Rewrote getFileStringSize

[charles-ramsey]

-Function now implements binary search instead of that if-clause mess. -Constants are now defined for 1 KB, 1 MB, 1 GB instead of repeated 1024 multiplication/division. -String contains a double formatted to 1 decimal place; the previous code attempted to print 3 decimal places but used integer division. -Removed the file_size <= 0 check. We should print "0 B" instead of "", and negative file_sizes should throw an error elsewhere.

[benwaffle]

consider #34

[charles-ramsey]

Sorry, I didn't see that. I'm new to git. Anyway, I think the log approach is cleaner, though the pow(1024,offset) call is unnecessary as the possible values [1,1024,1024^2,1024^3] could easily be precomputed and accessed like the suffix array.

I just did some benchmarking with generated normally-distributed data and while my code is faster, the difference isn't that significant. 1M iterations took 1.2s vs. 1.6s for the log version using the -O2 flag and g++.

[benwaffle]

I guess then it doesn't matter. We should choose the one that is easy to understand and is concise.

[benwaffle]

yet another suggestion

void scaledata(double prefix, double in, double *data, const char **unitp) {
    static int i;
    static const char units[2][8][4] = {
        { "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB" },
        { "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB" }
    };
    *data = in;
    for (i = 0; *data > prefix && i < 7; i++)
        *data /= prefix;
    *unitp = units[prefix==1024.0][i];
}

scaledata(1024, 2048, &outamns, &unitpointer);
outamnt = 2
outpointer = MiB

scaledata(1000, 2000, &outamns, &unitpointer);
outamnt = 2
outpointer = MB

[charles-ramsey]

That's equivalent to the logarithm method, only more difficult to read.

[fuyukaidesu]

We are merging this one or #44 ?