• Jos van Uden (13/13) Jan 28 2012 import std.stdio, std.stream, std.string, std.range;
• Mantis (6/19) Jan 28 2012 The same may be done without memory duplication by changing comparison
• Jos van Uden (18/39) Jan 28 2012 Good idea.
• Andrei Alexandrescu (5/46) Jan 28 2012 Could you please also compare with code that uses foreach with
• Jos van Uden (18/34) Jan 28 2012 After having added the upTo param, which is an improvement on both
• Joshua Niehus (3/17) Jan 28 2012 I ran this code on Mac OSX Lion using the "/usr/share/dict/words"
• Joshua Niehus (3/6) Jan 28 2012 found the problem: PEBKAC (problem exists between keyboard and
• Manfred Nowak (3/4) Jan 28 2012 Me got an unicode-error on that file on line 100.
• Jos van Uden (7/11) Jan 29 2012 The original file is encoded ISO 8859-1. You'll have to convert
• Andrei Alexandrescu (3/4) Jan 28 2012 if (line.walkLength(2) > 1) {
• Andrei Alexandrescu (5/18) Jan 28 2012 Thanks, Jos.

Jos van Uden <user domain.invalid> writes:

import std.stdio, std.stream, std.string, std.range;

void main() {
     int countPalindromes;
     auto infile = new BufferedFile("unixdict.txt");
     foreach (char[] line; infile) {
         if (line.walkLength > 1) {
             line.toLowerInPlace;
             if (line == line.dup.reverse)
                 countPalindromes++;
         }
     }
     writeln("palindromes found: ", countPalindromes);
}

Mantis <mail.mantis.88 gmail.com> writes:

28.01.2012 15:31, Jos van Uden пишет:
 import std.stdio, std.stream, std.string, std.range;

 void main() {
 int countPalindromes;
 auto infile = new BufferedFile("unixdict.txt");
 foreach (char[] line; infile) {
 if (line.walkLength > 1) {
 line.toLowerInPlace;
 if (line == line.dup.reverse)
 countPalindromes++;
 }
 }
 writeln("palindromes found: ", countPalindromes);
 }

The same may be done without memory duplication by changing comparison 
to this (equal function is in std.algorithm):

if (equal(line, retro(line)))

Hard to say without profiling if it will have actual impact on 
performance, but at least it should be more GC-friendly.

Jos van Uden <user domain.invalid> writes:

On 28-1-2012 14:57, Mantis wrote:
 28.01.2012 15:31, Jos van Uden пишет:
 import std.stdio, std.stream, std.string, std.range;

 void main() {
 int countPalindromes;
 auto infile = new BufferedFile("unixdict.txt");
 foreach (char[] line; infile) {
 if (line.walkLength > 1) {
 line.toLowerInPlace;
 if (line == line.dup.reverse)
 countPalindromes++;
 }
 }
 writeln("palindromes found: ", countPalindromes);
 }

 The same may be done without memory duplication by changing comparison
 to this (equal function is in std.algorithm):

 if (equal(line, retro(line)))

 Hard to say without profiling if it will have actual impact on
 performance, but at least it should be more GC-friendly.


Good idea.

It's also faster. I tried with a larger file (ukacd17.txt) which
has 240,000 entries.

---

import std.stdio, std.stream, std.string, std.range, std.algorithm;

void main() {
     int countPalindromes;
     auto infile = new BufferedFile("unixdict.txt");
     foreach (char[] line; infile) {
         if (line.walkLength > 1) {
             line.toLowerInPlace;
             if (equal(line, retro(line)))
                 countPalindromes++;
         }
     }
     writeln("palindromes found: ", countPalindromes);
}

Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:

On 1/28/12 8:21 AM, Jos van Uden wrote:
 On 28-1-2012 14:57, Mantis wrote:
 28.01.2012 15:31, Jos van Uden пишет:
 import std.stdio, std.stream, std.string, std.range;

 void main() {
 int countPalindromes;
 auto infile = new BufferedFile("unixdict.txt");
 foreach (char[] line; infile) {
 if (line.walkLength > 1) {
 line.toLowerInPlace;
 if (line == line.dup.reverse)
 countPalindromes++;
 }
 }
 writeln("palindromes found: ", countPalindromes);
 }

 The same may be done without memory duplication by changing comparison
 to this (equal function is in std.algorithm):

 if (equal(line, retro(line)))

 Hard to say without profiling if it will have actual impact on
 performance, but at least it should be more GC-friendly.


 Good idea.

 It's also faster. I tried with a larger file (ukacd17.txt) which
 has 240,000 entries.

 ---

 import std.stdio, std.stream, std.string, std.range, std.algorithm;

 void main() {
 int countPalindromes;
 auto infile = new BufferedFile("unixdict.txt");
 foreach (char[] line; infile) {
 if (line.walkLength > 1) {
 line.toLowerInPlace;
 if (equal(line, retro(line)))
 countPalindromes++;
 }
 }
 writeln("palindromes found: ", countPalindromes);
 }

Could you please also compare with code that uses foreach with 
stdin.byLine()?


Thanks,

Andrei

Jos van Uden <user domain.invalid> writes:

On 28-1-2012 16:07, Andrei Alexandrescu wrote:

 import std.stdio, std.stream, std.string, std.range, std.algorithm;

 void main() {
 int countPalindromes;
 auto infile = new BufferedFile("unixdict.txt");
 foreach (char[] line; infile) {
 if (line.walkLength > 1) {
 line.toLowerInPlace;
 if (equal(line, retro(line)))
 countPalindromes++;
 }
 }
 writeln("palindromes found: ", countPalindromes);
 }

 Could you please also compare with code that uses foreach with
 stdin.byLine()?

After having added the upTo param, which is an improvement on both
methods, the infile is (marginally) faster on my system than the byLine.

win7 64-bits i7 2600, dmd 2.057, optimizations enabled: -O -inline -release

--

import std.stdio, std.stream, std.string, std.range, std.algorithm;

void main() {
     int countPalindromes;
     auto infile = new BufferedFile("ukacd17.txt");
     foreach (char[] line; infile) {
         if (line.walkLength(2) > 1) {
             line.toLowerInPlace;
             if (equal(line, retro(line)))
                 countPalindromes++;
         }
     }
     writeln("palindromes found: ", countPalindromes);
}

"Joshua Niehus" <jm.niehus gmail.com> writes:

 import std.stdio, std.stream, std.string, std.range, 
 std.algorithm;

 void main() {
    int countPalindromes;
    auto infile = new BufferedFile("ukacd17.txt");
    foreach (char[] line; infile) {
        if (line.walkLength(2) > 1) {
            line.toLowerInPlace;
            if (equal(line, retro(line)))
                countPalindromes++;
        }
    }
    writeln("palindromes found: ", countPalindromes);
 }

I ran this code on Mac OSX Lion using the "/usr/share/dict/words" 
file and got 235834 words out of 235886.  I think something is 
wrong.

"Joshua Niehus" <jm.niehus gmail.com> writes:

 I ran this code on Mac OSX Lion using the 
 "/usr/share/dict/words" file and got 235834 words out of 
 235886.  I think something is wrong.

found the problem: PEBKAC (problem exists between keyboard and 
chair)
sorry:)

Manfred Nowak <svv1999 hotmail.com> writes:

Jos van Uden wrote:

 BufferedFile("ukacd17.txt");

Me got an unicode-error on that file on line 100.

-manfred

Jos van Uden <user domain.invalid> writes:

On 29-1-2012 7:55, Manfred Nowak wrote:
 Jos van Uden wrote:

 BufferedFile("ukacd17.txt");

 Me got an unicode-error on that file on line 100.

 -manfred

The original file is encoded ISO 8859-1. You'll have to convert
it to utf-8.

My earlier code sample used the unixdict.txt that I found
here http://www.puzzlers.org/pub/wordlists/unixdict.txt

I only changed to ukacd17.txt to test performance.

Jos

Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:

On 1/28/12 8:21 AM, Jos van Uden wrote:
 if (line.walkLength > 1) {

if (line.walkLength(2) > 1) {


Andrei

Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:

On 1/28/12 7:31 AM, Jos van Uden wrote:
 import std.stdio, std.stream, std.string, std.range;

 void main() {
 int countPalindromes;
 auto infile = new BufferedFile("unixdict.txt");
 foreach (char[] line; infile) {
 if (line.walkLength > 1) {
 line.toLowerInPlace;
 if (line == line.dup.reverse)
 countPalindromes++;
 }
 }
 writeln("palindromes found: ", countPalindromes);
 }

Thanks, Jos.

We still don't have the code that randomly rotates the homepage samples. 
Any volunteers? If not, I'll implement it in Javascript.


Andrei