RE: Crypt Producting Consistant Results with NE Strings

------=_NextPart_000_0000_01C51E82.87001B00
Content-Type: text/plain;
charset="us-ascii"
Content-Transfer-Encoding: 7bit

I accidently sent from my boss's email address ... however I found that
CRYPT seems to be only encoding the first 8 characters of any string. It's
too bad I didn't notice this earlier, is there anyway to work around it?

My application would work best (it's a menu with order history) if each
string sent had a unique counterpart. However, the menu is divided into
categories and some categories have the same item. Example: Pizza,Sicilian;
Pizza: Slices,Sicilian

The string I am encrypted is setup "Category,Item,Extras,Amount;etc..." and
represents one order for each item.

Thanks,
Andrew

Orig. Message
----------------------------------

Part of my application uses a crypt on a string to store it in a database,
but I started to notice non-unique hashes coming from similar strings. Here
are two strings that I tested and their resulting hash:

<MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
<MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">

RDNt2P.DhAVOA RDNt2P.DhAVOA

Is there a problem with using spaces or semicolons? I don't want to change
the salt or anything since I want to do lookups based on hashes.

Thanks
Andrew

------=_NextPart_000_0000_01C51E82.87001B00--

RE: Crypt Producting Consistant Results with NE Strings



Yes, CRYPT by definition uses only the first 8 characters for the hash. You
can use the built-in crypto_md5() function (compiled Miva Script). If still
working with uncompiled script, there is an MD5 function available on my
website - in the resource section.

Ivo
http://mivo.truxoft.com


-----Original Message-----
From: Andrew Rosborough

I accidently sent from my boss's email address ... however I found that
CRYPT seems to be only encoding the first 8 characters of any string. It's
too bad I didn't notice this earlier, is there anyway to work around it?

My application would work best (it's a menu with order history) if each
string sent had a unique counterpart. However, the menu is divided into
categories and some categories have the same item. Example: Pizza,Sicilian;
Pizza: Slices,Sicilian

The string I am encrypted is setup "Category,Item,Extras,Amount;etc..." and
represents one order for each item.

Thanks,
Andrew

Orig. Message
----------------------------------

Part of my application uses a crypt on a string to store it in a database,
but I started to notice non-unique hashes coming from similar strings. Here
are two strings that I tested and their resulting hash:

<MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
<MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">

RDNt2P.DhAVOA RDNt2P.DhAVOA

Is there a problem with using spaces or semicolons? I don't want to change
the salt or anything since I want to do lookups based on hashes.

Thanks
Andre

RE: Crypt Producting Consistant Results with NE Strings



Does your md5 function have any limit to the number of characters that
are used? I guess the string for an order could get really lengthy. If
there is still a limit to how many characters are used, I would probably
need to take a few characters spaced evenly apart from the string to
make a new hashable string but this would be overly complicated.

Thanks, I'll try setting up your script now and see how it works.
Andrew

On Wed, 2005-03-02 at 00:08 +0100, Ivo Truxa wrote:
> Yes, CRYPT by definition uses only the first 8 characters for the hash. You
> can use the built-in crypto_md5() function (compiled Miva Script). If still
> working with uncompiled script, there is an MD5 function available on my
> website - in the resource section.
>
> Ivo
> http://mivo.truxoft.com
>
>
> -----Original Message-----
> From: Andrew Rosborough
>
> I accidently sent from my boss's email address ... however I found that
> CRYPT seems to be only encoding the first 8 characters of any string. It's
> too bad I didn't notice this earlier, is there anyway to work around it?
>
> My application would work best (it's a menu with order history) if each
> string sent had a unique counterpart. However, the menu is divided into
> categories and some categories have the same item. Example: Pizza,Sicilian;
> Pizza: Slices,Sicilian
>
> The string I am encrypted is setup "Category,Item,Extras,Amount;etc..." and
> represents one order for each item.
>
> Thanks,
> Andrew
>
> Orig. Message
> ----------------------------------
>
> Part of my application uses a crypt on a string to store it in a database,
> but I started to notice non-unique hashes coming from similar strings. Here
> are two strings that I tested and their resulting hash:
>
> <MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
> <MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">
>
> RDNt2P.DhAVOA RDNt2P.DhAVOA
>
> Is there a problem with using spaces or semicolons? I don't want to change
> the salt or anything since I want to do lookups based on hashes.
>
> Thanks
> Andre
>
>

RE: Crypt Producting Consistant Results with NE Strings



I answered my last question by looking on your website where it says
that the md5 function handles any data length. However when I load
md5.mv, after a few uses of the test form it returns ffffffffffffff and
sometimes 0000000000000000. I'll keep looking and try to find out why.

Andrew

On Tue, 2005-03-01 at 19:45 -0500, Andrew Rosborough wrote:
> Does your md5 function have any limit to the number of characters that
> are used? I guess the string for an order could get really lengthy. If
> there is still a limit to how many characters are used, I would probably
> need to take a few characters spaced evenly apart from the string to
> make a new hashable string but this would be overly complicated.
>
> Thanks, I'll try setting up your script now and see how it works.
> Andrew
>
> On Wed, 2005-03-02 at 00:08 +0100, Ivo Truxa wrote:
> > Yes, CRYPT by definition uses only the first 8 characters for the hash. You
> > can use the built-in crypto_md5() function (compiled Miva Script). If still
> > working with uncompiled script, there is an MD5 function available on my
> > website - in the resource section.
> >
> > Ivo
> > http://mivo.truxoft.com
> >
> >
> > -----Original Message-----
> > From: Andrew Rosborough
> >
> > I accidently sent from my boss's email address ... however I found that
> > CRYPT seems to be only encoding the first 8 characters of any string. It's
> > too bad I didn't notice this earlier, is there anyway to work around it?
> >
> > My application would work best (it's a menu with order history) if each
> > string sent had a unique counterpart. However, the menu is divided into
> > categories and some categories have the same item. Example: Pizza,Sicilian;
> > Pizza: Slices,Sicilian
> >
> > The string I am encrypted is setup "Category,Item,Extras,Amount;etc..." and
> > represents one order for each item.
> >
> > Thanks,
> > Andrew
> >
> > Orig. Message
> > ----------------------------------
> >
> > Part of my application uses a crypt on a string to store it in a database,
> > but I started to notice non-unique hashes coming from similar strings. Here
> > are two strings that I tested and their resulting hash:
> >
> > <MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
> > <MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">
> >
> > RDNt2P.DhAVOA RDNt2P.DhAVOA
> >
> > Is there a problem with using spaces or semicolons? I don't want to change
> > the salt or anything since I want to do lookups based on hashes.
> >
> > Thanks
> > Andre
> >
> >

RE: Crypt Producting Consistant Results with NE Strings



No, MD5 has no designed limits - it is commonly used for making hashes over
entire files or directories. Just keep on mind that this Miva script
function is interpreted and Miva Script is not well suite for binary and
hexadecimal operations, hence it is rather slow for use on files. Tens or
hundreds of characters should not be of a problem, though.

Ivo
http://mivo.truxoft.com


-----Original Message-----
From: Andrew Rosborough [mailto:[email protected]]
Sent: Wednesday, March 02, 2005 1:45 AM
To: Ivo Truxa
Cc: [email protected]
Subject: RE: [meu] RE: Crypt Producting Consistant Results with NE Strings

Does your md5 function have any limit to the number of characters that
are used? I guess the string for an order could get really lengthy. If
there is still a limit to how many characters are used, I would probably
need to take a few characters spaced evenly apart from the string to
make a new hashable string but this would be overly complicated.

Thanks, I'll try setting up your script now and see how it works.
Andrew

On Wed, 2005-03-02 at 00:08 +0100, Ivo Truxa wrote:
> Yes, CRYPT by definition uses only the first 8 characters for the hash.
You
> can use the built-in crypto_md5() function (compiled Miva Script). If
still
> working with uncompiled script, there is an MD5 function available on my
> website - in the resource section.
>
> Ivo
> http://mivo.truxoft.com
>
>
> -----Original Message-----
> From: Andrew Rosborough
>
> I accidently sent from my boss's email address ... however I found that
> CRYPT seems to be only encoding the first 8 characters of any string.
It's
> too bad I didn't notice this earlier, is there anyway to work around it?
>
> My application would work best (it's a menu with order history) if each
> string sent had a unique counterpart. However, the menu is divided into
> categories and some categories have the same item. Example:
Pizza,Sicilian;
> Pizza: Slices,Sicilian
>
> The string I am encrypted is setup "Category,Item,Extras,Amount;etc..."
and
> represents one order for each item.
>
> Thanks,
> Andrew
>
> Orig. Message
> ----------------------------------
>
> Part of my application uses a crypt on a string to store it in a database,
> but I started to notice non-unique hashes coming from similar strings.
Here
> are two strings that I tested and their resulting hash:
>
> <MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
> <MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">
>
> RDNt2P.DhAVOA RDNt2P.DhAVOA
>
> Is there a problem with using spaces or semicolons? I don't want to
change
> the salt or anything since I want to do lookups based on hashes.
>
> Thanks
> Andre

RE: Crypt Producting Consistant Results with NE Strings



Thanks for helping me with this. It turns out it won't be hard to
change the crypt function and make everything work again. The hardest
part will be correctly calling the md5 script with MvDO.

I tried <MvDO FILE="/miva/md5.mv" NAME="l.checksum" VALUE="{calcMD5('''
$ l.completeorder $ ''')}"> but it is returning the same string each
time.

When I call the function, it rarely returns an actual hash. That's why
I added apostrophe's. The script is running on earthlink servers Miva
3.9408

Andrew



On Tue, 2005-03-01 at 19:57 -0500, Andrew Rosborough wrote:
> I answered my last question by looking on your website where it says
> that the md5 function handles any data length. However when I load
> md5.mv, after a few uses of the test form it returns ffffffffffffff and
> sometimes 0000000000000000. I'll keep looking and try to find out why.
>
> Andrew
>
> On Tue, 2005-03-01 at 19:45 -0500, Andrew Rosborough wrote:
> > Does your md5 function have any limit to the number of characters that
> > are used? I guess the string for an order could get really lengthy. If
> > there is still a limit to how many characters are used, I would probably
> > need to take a few characters spaced evenly apart from the string to
> > make a new hashable string but this would be overly complicated.
> >
> > Thanks, I'll try setting up your script now and see how it works.
> > Andrew
> >
> > On Wed, 2005-03-02 at 00:08 +0100, Ivo Truxa wrote:
> > > Yes, CRYPT by definition uses only the first 8 characters for the hash. You
> > > can use the built-in crypto_md5() function (compiled Miva Script). If still
> > > working with uncompiled script, there is an MD5 function available on my
> > > website - in the resource section.
> > >
> > > Ivo
> > > http://mivo.truxoft.com
> > >
> > >
> > > -----Original Message-----
> > > From: Andrew Rosborough
> > >
> > > I accidently sent from my boss's email address ... however I found that
> > > CRYPT seems to be only encoding the first 8 characters of any string. It's
> > > too bad I didn't notice this earlier, is there anyway to work around it?
> > >
> > > My application would work best (it's a menu with order history) if each
> > > string sent had a unique counterpart. However, the menu is divided into
> > > categories and some categories have the same item. Example: Pizza,Sicilian;
> > > Pizza: Slices,Sicilian
> > >
> > > The string I am encrypted is setup "Category,Item,Extras,Amount;etc..." and
> > > represents one order for each item.
> > >
> > > Thanks,
> > > Andrew
> > >
> > > Orig. Message
> > > ----------------------------------
> > >
> > > Part of my application uses a crypt on a string to store it in a database,
> > > but I started to notice non-unique hashes coming from similar strings. Here
> > > are two strings that I tested and their resulting hash:
> > >
> > > <MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
> > > <MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">
> > >
> > > RDNt2P.DhAVOA RDNt2P.DhAVOA
> > >
> > > Is there a problem with using spaces or semicolons? I don't want to change
> > > the salt or anything since I want to do lookups based on hashes.
> > >
> > > Thanks
> > > Andre
> > >
> > >

RE: Crypt Producting Consistant Results with NE Strings



Make sure you copied the plain-text version of the script as adviced at the
top of the page, not the syntax-coloured one - that usually slightly differs
due to the imperfection of the MvCALL parser. What strings give you result
ffffffffffffff and 0000000000000?

Ivo Truxa

| http://miva.truxoft.com
| Advanced Miva Merchant modules



-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf
Of Andrew Rosborough
Sent: Wednesday, March 02, 2005 2:57 AM
To: [email protected]
Cc: [email protected]
Subject: RE: [meu] RE: Crypt Producting Consistant Results with NE Strings

Thanks for helping me with this. It turns out it won't be hard to
change the crypt function and make everything work again. The hardest
part will be correctly calling the md5 script with MvDO.

I tried <MvDO FILE="/miva/md5.mv" NAME="l.checksum" VALUE="{calcMD5('''
$ l.completeorder $ ''')}"> but it is returning the same string each
time.

When I call the function, it rarely returns an actual hash. That's why
I added apostrophe's. The script is running on earthlink servers Miva
3.9408

Andrew



On Tue, 2005-03-01 at 19:57 -0500, Andrew Rosborough wrote:
> I answered my last question by looking on your website where it says
> that the md5 function handles any data length. However when I load
> md5.mv, after a few uses of the test form it returns ffffffffffffff and
> sometimes 0000000000000000. I'll keep looking and try to find out why.
>
> Andrew
>
> On Tue, 2005-03-01 at 19:45 -0500, Andrew Rosborough wrote:
> > Does your md5 function have any limit to the number of characters that
> > are used? I guess the string for an order could get really lengthy. If
> > there is still a limit to how many characters are used, I would probably
> > need to take a few characters spaced evenly apart from the string to
> > make a new hashable string but this would be overly complicated.
> >
> > Thanks, I'll try setting up your script now and see how it works.
> > Andrew
> >
> > On Wed, 2005-03-02 at 00:08 +0100, Ivo Truxa wrote:
> > > Yes, CRYPT by definition uses only the first 8 characters for the
hash. You
> > > can use the built-in crypto_md5() function (compiled Miva Script). If
still
> > > working with uncompiled script, there is an MD5 function available on
my
> > > website - in the resource section.
> > >
> > > Ivo
> > > http://mivo.truxoft.com
> > >
> > >
> > > -----Original Message-----
> > > From: Andrew Rosborough
> > >
> > > I accidently sent from my boss's email address ... however I found
that
> > > CRYPT seems to be only encoding the first 8 characters of any string.
It's
> > > too bad I didn't notice this earlier, is there anyway to work around
it?
> > >
> > > My application would work best (it's a menu with order history) if
each
> > > string sent had a unique counterpart. However, the menu is divided
into
> > > categories and some categories have the same item. Example:
Pizza,Sicilian;
> > > Pizza: Slices,Sicilian
> > >
> > > The string I am encrypted is setup
"Category,Item,Extras,Amount;etc..." and
> > > represents one order for each item.
> > >
> > > Thanks,
> > > Andrew
> > >
> > > Orig. Message
> > > ----------------------------------
> > >
> > > Part of my application uses a crypt on a string to store it in a
database,
> > > but I started to notice non-unique hashes coming from similar strings.
Here
> > > are two strings that I tested and their resulting hash:
> > >
> > > <MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
> > > <MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">
> > >
> > > RDNt2P.DhAVOA RDNt2P.DhAVOA
> > >
> > > Is there a problem with using spaces or semicolons? I don't want to
change
> > > the salt or anything since I want to do lookups based on hashes.
> > >
> > > Thanks
> > > Andre

RE: Crypt Producting Consistant Results with NE Strings



> Thanks for helping me with this. It turns out it won't be hard to
> change the crypt function and make everything work again. The hardest
> part will be correctly calling the md5 script with MvDO.
>
> I tried <MvDO FILE="/miva/md5.mv" NAME="l.checksum" VALUE="{calcMD5('''
> $ l.completeorder $ ''')}"> but it is returning the same string each
> time.

Use <MvDO FILE="/miva/md5.mv" NAME="l.checksum"
VALUE="{calcMD5(l.completeorder)}"> so that you aren't simply hashing the
string constant 'l.completeorder'. I've been using Ivo's function for a
couple of years without problems.

> When I call the function, it rarely returns an actual hash. That's why
> I added apostrophe's. The script is running on earthlink servers Miva
> 3.9408
>
> Andrew
>
>
>
> On Tue, 2005-03-01 at 19:57 -0500, Andrew Rosborough wrote:
>> I answered my last question by looking on your website where it says
>> that the md5 function handles any data length. However when I load
>> md5.mv, after a few uses of the test form it returns ffffffffffffff and
>> sometimes 0000000000000000. I'll keep looking and try to find out why.
>>
>> Andrew
>>
>> On Tue, 2005-03-01 at 19:45 -0500, Andrew Rosborough wrote:
>> > Does your md5 function have any limit to the number of characters that
>> > are used? I guess the string for an order could get really lengthy.
>> If
>> > there is still a limit to how many characters are used, I would
>> probably
>> > need to take a few characters spaced evenly apart from the string to
>> > make a new hashable string but this would be overly complicated.
>> >
>> > Thanks, I'll try setting up your script now and see how it works.
>> > Andrew
>> >
>> > On Wed, 2005-03-02 at 00:08 +0100, Ivo Truxa wrote:
>> > > Yes, CRYPT by definition uses only the first 8 characters for the
>> hash. You
>> > > can use the built-in crypto_md5() function (compiled Miva Script).
>> If still
>> > > working with uncompiled script, there is an MD5 function available
>> on my
>> > > website - in the resource section.
>> > >
>> > > Ivo
>> > > http://mivo.truxoft.com
>> > >
>> > >
>> > > -----Original Message-----
>> > > From: Andrew Rosborough
>> > >
>> > > I accidently sent from my boss's email address ... however I found
>> that
>> > > CRYPT seems to be only encoding the first 8 characters of any
>> string. It's
>> > > too bad I didn't notice this earlier, is there anyway to work around
>> it?
>> > >
>> > > My application would work best (it's a menu with order history) if
>> each
>> > > string sent had a unique counterpart. However, the menu is divided
>> into
>> > > categories and some categories have the same item. Example:
>> Pizza,Sicilian;
>> > > Pizza: Slices,Sicilian
>> > >
>> > > The string I am encrypted is setup
>> "Category,Item,Extras,Amount;etc..." and
>> > > represents one order for each item.
>> > >
>> > > Thanks,
>> > > Andrew
>> > >
>> > > Orig. Message
>> > > ----------------------------------
>> > >
>> > > Part of my application uses a crypt on a string to store it in a
>> database,
>> > > but I started to notice non-unique hashes coming from similar
>> strings. Here
>> > > are two strings that I tested and their resulting hash:
>> > >
>> > > <MvEVAL EXPR = "{'Pizza,Large Plain,,1;' CRYPT 'RD'}">
>> > > <MvEVAL EXPR = "{'Pizza,Large White,,1;' CRYPT 'RD'}">
>> > >
>> > > RDNt2P.DhAVOA RDNt2P.DhAVOA
>> > >
>> > > Is there a problem with using spaces or semicolons? I don't want to
>> change
>> > > the salt or anything since I want to do lookups based on hashes.
>> > >
>> > > Thanks
>> > > Andre
>> > >
>> > >