• 0

2 Questions about Hash Strings


Question

When I generate a hash (say SHA256 or SHA512) , the hash string is composed of a combination of a-f and 0-9.

 

  1. Is there a way to generate a hash string that is composed of a-z, A-Z and 0-9? 
  2. Is there away to control what characters are used so if I only wanted m-z, A-L, 0-9 and "-_=*^#@!()[]{}<>;:,.?" that would be a possibility?
Link to comment
https://www.neowin.net/forum/topic/1369666-2-questions-about-hash-strings/
Share on other sites

11 answers to this question

Recommended Posts

  • 0

You could base36 encode the hash output to give you a string composed of a-z0-9 (or write a simple custom cipher to map to whatever set of characters you want) but I can't think of a reason why you would want to do this?

  • 0

Understanding why it's needed is my business. But thank you.

 

Secondly ,PHP does this when creating a session. You are able to customize it's sid_bits_per_character to 6, which does a-zA-Z and 0-9; thus I assumed there is a method to specify what values you want to be included as the components of the hash.

  • 0

You assumed wrong. What PHP is doing is completely independent of the hashing method, it's simply taking the bits returned from (any) hashing method and rather than displaying them as a hexadecimal representation it's encoding them into a string using a character set of their choosing, just as I said you could do:

 


static char hexconvtab[] = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ,-";

static void bin_to_readable(unsigned char *in, size_t inlen, char *out, size_t outlen, char nbits) /* {{{ */
{
	unsigned char *p, *q;
	unsigned short w;
	int mask;
	int have;

	p = (unsigned char *)in;
	q = (unsigned char *)in + inlen;

	w = 0;
	have = 0;
	mask = (1 << nbits) - 1;

	while (outlen--) {
		if (have < nbits) {
			if (p < q) {
				w |= *p++ << have;
				have += 8;
			} else {
				/* Should never happen. Input must be large enough. */
				ZEND_ASSERT(0);
				break;
			}
		}

		/* consume nbits */
		*out++ = hexconvtab[w & mask];
		w >>= nbits;
		have -= nbits;
	}

	*out = '\0';
}

https://github.com/php/php-src/blob/master/ext/session/session.c#L269

 

 

I asked why because I hope you're not using this for security purposes, based on the fact that you had to ask this question in the first place you're more likely to end up reducing security rather than increasing it. 

Edited by ZakO
  • Like 1
  • Thanks 1
  • 0

1. No you can only have one of A-Z or a-z.  This can be done by encoding the string that your hashing algorithm to something that is hex.

2. No, this defeats the point of a hash.  I can't think of a good reason for doing this.

  • 0
  On 05/08/2018 at 11:20, Fahim S. said:

1. No you can only have one of A-Z or a-z.  This can be done by encoding the string that your hashing algorithm to something that is hex.

2. No, this defeats the point of a hash.  I can't think of a good reason for doing this.

Expand  

Thanks.

It's funny when I make inquiries, rather than answering I am offered personal opinions of understanding.  

 

While you offered some answers, you added your ego (or lack of worldly experience) in to the mix.  You do not need to know why I want something.  The comment of "I can't think of a good reason for doing this" is naive and immature.  Of course you cannot think of a good reason to do this; it's because you haven't lived my life; surely you understand that. But mostly that comment is completely a relevant.

 

In future, just answer the question and don't interject your immaturity in to your response.

 

Cheers mate.

Edited by Brian Miller
  • 0
  On 05/08/2018 at 10:51, ZakO said:

You assumed wrong. What PHP is doing is completely independent of the hashing method, it's simply taking the bits returned from (any) hashing method and rather than displaying them as a hexadecimal representation it's encoding them into a string using a character set of their choosing, just as I said you could do:

  


static char hexconvtab[] = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ,-";

static void bin_to_readable(unsigned char *in, size_t inlen, char *out, size_t outlen, char nbits) /* {{{ */
{
	unsigned char *p, *q;
	unsigned short w;
	int mask;
	int have;

	p = (unsigned char *)in;
	q = (unsigned char *)in + inlen;

	w = 0;
	have = 0;
	mask = (1 << nbits) - 1;

	while (outlen--) {
		if (have < nbits) {
			if (p < q) {
				w |= *p++ << have;
				have += 8;
			} else {
				/* Should never happen. Input must be large enough. */
				ZEND_ASSERT(0);
				break;
			}
		}

		/* consume nbits */
		*out++ = hexconvtab[w & mask];
		w >>= nbits;
		have -= nbits;
	}

	*out = '\0';
}

https://github.com/php/php-src/blob/master/ext/session/session.c#L269

 

 

I asked why because I hope you're not using this for security purposes, based on the fact that you had to ask this question in the first place you're more likely to end up reducing security rather than increasing it. 

Expand  

 

Thanks dude, that's what I thought too.  I like your idea of the Base encoding it, I may use Base56.

 

The reason I had asked is because I wanted to learn about forming such strings, and not necessarily for any foolish attempt at security.  The formation of BitCoin addresses such as "1BoNtSLRHtKNngkdx3e0bR7gb53L3TtpYt" first peeked my curiosity, then when I discovered PHP sessions can also include  a-zA-Z and 0-9 when setting it's sid_bits_per_character to 6 prompted me to enquire with learned people here.

 

  • 0
  On 06/08/2018 at 03:26, Brian Miller said:

Thanks.

It's funny when I make inquiries, rather than answering I am offered personal opinions of understanding.  

 

While you offered some answers, you added your ego (or lack of worldly experience) in to the mix.  You do not need to know why I want something.  The comment of "I can't think of a good reason for doing this" is naive and immature.  Of course you cannot think of a good reason to do this; it's because you haven't lived my life; surely you understand that. But mostly that comment is completely a relevant.

 

In future, just answer the question and don't interject your immaturity in to your response.

 

Cheers mate.

Expand  

Err... ok. 

 

As we are offering tips to one-another let me give you one: a bit of context can go a long way in getting an answer.  Developers like solving problems, and without detail of the underlying problem it is difficult to help. 

 

The comment was intended to probe for context (suggest you read about the 5 whys) so that I can try my best (within the bounds of my knowledge) to help you come to an answer quicker.  I apologise for the negative impression that you drew from it.

  • Like 2
  • 0
  On 06/08/2018 at 03:26, Brian Miller said:

Thanks.

It's funny when I make inquiries, rather than answering I am offered personal opinions of understanding.  

 

While you offered some answers, you added your ego (or lack of worldly experience) in to the mix.  You do not need to know why I want something.  The comment of "I can't think of a good reason for doing this" is naive and immature.  Of course you cannot think of a good reason to do this; it's because you haven't lived my life; surely you understand that. But mostly that comment is completely a relevant.

 

In future, just answer the question and don't interject your immaturity in to your response.

 

Cheers mate.

Expand  

 

Mate, I see your post count and reputation, but this doesn't mean you should act like an a**hole  You are not paying those people to have such expectations for their answers. From my point of view those were relevant and polite answers and doing their best to help you.

 

...and yes adding my ego is perfectly fine on public forum.

 

Have a nice day!

  • 0

Hello,

 

SHA-256 and SHA-512 output their results in hexadecimal notation, which is why you see 0-9 and a-f used in the results--those are the sixteen digits which compose hexadecimal notation.

 

Instead of having to re-write the hashing algorithms to provide your own numbering system, perhaps it would be better to use something like SSDeep, instead, which supports a larger encoding set?

 

Regards,

 

Aryeh Goretsky

 

  • 0
  On 06/08/2018 at 03:26, Brian Miller said:

You do not need to know why I want something.

Expand  

You clearly do not understand how forums work... That you got the answers you got is way more than I would ever in a million years given you..  With such a comment when asked why..

  • 0

Just encode the hash, shortest practical encoding I can find is base85.

 

Encoding: input -> SHA256/512 -> base85

Decoding: base85 -> SHA256/512 -> Find input data with hash

 

There are multiple common base encodings: base2(1), base10(2), base16(3), base32, base36, base58, base64(4), base85, base91(5), base128(6)

 

Above base encodings have a default character set of X characters that are being used for encoding, but it's possible to replace those with your own character set.

 

  1. Base encoding of binary data (0 & 1)
  2. Base encoding of a decimal number (0-9)
  3. Base encoding of hexidecimal string like a SHA512 hash (A-F0-9)
  4. Base encoding commonly used for encoding binary data to a string to embed it in websites
  5. Base encoding with most printable characters
  6. Base encoding of a byte and ascii string

 

BUT

If you actually try to encode your hash you will find the string doesn't become shorter ?

 

Original:

seahorsepip

SHA512: 

F9AA2F6D639C026E3325F31247E8253987D6EC6EEC7E93764F9F3CC25D08FABA7DF95FAF94779CACF22D72F96EEE88D46C90A8CE727944218A1DC272EDA29084

base85: 

mMA+2gdBe&hzWuXfFUKPgCZ^zmL=+Og=E&.gbQ<Fi5:k-mme$4mmf15iwSPLg!6F{gEBI+hafbTmNovbh:*a}hax(3iw-VNiyu81mLV.!h.)&.hBQ56i5<q.hBxFUk@.k2h.)rLg=c%xi6/n!lOZOQmmoA%iwrAH

 

Why doesn't it become shorter?

When you create a hash from data it returns a hash string in the hexadecimal format, also known as base16.

So when you encode the hash as a string using base85 you actually tell the base85 encoder that the input is an ascii string (base128), so that means you're encoding a base128 string to a base85 string which results in a longer instead of shorter string!

 

How to fix this?

Make sure to actually let the base85 encoder know that the input format is base16.

So to do that you can actually convert the hex string to bytes(base128) with hex2bin in php for example and use those bytes as input for the base85 encoder.

This means the encoding would be: input -> SHA256/512(base16) -> bytes(base128) -> base85

 

Original:

seahorsepip

SHA512: 

F9AA2F6D639C026E3325F31247E8253987D6EC6EEC7E93764F9F3CC25D08FABA7DF95FAF94779CACF22D72F96EEE88D46C90A8CE727944218A1DC272EDA29084

base85: 

}kM8+w1iQrgBqS9n9zlVHT$*f)0+dwpOf+^t)QqrEFElNLY@U0[?1[TzTJtgy(>QvA^p4@IxfMO)v]X}

And even shorter base91 (only 1 char shorter in this example):

?e#a_*!Og$d0Rh"Y4Qx.=8}^zpmb~^B4aGWI;`W=?}5&b`B3w0Exl`S[GYF#fG9.1,vcLH]LR%LxhzQ

 

And the encoded string is now shorter :D

 

Php libraries to do this:

hex2bin: http://php.net/hex2bin

base85: https://github.com/tuupola/base85

 

So php code to create a shorter hash:

$shorterHash = $base85->encode(hex2bin(hash($file)))

 

Update:

Seems like you wanted to create a custom base56 encoding, to do that we could manually create functions encode and decode it:

$base56_digits = '0123456789ABCDEFGHIJKLMNOPQRSTVWXYZabcdefghijklmnopqrstv';
$custom_digits = 'mnopqrstvwxyzABCDEFGHIJKL0123456789-_=*^#@!()[]{}<>;:,.?';

function encode($base16) {
    global $base56_digits, $custom_digits;

    $base56 = base_convert($base16, 16, 56);
    $custom = strtr($base56, $base56_digits, $custom_digits);

    return $custom;
}

function decode($custom) {
    global $base56_digits, $custom_digits;

    $base56 = strtr($custom, $custom_digits, $base56_digits);
    $base16 = base_convert($base56, 56, 16);

    return $base16;
}

But above doesn't work since php base_convert is limited to base36 :(

Instead you can use a magnificent 3rd party library: https://github.com/ArtBIT/base_convert

 

And then you have:

$custom_digits = 'mnopqrstvwxyzABCDEFGHIJKL0123456789-_=*^#@!()[]{}<>;:,.?';

function encode($base16) {
    global $custom_digits;

    return math\base_convert($base16, 16, $custom_digits);
}

function decode($custom) {
    global $custom_digits;

    return math\base_convert($custom, $custom_digits, 16);
}

Original:

seahorsepip

SHA512: 

F9AA2F6D639C026E3325F31247E8253987D6EC6EEC7E93764F9F3CC25D08FABA7DF95FAF94779CACF22D72F96EEE88D46C90A8CE727944218A1DC272EDA29084

Custom base56: 

n<^(q8}=_G@x0;B1]K6zD-DF*96yE-6L#_>K8vJ},vCz02m,8yB][4qA^12>.pw>2-?_m,{0L<qFCK:K,2@04)3s:

 

TL;DR

All data is encoded in a specific base, data can be represented as a shorter string by increasing it's base and can be respresented with a smaller character dictionary by decreasing it's base.

 

Oftopic:

  Quote

Stop the bickering back and forth, we're here to learn things and help each other, if someone doesn't want to share why he wants to do something then that's his right.

Though that doesn't mean that you have to be rude about it, if you don't want to share the why, let others know in a respectful manner.

Expand  

 

  • Like 2
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • Rubbish talentless people hired on basis of filling a certain quota will always lead to rubbish company decisions. I am stumped at the thought that people are still OK with DEI after the hell that the gaming industry is going through.
    • I mean, to say that Xbox lost is soul, is the exact same argument you could make about Sony. Clearly, the console space is not the as it was a decade ago, as it was a decade before that and another before that. Generational shifts and market trends clearly defined the space we're currently at. One thing is clear, gaming is not going anywhere, it only continues to grow, for better or for worse. Microsoft pivoted to a publisher focused, and now they want to dominate console streaming, which they seem to have a considerable leg up over Sony on this regard as Internet reliability and speeds keep improving, console hardware might not be a feasible model unless you follow the Nintendo route of making the hardware at a profit (considering its limitations). Also, it seems now that Microsoft pivoted the Xbox branding, that could open the door for 3rd party hardware companies make Consoles that run Windows-Xbox OS, which is something interesting to consider, which would convert Xbox into an Android like approach.
    • UK planning measures to stop illegal content from going viral by Usama Jawad Image via Pixabay A few days ago, we learned that the UK's digital regulator Ofcom has mandated adult websites like Pornhub to put robust age check mechanisms in place to ensure that age-inappropriate content like pornography is not readily available to minors. Now, the government-approved authority has set its sights on a new target. According to an announcement on its website, Ofcom is now holding a consultation period for a set of measures it is proposing to curb the spread of illegal content online. These methods are a part of Ofcom's "Year of Action" initiative, where the regulator is taking decisive practical actions to ensure the online safety of the UK's citizens. Ofcom is proposing the implementation of better recommender systems and crisis response protocols to restrict illegal content from going viral online. In addition, it has also proposed the utilization of proactive mechanisms like hash checks for illegal images to nip the problem in the bud and not even allow any such visual content to be published online. Furthermore, the regulator will be asking tech firms and social media platform owners to leverage artificial intelligence systems in order to detect content like fraud and suicide. People who regularly engage in the practice of publication of illegal content online will also be punished with new sanctions. Lastly, Ofcom wants to build upon its existing child protection codes and capabilities online. It is planning to do this by placing restrictions on interactions between minors and streamers during livestreams. It will urge website owners to block users who share child sexual abuse material (CSAM) and is also proposing the development of AI tools to detect grooming activities and CSAM. The consultation period is open until 5pm on October 20, 2025. Ofcom is soliciting feedback from "service providers", law enforcement authorities, and the public until that deadline to finalize its proposal.
    • What are you talking about? I have had 4 Samsung fold devices and they all have the crease. It is not as noticable when facing straight on but its still noticable.
  • Recent Achievements

    • Week One Done
      TIGOSS earned a badge
      Week One Done
    • First Post
      henryj earned a badge
      First Post
    • First Post
      CarolynHelen earned a badge
      First Post
    • Reacting Well
      henryj earned a badge
      Reacting Well
    • Community Regular
      Primey_ went up a rank
      Community Regular
  • Popular Contributors

    1. 1
      +primortal
      480
    2. 2
      +FloatingFatMan
      194
    3. 3
      ATLien_0
      164
    4. 4
      Xenon
      81
    5. 5
      Som
      76
  • Tell a friend

    Love Neowin? Tell a friend!