Help me make a program please..

[sam707]

video wanted?? Ok i am definitively MEAN

[Martin_fr]

in pseudo code

counting the recursion as one "loop", as it iterates every char once.

Code: Pascal  [Select][+][-]

1. cnt: array [char] of integer;
2.  
3. function filter(idx, ): string;
4. begin
5.    result := '';
6.    if idx > length(source) then exit; // fpc stores length, so no loop
7.  
8.    chr= source[idx]
9.    inc(cnt[chr]);
10.    result = filter(idx+1);
11.    if cnt[chr] = 1 then
12.       result = chr + result;  // ok, the mem move could be counted as a loop
13. end
14.  
15. res := filter(1)
16.  
17.  

to avoid the mem move when joining strings (chr + result), you can do it backward.
Start with
  res := filter(length(source))
Allocate enough mem for the result, then you can write the chr to the string starting from pos 1, into the already allocated memory. A final setlength would shorten the string, but not need to move it. So that would be one loop only.

If you must need eat something, then google "edible shoes".

[molly]

@josh:
I think i found a small issue inside your code.

I got a wrong result with testing TS example "asda", which returned "asd"

Code: Pascal  [Select][+][-]

1. function Dups(s:shortstring):String;
2. var i,k:Integer;
3.   ch:char;
4.   nw:string;
5.   le:Byte absolute s[0];
6. begin
7.   result:='';
8.   i:=1;
9.   k:=1;
10.   nw:=s[1];
11.   repeat
12.     if i=k then inc(k)  // do not check current char with itself
13.     else
14.     begin
15.       ch:=s[i];
16.       if ch=s[k] then nw:='';
17.       inc(k);
18.     end;
19.     if k>=le then  // Should this not read: if k> le ?
20.     begin
21.       inc(i);
22.       result:=result+nw;
23.       nw:=s[i];
24.       k:=1;
25.     end;
26.   until i>le;
27. end;        
28.  

Line that afaik causes the issue, highlighted in above code.

PS: in case that was on purpose e.g. to not make it too easy for TS, then please let me know and i'll remove.

[eny]

Code: Pascal  [Select][+][-]

1. const
2.   S = 'the quick brown fox jumped over the lazy dog';
3.  

jumps 

[molly]

Hey... even when it's possible... "the earth is spinning around its axes... yeah, we've got a loop there so I'm not going to eat my shoe".

There will always be sore losers. No offense intended to sam707.

Especially playing the asm card is just plain vile. As if i am responsible for whatever the compiler produces. And in case RTL does an API call and that code contains a loop then... i had the impression we were talking about Pascal loops here.

[sam707]

dang 

@Martin_fr is about to make me eat a shoe

yup .. I was asking for an "inplace" solution (because by past we spared mem a lot)

but if you have a secondary buffer it is possible to do it, in one loop, still with the help  of the  'Set' checker

OMG my stomach HURTS hahahaha

for n = 1 to len
 if char[n] not in set then
 add to buffer and add to set
   otherwise do not add anywhere and
 continue loop in any case
...
 check if 1st char in set, then do final remove (delete(buffer,1,1) in case of
 return buffer

ima old looser gona eat a shoe all by myself hahahaha
I just killed meh myself and I

[Josh]

Hi Molly,

Well spotted, was an error, should have tested a bit more.

just in case addition of strings causes a loop, I have modified the routine below.
As I have not used shortstring and this 0 byte index for size of string before; is it safe to create and add to a shortstring this way. ie does a shortstring assign a 256 byte chunk of memory when it is declared?

Code: Pascal  [Select][+][-]

1. function Dups(s:shortstring):String;
2. var i,k,sz:Integer;
3.   ch:char;
4.   nw:string;
5.   le:Byte absolute s[0];
6.   ans:shortstring;
7. begin
8.   result:='';
9.   if s='' then exit;  // exit if s is blank
10.   i:=1;
11.   k:=1;
12.   nw:=s[1];
13.   sz:=0;
14.   ans[0]:=chr(sz);
15.   repeat
16.     if i=k then inc(k)  // do not check current char with itself
17.     else
18.     begin
19.       ch:=s[i];
20.       if ch=s[k] then nw:='';
21.       inc(k);
22.     end;
23.     if k>le then
24.     begin
25.       inc(i);
26.       if nw<>'' then
27.       begin
28.         inc(sz);
29.         ans[sz]:=ch;
30.         ans[0]:=chr(sz);
31.       end;
32.      // result:=result+nw;
33.       nw:=s[i];
34.       k:=1;
35.     end;
36.   until i>le;
37.   result:=ans;
38. end;              
39.  

[sam707]

@molly noone is going to offend an old gamer, you are welcome as we are kidding

[sam707]

 

I'm thinking... hmmm

now the HAPPY one should be @irppaann with his newbie school assignment he's gonna write 3 books

[molly]

Hi Molly,

Well spotted, was an error, should have tested a bit more.

No problem.

Another one for you to wrap it up: pass an empty string and see what happens 

As I have not used shortstring and this 0 byte index for size of string before; is it safe to create and add to a shortstring this way. ie does a shortstring assign a 256 byte chunk of memory when it is declared?

According to documentation:

If the size of the string is not specified, 255 is taken as a default. The actual length of the string can be obtained with the Length standard runtime routine. For example in

and

For short strings, the length is stored in the character at index 0. Old Turbo Pascal code relies on this, and it is implemented similarly in Free Pascal.

Despite this, to write portable code, it is best to set the length of a shortstring with the SetLength call, and to retrieve it with the Length call. These functions will always work, whatever the internal representation of the shortstrings or other strings in use: this allows easy switching between the various string types.

So, to answer your question: yes: your code will do just fine. It is an array of 255 characters (while rest of code depends on the length byte to obtain real number of characters from it).

You can even use a absolute byte variable to you ans shortstring variable and change that iin case that is more convenient *wink*

PS: i was already impressed with your code using the length function. This shortstring stuff is from old TP days and in the ends is very limited (and might not even be future proof).

[molly]

@molly noone is going to offend an old gamer, you are welcome as we are kidding

No problem, you are correct there......  (but, it still hurts       )

...
 check if 1st char in set, then do final remove (delete(buffer,1,1) in case of
 return buffer

I was under the impression that delete() function also uses a loop to accomplish its task ?

[Josh]

So my final tweaked code below

Code: Pascal  [Select][+][-]

1. function Dups(s:shortstring):String;
2. var i,k:Integer;
3.   ch:char;
4.   nw:string;
5.   le:Byte absolute s[0];
6.   ans:shortstring;
7.   sz:Byte absolute ans[0];
8.  
9. begin
10.   if le<=1 then
11.   begin
12.     result:=s;
13.     exit;
14.   end;
15.   result:='';
16.   i:=1;
17.   k:=1;
18.   nw:=s[1];
19.   sz:=0;
20.   repeat
21.     if i=k then inc(k)  // do not check current char with itself
22.     else
23.     begin
24.       ch:=s[i];
25.       if ch=s[k] then nw:='';
26.       inc(k);
27.     end;
28.     if k>le then
29.     begin
30.       inc(i);
31.       if nw<>'' then
32.       begin
33.         inc(sz);
34.         ans[sz]:=ch;
35.       end;
36.       nw:=s[i];
37.       k:=1;
38.     end;
39.   until i>le;
40.   result:=ans;
41. end;            
42.  

[sam707]

@josh nice impressive oldschool snippet

but k is multiple loops in i primary loop!

so far my prefered snippet still is the one from @Bart which use a Set

dont panic, it's impossible in one only loop because EACH chr needs to be checked againts all the others

i'll spare my money and not buy new shoes 

[molly]

@josh:
neat 

Only goes wrong now for passing a single character for string. I'm currently too tired to see what goes wrong. I'll have a fresh look after some good nights rest 

[sam707]

'in fine'
each not read chr of the array/string needs a loop against the others remaining forward, but... the author of the post didn't ask for 'one loop only', its a @Bart assertion ... so its all good