: 10

POPULATION
Population: 4

(and so on)

and then have a program that reads through that file, making new maps for the categories, then adding keys and values to those maps based on the things under them, and then stop reading for that category when it hits whitespace. So should I make a map <string, data> or something to accomplish that?

[olemars]

There are dozens of ways to achieve that. You just need to pick one that fits what you intend to do with the categories. What you want them to be, if there is a fixed set you have in mind or it's meant to be a more generic concept, etc.

Also, whitespace can be a tricky thing to use as a delimiter.

[MagmaMcFry]

[C++]
Would it be possible for me to make my program generate maps with the name of the thing they contain based on strings in a .dat file? For instance, if there are a set of values that I would categorize as "structures" I would like to have the text file set up like this:

Spoiler (click to show/hide)

STRUCTURES
Towns: 0
Farms: 10

POPULATION
Population: 4

(and so on)

and then have a program that reads through that file, making new maps for the categories, then adding keys and values to those maps based on the things under them, and then stop reading for that category when it hits whitespace. So should I make a map <string, data> or something to accomplish that?

What you want is a map<string, map<string, int> >. Yes, that is possible. I assume you are doing this so you can print the data in order, right? Well, that's kinda not the best way to do that. Here's an alternative way to do this: Just make a format file that specifies the order of all the data values:

Code: [Select]

STRUCTURES
Towns: $towns
Farms: $farms

POPULATION
Population: $pop

(and so on)Read this file into memory at the start of your program and save it as a vector<vector<string> >, saving each line as a vector<string> of words. When you want to print your data, you just do this:

Code: [Select]

// format is the vector<vector<string> >.
for (vector<string>& formatLine : format) {
  for (string word : formatLine) {
    // If a format word begins with a $, replace it by the data value of that name (minus the $ of course).
    if (word.front() == '$') word = data[word.substr(1)];
    // os is your output stream.
    os << word << " ";
  }
  os << endl;
}

This also means that you don't need to save the description of the variable in the name of the variable, so you can write data["pop"] instead of data["Population:"].

[Telgin]

Figured out my Fortran problem, and I don't know if I should blame it or Linux.  There was a second empty file with the same name in the directory, except its file name apparently ended in an unprintable character (ls listed it as 3dminput.grd?).  I deleted that file and reran the program twice and it started working.  No idea why it took two runs... or where the file came from.

Normally I'd just cast the blame on whoever created that file, but I still want to know why Fortran was happy to tell me the correct file size and tell me that it had opened the file, when apparently it opened that other empty one.

Again, not sure if this is a quirk of Linux's file I/O system or PGI's Fortran implementation... considering that Intel's compiler produced a working binary it's probably the implementation's fault somehow.  I've noticed other interesting differences between the two compilers, such as Intel's compiler allowing you to initialize unallocated arrays, where PGI's crashes (sensibly in my opinion).  Not sure who's to blame there either.  Maybe the Fortran spec allows that.

[moghopper]

I was going to start my own thread, but this seems like a better place to post.

I'm starting an indie gaming company, and I'm currently looking for a coder to help with the development of 2D games.

The ideal candidate will be proficient in:

C++
SDL


Any kind of sound/music experience is also welcome.

If you are interested, Pm me

[miauw62]

This isn't an advertising thread, so not quite the right place.

[moghopper]

This isn't an advertising thread, so not quite the right place.

I just figured it would be the best place to find interested and dedicated individuals.

[cerapa]

This isn't an advertising thread, so not quite the right place.

I just figured it would be the best place to find interested and dedicated individuals.

Make a thread in creative projects.

Possibly with a better outline than "2D game". Something to make people want to work with and for you.

[moghopper]

This isn't an advertising thread, so not quite the right place.

I just figured it would be the best place to find interested and dedicated individuals.

Make a thread in creative projects.

Possibly with a better outline than "2D game". Something to make people want to work with and for you.

Fair enough. I'll do that then.

[ECrownofFire]

Figured out my Fortran problem, and I don't know if I should blame it or Linux.  There was a second empty file with the same name in the directory, except its file name apparently ended in an unprintable character (ls listed it as 3dminput.grd?).  I deleted that file and reran the program twice and it started working.  No idea why it took two runs... or where the file came from.

Normally I'd just cast the blame on whoever created that file, but I still want to know why Fortran was happy to tell me the correct file size and tell me that it had opened the file, when apparently it opened that other empty one.

Again, not sure if this is a quirk of Linux's file I/O system or PGI's Fortran implementation... considering that Intel's compiler produced a working binary it's probably the implementation's fault somehow.  I've noticed other interesting differences between the two compilers, such as Intel's compiler allowing you to initialize unallocated arrays, where PGI's crashes (sensibly in my opinion).  Not sure who's to blame there either.  Maybe the Fortran spec allows that.

Files in Linux consist of a series of bytes, usually assumed to be UTF-8. A few searches tell me that Fortran does not null-terminate strings, so my (completely in the dark) guess is that something somehow overran the string for some bizarre reason.