# Sign that word 2!

Not that long ago, I posted a challenge called Sign that word!. In the challenge, you must find the signature of word, which is the letters put in order (e.g. The signature of this is hist). Now then, that challenge did quite well, but there was one key issue: it was WAY too easy (see the GolfScript answer). So, I've posted a similar challenge, but with more rules, most of which have been suggested by PPCG users in the comments on the previous puzzle. So, here we go!

## Rules

1. Your program must take an input, then output the signature to STDOUT or the equivalent in whatever language your using.
2. You are not allowed to use built-in sorting functions, so stuff like $ in GolfScript is not allowed. 3. Multicase must be supported - your program must group letters of both uppercase and lowercase together. So the signature of Hello is eHllo, not Hello as you are given by the GolfScript answer on the first version. 4. There must be a free interpreter/compiler for your program, which you should link to. ## Scoring Your score is your byte count. Lowest byte count wins. ## Leaderboard Here is a Stack Snippet to generate both a regular leaderboard and an overview of winners by language. To make sure that your answer shows up, please start your answer with a headline, using the following Markdown template: # Language Name, N bytes  where N is the size of your submission. If you improve your score, you can keep old scores in the headline, by striking them through. For instance: # Ruby, <s>104</s> <s>101</s> 96 bytes  var QUESTION_ID=55090;function answersUrl(e){return"http://api.stackexchange.com/2.2/questions/"+QUESTION_ID+"/answers?page="+e+"&pagesize=100&order=desc&sort=creation&site=codegolf&filter="+ANSWER_FILTER}function getAnswers(){jQuery.ajax({url:answersUrl(page++),method:"get",dataType:"jsonp",crossDomain:!0,success:function(e){answers.push.apply(answers,e.items),e.has_more?getAnswers():process()}})}function shouldHaveHeading(e){var a=!1,r=e.body_markdown.split("\n");try{a|=/^#/.test(e.body_markdown),a|=["-","="].indexOf(r[1][0])>-1,a&=LANGUAGE_REG.test(e.body_markdown)}catch(n){}return a}function shouldHaveScore(e){var a=!1;try{a|=SIZE_REG.test(e.body_markdown.split("\n")[0])}catch(r){}return a}function getAuthorName(e){return e.owner.display_name}function process(){answers=answers.filter(shouldHaveScore).filter(shouldHaveHeading),answers.sort(function(e,a){var r=+(e.body_markdown.split("\n")[0].match(SIZE_REG)||[1/0])[0],n=+(a.body_markdown.split("\n")[0].match(SIZE_REG)||[1/0])[0];return r-n});var e={},a=1,r=null,n=1;answers.forEach(function(s){var t=s.body_markdown.split("\n")[0],o=jQuery("#answer-template").html(),l=(t.match(NUMBER_REG)[0],(t.match(SIZE_REG)||[0])[0]),c=t.match(LANGUAGE_REG)[1],i=getAuthorName(s);l!=r&&(n=a),r=l,++a,o=o.replace("{{PLACE}}",n+".").replace("{{NAME}}",i).replace("{{LANGUAGE}}",c).replace("{{SIZE}}",l).replace("{{LINK}}",s.share_link),o=jQuery(o),jQuery("#answers").append(o),e[c]=e[c]||{lang:c,user:i,size:l,link:s.share_link}});var s=[];for(var t in e)e.hasOwnProperty(t)&&s.push(e[t]);s.sort(function(e,a){return e.lang>a.lang?1:e.lang<a.lang?-1:0});for(var o=0;o<s.length;++o){var l=jQuery("#language-template").html(),t=s[o];l=l.replace("{{LANGUAGE}}",t.lang).replace("{{NAME}}",t.user).replace("{{SIZE}}",t.size).replace("{{LINK}}",t.link),l=jQuery(l),jQuery("#languages").append(l)}}var ANSWER_FILTER="!t)IWYnsLAZle2tQ3KqrVveCRJfxcRLe",answers=[],page=1;getAnswers();var SIZE_REG=/\d+(?=[^\d&]*(?:&lt;(?:s&gt;[^&]*&lt;\/s&gt;|[^&]+&gt;)[^\d&]*)*$)/,NUMBER_REG=/\d+/,LANGUAGE_REG=/^#*\s*([^,]+)/;
body{text-align:left!important}#answer-list,#language-list{padding:10px;width:290px;float:left}table thead{font-weight:700}table td{padding:5px}
<script src="https://ajax.googleapis.com/ajax/libs/jquery/2.1.1/jquery.min.js"></script><link rel="stylesheet" type="text/css" href="//cdn.sstatic.net/codegolf/all.css?v=83c949450c8b"><div id="answer-list"> <h2>Leaderboard</h2> <table class="answer-list"> <thead> <tr><td></td><td>Author</td><td>Language</td><td>Size</td></tr></thead> <tbody id="answers"> </tbody> </table></div><div id="language-list"> <h2>Winners by Language</h2> <table class="language-list"> <thead> <tr><td>Language</td><td>User</td><td>Score</td></tr></thead> <tbody id="languages"> </tbody> </table></div><table style="display: none"> <tbody id="answer-template"> <tr><td>{{PLACE}}</td><td>{{NAME}}</td><td>{{LANGUAGE}}</td><td>{{SIZE}}</td><td><a href="{{LINK}}">Link</a></td></tr></tbody></table><table style="display: none"> <tbody id="language-template"> <tr><td>{{LANGUAGE}}</td><td>{{NAME}}</td><td>{{SIZE}}</td><td><a href="{{LINK}}">Link</a></td></tr></tbody></table>

• Must lowercase and uppercase be ordered? For example, for ThHihs, can we output hHhist or do we have to output hhHist or Hhhist? Aug 21, 2015 at 7:20
• @Kslkgh I do not wish to install a GolfScript interpreter in order to understand the question. The requirements should be clearly specified in the question itself. Aug 21, 2015 at 7:29
• @feersum You don't need to. There is a link in the answer to an online interpreter. But I will make it clearer.
– user42003
Aug 21, 2015 at 7:35
• Correctly handling lower/upper case in Unicode is scary, so is this question voluntarily limited to ASCII letters: [a-zA-Z]? Aug 21, 2015 at 9:03
• You forgot to close the parenthesis after "see this GolfScript answer". xkcd.com/859 Aug 21, 2015 at 9:24

# Pyth, 10 bytes

sm@+drd1zG


Try it online: Demonstration

### Explanation:

             implicit: z = input string
m       G   map each letter (variable d) of the alphabet to:
+drd1        d + upper(d)
@     z       filter z for these two letters
s            sum, join to a string


f s=[x|(a,b)<-zip['a'..'z']['A'..],x<-s,x==a||x==b]


The zip creates a list of pairs of characters [('a','A'), ...('z','Z')]. Because of truncation, the second endpoint doesn't need to be specified. For each pair in the list, we take the letters in the input string s that are either of the two characters in the pair.

# Python 3, 72 70 bytes

s=input()
print("".join(d*(ord(d)&31==c)for c in range(27)for d in s))


Assumes the input consists of only [a-zA-Z].

(-2 bytes thanks to @xnor)

• I think you can do "".join(c*b ...) for "".join(c ... if b) the same way sum(n ... if b) can be sum(n*b ...).
– xnor
Aug 22, 2015 at 1:46
• Collapsing the two loops into a single loop with a sentinel, and compressing the equality check: c=1 for d in(input()+'~')*26:print(d[ord(d)&32^c:],end='');c+=d>'z' (67)
– xnor
Aug 22, 2015 at 2:20
• @xnor I've added the first tip (thanks!) but I think you should post the second as a separate answer :) Aug 22, 2015 at 3:24

# GOTO++, 531 bytes

niveaugourou 0
s=ENTRETONTEXTE()
§2 a=LeCaracNumero()&s *(1)
n=*(1)
costaud i=*(2)/&i infeg NombreDeLettres(&s)/i=+*(1)
b=LeCaracNumero()&s &i
c=&b
d=&a
GOTONULPOURLESNULS %4 }&b sup *(96){
c=-*(32)
§4 GOTONULPOURLESNULS %5 }&a sup *(96){
d=-*(32)
§5 GOTONULPOURLESNULS %1 }&c inf &d{
a=&b
n=&i
§1 faiblard
GOTOPRINTDUTEXTE()&a
s=Marijuana()}BOITEAPINGOUINS()}PrendsUnMorceau()&s *(0) &n{ }PrendsUnMorceau()&s }&n+*(1){ *(0){{ «»
GOTONONNULPOURLESNULS %3 }NombreDeLettres(&s) eg *(1){
GOTOPASMALIN %2
§3 GOTOPRINTDUTEXTE()&s


GOTO++ Project page

Here's a slighlty more readable and commented version of the code(Note that GOTO starts a comment in GOTO++):

niveaugourou 0                                          GOTO Allow every keyword to be used
§2 a=LeCaracNumero()&s *(1)                             GOTO Get first char in s
n=*(1)
costaud i=*(2)/&i infeg NombreDeLettres(&s)/i=+*(1)     GOTO Loop on every char of s
b=LeCaracNumero()&s &i                                  GOTO b = i-th char
c=&b
d=&a
GOTONULPOURLESNULS %4 }&b sup *(96){                    GOTO If b is uppercase, goto §4
c=-*(32)                                                GOTO Get the uppercase ASCII value of b
§4 GOTONULPOURLESNULS %5 }&a sup *(96){                 GOTO same as above but with a
d=-*(32)
§5 GOTONULPOURLESNULS %1 }&c inf &d{                    GOTO If b is after a in alphabetical order, goto §1 (next loop iteration)
a=&b                                                    GOTO Else replace a by b
n=&i
§1 faiblard                                             GOTO End loop
GOTOPRINTDUTEXTE()&a                                    GOTO Print the value of a
t=PrendsUnMorceau()&s *(0) &n                           GOTO Get the part of s before a
u=PrendsUnMorceau()&s }&n+*(1){ *(0)                    GOTO Get the part of s after a
e=BOITEAPINGOUINS()&t &u                                GOTO Create an array of penguins containing the two substrings
s=Marijuana()&e «»                                      GOTO Concatenate the penguins in the array
GOTONONNULPOURLESNULS %3 }NombreDeLettres(&s) eg *(1){  GOTO If s is one char long, goto §3
GOTOPASMALIN %2                                         GOTO Else goto §2
§3 GOTOPRINTDUTEXTE()&s                                 GOTO Print the last char

• Marks of a great programming language: all commands are in French and the official website says "Send your logo et look dumb." Aug 22, 2015 at 19:06
• @AlexA. Mark of a great programming language: function(arg) and function() arg are both valid. Also, priority brackets are } { and not the boring ( ) Aug 22, 2015 at 19:19

# Pyth, 15 14 bytes

s*V/LzJ.irG1GJ


Thanks for isaacg for removing 1 byte.

I don't know much about Pyth yet, so this may not be golfed well.

Try it here.

• One easy golf: s is the same as jk on a list of strings. Aug 21, 2015 at 8:08

# JavaScript (ES6), 71 74

Limited to A-Za-z (see comment by @Matthieu M)

Edit Too used to compose a single expression with commas, to avoid 'return'. Here an output is required, so I can use a simple for and forget about commas.

Using array comprehension the byte count is 73, but that is not valid EcmaScript 6 anymore

Usual note: test running the snippet on any EcmaScript 6 compliant browser (notably not Chrome not MSIE. I tested on Firefox, Safari 9 could go)

f=w=>{v=[];for(c of w)v[n=parseInt(c,36)]=(v[n]||'')+c;alert(v.join)}
<input id=I value='Hellzapoppin'><button onclick=f(I.value)>-></button>

## Javascript, 112 194 bytes

r=[];t=[];a=s.split('').map(function(x){t[i=parseInt(x,36)]?t[i].push(x):t[i]=[x];return i;});while(l=a.length)r[l-1]=t[a.splice(a.indexOf(Math.max.apply({},a)),1)].pop();console.log(r.join(''))


This is far away from "golfed" but I'm a bit busy right now, just editted to remove sort.

• @frikinside While it might not help this solution since sorting was disallowed, Javascript ES6 arrow functions could help shorten this. (Not all browsers support ES6 yet, try Firefox). Maybe it'll help your future solutions though! :) An example of what it might look like: console.log(input.split('').sort((a,b)=>a.toLowerCase().localeCompare(b.toLowerCase())).join('')) Aug 21, 2015 at 14:36
• @UndefinedFunction Actually I choose to use plain javascript on purpose, for the "challenge" but thx a lot for the nice suggestion! Aug 24, 2015 at 11:19
• @edc65 actually I was aware of that in a previous comment (like u said) and I did not forget it, I wasn't have time until now. Aug 24, 2015 at 11:23

## Python 3, 64

A small improvement on Sp3000's answer, which uses the idea of iterating iterating the character indices, and for each one, iterating through the input to take characters that match up to case.

c=1
for d in(input__+'~')*26:print(end=d[ord(d)&31^c:]);c+=d>'z'


This uses a single loop, looping through the input 26 times. The separator ~ is used to know when to go to the next character index c. To whether character d matches value c up to case, the last five bits of the bit-value of d are xor-ed with c, with a 0 indicating a match.

Then, the character d is printed exactly when the result is 0, with an empty string otherwise.

## Python 2.7, 114 106 bytes

l=[0]*123
for e in raw_input():l[ord(e)]+=1
print''.join(chr(j)*l[j]for i in range(26)for j in(i+65,i+97))


Logs the presence of a char in a 123 length array(for including both A-Z and a-z ranges) and then iterates through it to get the non-zero entries.
Inefficient, but more efficient than brute forcing it(but longer :().

Testing it-

<< HelloWorldhi
>> deHhillloorW

• When I run it with HelloWorldhi, I get ['d', 'e', 'H', 'h', 'i', 'lll', 'oo', 'r', 'W']
– Blue
Aug 21, 2015 at 11:09
• @muddyfish oops my bad. I pasted the wrong code while trying out a couple of things. Thanks for that :) Aug 21, 2015 at 11:27
• You don't need the if l[j] either.
– Blue
Aug 21, 2015 at 11:47
• @muddyfish Nice, they become empty strings(and are removed when joined). -8 bytes. Thanks :) Aug 21, 2015 at 11:53
• @SirParselot Yeah, but that seemed like modifying the input as given in the question, so didn't do that Aug 21, 2015 at 13:17

# PHP, 275 270 bytes

<?php
for($i=65;$i<123;$i++){$v[$i]=chr($i);}foreach(str_split($argv[1])as$c){$a=array_search($c,$v);if($a<97){$p[]=($a+32);$z[]=$a;}else{$p[]=$a;}}foreach($p as$chr){$m=min($p);if($z[0]+32==$m){echo chr($m-32);unset($z[0]);}else{echo chr($m);}unset($p[array_search($m,$p)]);}


Explanation:
The code generates an array with every letter in the alphabet, which has its ASCII Value as Array Key. Afterwards the code generates a new array which contains the ASCII Values of the Input. Then the lowest Value gets printed out and gets removed.

Usage:
Call the script with an argument: php -d error_reporting=0 script.php Hello

Ungolfed Version:

<?php
$input =$argv[1];
$valueArray = []; for($i=65;$i<123;$i++) {
$valueArray[$i] = chr($i); }$new = str_split($input); foreach($new as $char) { if(array_search($char, $valueArray)<97) {$newArray[] = (array_search($char,$valueArray)+32);
$checkArray[] = array_search($char, $valueArray); } else {$newArray[] = array_search($char,$valueArray);
}
}
foreach($newArray as$chr) {
if($checkArray[0]+32 == min($newArray)) {
$string .= chr(min($newArray)-32);
unset($checkArray[0]); } else {$string .= chr(min($newArray)); }$key = array_search(min($newArray),$newArray);
unset($newArray[$key]);
}

echo $string;  Any advices are greatly appreciated. • Uppercase letters are not working ... need to fix this. Aug 21, 2015 at 8:16 • Fixed that the code also works when the Input has Uppercase letters. Aug 21, 2015 at 14:31 # Haskell, 83 53 bytes import Data.Char f y=[c|x<-[' '..],c<-y,toLower c==x]  Usage: f "HelloWorldhi" -> "deHhillloorW". How it works: let y be the input string [ |x<-[' '..] ] -- for every x from Space to the last Unicode character ,c<-y -- loop through all character c from the input string c ,toLower c==x -- and keep those where the lowercase version equals x  Edit: 30 bytes saved, imagine that! Thanks @Mauris. • Why not just import Data.Char;f y=[c|x<-[' '..],c<-y,toLower c==x] (53 bytes)? (This will take some time to terminate, because length [' '..] == 1114080 -- but it's finite.) – Lynn Aug 21, 2015 at 22:07 • @Mauris: Wow! Btw: 8.5sec on a 4 year old laptop is not too long. – nimi Aug 21, 2015 at 22:48 # Python 3, 61 bytes A new answer for a different technique! z=['']*42 for c in input():z[ord(c)&31]+=c print(*z,sep='')  Noting that ord('a')&31==ord('A')&31 and that ord('z')&31==ord('Z')&31, we can simply create an array of empty strings and for each character add it to the array index of its ASCII value &31. When you print it, it will be sorted. Limited to input a-zA-Z. # Python 3, 97 92 bytes from itertools import*;print(*min(permutations(input()),key=lambda z:str(z).lower()),sep='')  The best way to sort is clearly to generate all permutations, and then pick the minimum, which just happens to be sorted :) The strings are lowercased before comparison to abide by the 'case-aware' rules. Warning: may be very slow with big strings. An interpreter is located here. # Python 3, 118 bytes i=input();i,x=map(list,(i,i.lower())) while x:q=min(x);x.remove(q);q=[q.upper(),q][q in i];i.remove(q);print(q,end="")  Could be golfed much shorter, I know • You can replace if q not in i: with if~-(q in i):. Aug 21, 2015 at 10:41 • I do because I'm calling the .remove() function on it. – Blue Aug 21, 2015 at 10:57 # Powershell, 164 Bytes I'm sure there's a cleaner way to do this, but I couldn't come up with anything else. Just takes the input as a character array, does an insertion sort, and spits out the output. Loses horribly, even to other non-golfing languages. ### Code: $a=[char[]]($args[0]);For($u=1;$u-lt$a.Count;$u++){$n=$a[$u];$l=$u;while(($l-gt0)-and((""+$a[$l-1]).CompareTo(""+$n)-gt0)){$a[$l]=$a[$l-1];$l--}$a[$l]=$n};$a-join''  ### Usage: PS C:\scripts> .\sign-word-2.ps1 tTHhis hHistT  ### Expanded & Explained: $a=[char[]]($args[0]) # Takes command-line argument, recasts as char array For($u=1;$u-lt$a.Count;$u++){ # Performs a quick-n-dirty insertion sort$n=$a[$u]
$l=$u
while(($l-gt0)-and((""+$a[$l-1]).CompareTo(""+$n)-gt0)){
# Ugly, ugly code here. String.CompareTo(String) is case-insensitive, but
# because we cast as a char[], Char.CompareTo(Char) is case-sensitive ...
# So, need to do an on-the-fly re-casting as a string with ""+
$a[$l]=$a[$l-1]
$l-- }$a[$l]=$n
}
$a-join'' # Without the -join'', it would print out the chars with a space between  # Julia, 61 bytes f=s->s>""?(k=indmax([s...]%32);f(s[k+1:end]s[1:k-1])s[k:k]):s  Julia will display it as a string output if you call it in the REPL. If it has to print out to STDOUT, then it needs 78 bytes: x->(f=s->s>""?(k=indmax([s...]%32);f(s[k+1:end]s[1:k-1])s[k:k]):s;print(f(x)))  An interpreter for Julia can be found here. Another one, which I've already put some code into, is here. Note that, with the second one, you will need to make the terminal (at the bottom) visible by dragging the boundary up. Clicking "execute" will make it run in the terminal at the normal command line (and thus won't show the output if called without println). Alternatively, you can just type julia into the terminal itself, then handle everything inside the REPL that will come up. And for a bit of extra fun, here are some other implementations Gnome Sort (83 bytes): s->(for m=2:endof(s),n=m:-1:2 s[n]%32<s[n-1]%32&&(s=s[[1:n-2,n,n-1,n+1:end]])end;s)  My own sorting algorithm (84 bytes): s->(k=1;while length(k)>0 k=find(diff([s...]%32).<0);s=s[setdiff(1:end,k)]s[k]end;s)  # Scala, 82 bytes print((""/:args(0)){case(s,c)=>val(a,b)=s.span(h=>{if(h<97)32 else 0}+h<c);a+c+b})  from command line: $ scala -e 'print((""/:args(0)){case(s,c)=>val(a,b)=s.span(h=>{if(h<97)32 else 0}+h<c);a+c+b})' Hello
eHllo


probably can be golfed a bit further... just implementing insertion sort using fold.

# x86 machine code, 51 42 bytes

00000000  b3 82 89 da 8b 07 80 fc  0d 74 12 b9 20 20 09 c1  |.........t..  ..|
00000010  38 e9 7e 06 86 c4 89 07  31 d2 43 eb e7 85 d2 74  |8.~.....1.C....t|
00000020  df c6 47 01 24 b4 09 cd  21 c3                    |..G.$...!.| 0000002a  Bubble sort, with some register reuse tricks to shave bytes here and there; the .COM file runs in DosBox, receives the input from command line and prints the output to standard output. Commented assembly:  org 100h section .text start: ; bubble sort - external loop ext: ; start from the first character (assume bh=0, true on every DOS) mov bl,82h ; "not-swapped" flag - 82h => no swaps in current iteration; ; 0 => a swap happened (the 82h will come in handy later) mov dx,bx ; bubble sort - internal loop int: ; read 2 characters at time in the full ax mov ax,word[bx] ; al ah ; ^[bx] ^[bx+1] ; check if we are at the end (the command line is CR terminated) cmp ah,0dh je skip ; make uppercase in cx mov cx,2020h or cx,ax ; compare cmp cl,ch jle next ; wrong order - swap and rewrite xchg al,ah mov word[bx],ax ; mark that we did a swap xor dx,dx next: ; next character inc bx jmp int skip: ; loop as far as we swapped something test dx,dx jz ext end: ;$-terminate the string
mov byte[bx+1],'$' ; print ; dx already contains the location of the string, since that's the ; flag value we used for "no swaps" mov ah,9 int 21h ret  # Java (JDK 10), 125 bytes s->{for(int i=s.length,j;i-->1;)for(j=i;j-->0;)s[i]^=(s[i]&95)<(s[j]&95)?s[j]^(s[j]=s[i]):0;System.out.print(new String(s));}  Try it online! Using a naive sort. # Perl, 88 bytes @_=/./g;a:{for(0..@_-2){@_[$_,$_+1]=@_[$_+1,$_],redo a if uc$_[$_]gt uc$_[$_+1]}}print@_  Just a simple Bubble Sort. Call with -n option to pass the text. e.g: echo "tThHiIsS" | perl -n sort2.pl  Output: hHiIsStT  # PHP, 106 bytes The code: $c=count_chars($argv[1]);$r=str_repeat;for($i=64;++$i<91;)echo$r(chr($i),$c[$i]),$r(chr($i+32),$c[$i+32]);


There is nothing special in the code; count_chars() produces an array indexed by ASCII codes that contains the number of occurrences for each ASCII character. The rest is a dull iteration over this array.

Example of execution:

$php -d error_reporting=0 sign.php qwertyuioplkjhgfdsazxcvbnmQWERTYUIOPLKasdfJHGFDSAZXCVBNM AaaBbCcDddEeFffGgHhIiJjKkLlMmNnOoPpQqRrSssTtUuVvWwXxYyZz  An extra byte can be saved using PHP 7: replace $c[$i] with ($c=count_chars($argv[1]))[$i] and remove the assignment of $c from the beginning of the program. # Haskell, 74 bytes l=(mod32).fromEnum f=foldr(#)"" e#[]=[e] e#a@(h:t)|l e<l h=e:a|1<2=h:e#t  Completely different from my other answer. This time it's a simple Insertion sort. # Pip, 18 14 bytes GitHub repository for Pip Seems there's no competing with Pyth, but this is pretty respectable. FcAZ OcQUC_FIa  Works only on strings containing a-zA-Z. For each letter of the alphabet, uses a filter operation to grab the letters from the input string that equal that letter case-insensitively:  a <- cmdline arg, AZ <- string containing uppercase alphabet (implicit) FcAZ For each character c in AZ: FIa Filter characters of a on the following lambda function: UC_ Uppercase of character... Qc ... is equal to c O Output the resulting list, joined on empty string by default  Two notes: • The space is necessary; otherwise, the sequence AZO would scan as A ZO instead of AZ O; • The program doesn't output a trailing newline. To add one, put an x at the end of the code (thereby printing an empty string after the loop is done). Sample run (using the x variant): dlosc@dlosc:~/pip$ pip -e "FcAZ OcQUC_FIax" "HelLo wOrld"
deHlLloOrw


# APL (Dyalog Extended), 8 bytes

s⊇⍨⍋⌈s←⍞


Try it online!

(If a function were OK, it would have been 5 bytes ⍋⍤⌈⊇⊢.)

⍋ can be controversial here, because the challenge says "built-in sorting functions not allowed" and ⍋, called "Grade", is definitely related to sorting but it doesn't sort the input by itself. Instead, it gives a list of indices which, when used to index into the original input, results in the sorted array:

      x←10 50 20 40 75
⍋x
1 3 4 2 5
x[⍋x]
10 20 40 50 75


This can be also used to implement "sort-by", if we index into a different array other than x:

      'abcde'[⍋x]
acdbe


### How it works

s⊇⍨⍋⌈s←⍞  ⍝ Full program
s←⍞  ⍝ Take a line of string input and assign to s
⌈     ⍝ Uppercase the letters
⍋      ⍝ Grade up; find the sorting order
s⊇⍨       ⍝ Index into the original string s,
⍝ essentially sorting the string in the case-insensitive order


If ⍋ is not acceptable:

# APL (Dyalog Extended), 12 bytes

∊⍞∘∩¨⎕A,¨⌊⎕A


Try it online!

Set intersection in APL is essentially "erase the items in the left argument that do not appear in the right argument", so the duplicates in the left argument is preserved.

### How it works

∊⍞∘∩¨⎕A,¨⌊⎕A  ⍝ Full program.
⎕A,¨⌊⎕A  ⍝ Pair each char in ⎕A (uppercase letters) with its lowercase
⍞∘∩¨         ⍝ Take intersection of entire input and each pair of chars
∊             ⍝ Flatten


# Perl 5-a, 31 bytes

for$i(a..z){print grep/$i/i,@F}


Try it online!

# 05AB1E, 8 bytes

œ.ΔlÇ¥dP


Explanation:

œ         # Get all permutations of the (implicit) input-string
.Δ       # Find the first permutation which is truthy for:
l      #  Convert it to lowercase
Ç     #  Convert each character to its ASCII codepoint
¥    #  Get the deltas / forward-differences of those codepoint integer
d   #  Check that each difference is >= 0
P  #  And check that all of them are truthy
# (after which the found string is output implicitly as result)