Test cases

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

Test cases

Rob Withers
I was thinking of splitting the Test Case up into multiple test cases  
that can be run independently.  There are a few tests that take a  
very long time, including the ElGamal test cases and one of the tests  
for SHA256.  I was to create the following test cases:

DSATestCase
RSATestCase
DESTestCase
HashFunctionTestCase
ElGamalTestCase
RijndaelTestCase
CipherBlockModeTestCase


Then ElGamal can take a long time.  Is it ok to remove the SHA256  
test case of  hashing a 1000000 $a string?  The function is already  
tested with the other two tests.

Does anyone have an issue with my doing this?

cheers,
Robert
_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

RE: Test cases

Ron Teitelbaum
Rob,

We need to add standard test to all cryptographic functions.  This is one of
those tests.  I would be ok with me if we have multiple test classes.  We
could have basic tests for us, and a separate set of standard tests.  It
would be ok with me to move this test to a standard test class.

Ron

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of
> Robert Withers
> Sent: Wednesday, October 04, 2006 10:51 PM
> To: Cryptography Team Development List
> Subject: [Cryptography Team] Test cases
>
> I was thinking of splitting the Test Case up into multiple test cases
> that can be run independently.  There are a few tests that take a
> very long time, including the ElGamal test cases and one of the tests
> for SHA256.  I was to create the following test cases:
>
> DSATestCase
> RSATestCase
> DESTestCase
> HashFunctionTestCase
> ElGamalTestCase
> RijndaelTestCase
> CipherBlockModeTestCase
>
>
> Then ElGamal can take a long time.  Is it ok to remove the SHA256
> test case of  hashing a 1000000 $a string?  The function is already
> tested with the other two tests.
>
> Does anyone have an issue with my doing this?
>
> cheers,
> Robert
> _______________________________________________
> Cryptography mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography


_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

Re: Test cases

Rob Withers
Ron,

It makes perfect sense to do so.   At least splitting the test cases  
up will mean we can focus on a specific set of test cases rather than  
pay the whole cost each time.  There are a set of test cases using  
Rijndael that are failing.  Any ideas on them?

cheers,
Robert

On Oct 5, 2006, at 6:13 AM, Ron Teitelbaum wrote:

> Rob,
>
> We need to add standard test to all cryptographic functions.  This  
> is one of
> those tests.  I would be ok with me if we have multiple test  
> classes.  We
> could have basic tests for us, and a separate set of standard  
> tests.  It
> would be ok with me to move this test to a standard test class.
>
> Ron
>
>> -----Original Message-----
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of
>> Robert Withers
>> Sent: Wednesday, October 04, 2006 10:51 PM
>> To: Cryptography Team Development List
>> Subject: [Cryptography Team] Test cases
>>
>> I was thinking of splitting the Test Case up into multiple test cases
>> that can be run independently.  There are a few tests that take a
>> very long time, including the ElGamal test cases and one of the tests
>> for SHA256.  I was to create the following test cases:
>>
>> DSATestCase
>> RSATestCase
>> DESTestCase
>> HashFunctionTestCase
>> ElGamalTestCase
>> RijndaelTestCase
>> CipherBlockModeTestCase
>>
>>
>> Then ElGamal can take a long time.  Is it ok to remove the SHA256
>> test case of  hashing a 1000000 $a string?  The function is already
>> tested with the other two tests.
>>
>> Does anyone have an issue with my doing this?
>>
>> cheers,
>> Robert
>> _______________________________________________
>> Cryptography mailing list
>> [hidden email]
>> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
>> cryptography
>
>
> _______________________________________________
> Cryptography mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
> cryptography

_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

Re: Re: Test cases

Lukas Renggli
> It makes perfect sense to do so.   At least splitting the test cases
> up will mean we can focus on a specific set of test cases rather than
> pay the whole cost each time.

Or use a testing framework, that allows to tag tests to belong to
different categories/groups, such as #cryptographyStandardTest and
#slowTest ... and then everybody can chose what tests to run and what
to ignore ;-)

Lukas

--
Lukas Renggli
http://www.lukas-renggli.ch
_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

RE: Re: Test cases

Ron Teitelbaum
Agreed!  Does SUnit support that?

Ron

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of
> Lukas Renggli
> Sent: Thursday, October 05, 2006 10:08 AM
> To: Cryptography Team Development List
> Subject: Re: Re: [Cryptography Team] Test cases
>
> > It makes perfect sense to do so.   At least splitting the test cases
> > up will mean we can focus on a specific set of test cases rather than
> > pay the whole cost each time.
>
> Or use a testing framework, that allows to tag tests to belong to
> different categories/groups, such as #cryptographyStandardTest and
> #slowTest ... and then everybody can chose what tests to run and what
> to ignore ;-)
>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
> _______________________________________________
> Cryptography mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography


_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

Re: Test cases

Rob Withers
In reply to this post by Rob Withers
Load the latest Cryptography-Tests

On Oct 5, 2006, at 6:40 AM, Robert Withers wrote:

> Ron,
>
> It makes perfect sense to do so.   At least splitting the test  
> cases up will mean we can focus on a specific set of test cases  
> rather than pay the whole cost each time.  There are a set of test  
> cases using Rijndael that are failing.  Any ideas on them?
>
> cheers,
> Robert
>
> On Oct 5, 2006, at 6:13 AM, Ron Teitelbaum wrote:
>
>> Rob,
>>
>> We need to add standard test to all cryptographic functions.  This  
>> is one of
>> those tests.  I would be ok with me if we have multiple test  
>> classes.  We
>> could have basic tests for us, and a separate set of standard  
>> tests.  It
>> would be ok with me to move this test to a standard test class.
>>
>> Ron
>>
>>> -----Original Message-----
>>> From: [hidden email]
>>> [mailto:[hidden email]] On  
>>> Behalf Of
>>> Robert Withers
>>> Sent: Wednesday, October 04, 2006 10:51 PM
>>> To: Cryptography Team Development List
>>> Subject: [Cryptography Team] Test cases
>>>
>>> I was thinking of splitting the Test Case up into multiple test  
>>> cases
>>> that can be run independently.  There are a few tests that take a
>>> very long time, including the ElGamal test cases and one of the  
>>> tests
>>> for SHA256.  I was to create the following test cases:
>>>
>>> DSATestCase
>>> RSATestCase
>>> DESTestCase
>>> HashFunctionTestCase
>>> ElGamalTestCase
>>> RijndaelTestCase
>>> CipherBlockModeTestCase
>>>
>>>
>>> Then ElGamal can take a long time.  Is it ok to remove the SHA256
>>> test case of  hashing a 1000000 $a string?  The function is already
>>> tested with the other two tests.
>>>
>>> Does anyone have an issue with my doing this?
>>>
>>> cheers,
>>> Robert
>>> _______________________________________________
>>> Cryptography mailing list
>>> [hidden email]
>>> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
>>> cryptography
>>
>>
>> _______________________________________________
>> Cryptography mailing list
>> [hidden email]
>> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
>> cryptography
>
> _______________________________________________
> Cryptography mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
> cryptography

_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

Re: Test cases

Rob Withers
I just split the tests apart and made some changes to speed things up.

DiffieHellman:  now only generate a 64 bit prime/generator, otherwise  
test with a pre-generated prime/generator pair
ElGamal:  now only generate a 64 bit keys, otherwise test with a  
single 512 bit pre-generated keys
SHA256:  commented out the test of a string of 1000000 $a.   When  
primitive support arrives, we can uncomment.

All tests run in under 3 seconds of so, so there is good responsive  
feedback.

cheers,
Robert

On Oct 5, 2006, at 7:22 AM, Robert Withers wrote:

> Load the latest Cryptography-Tests
>
> On Oct 5, 2006, at 6:40 AM, Robert Withers wrote:
>
>> Ron,
>>
>> It makes perfect sense to do so.   At least splitting the test  
>> cases up will mean we can focus on a specific set of test cases  
>> rather than pay the whole cost each time.  There are a set of test  
>> cases using Rijndael that are failing.  Any ideas on them?
>>
>> cheers,
>> Robert
>>
>> On Oct 5, 2006, at 6:13 AM, Ron Teitelbaum wrote:
>>
>>> Rob,
>>>
>>> We need to add standard test to all cryptographic functions.  
>>> This is one of
>>> those tests.  I would be ok with me if we have multiple test  
>>> classes.  We
>>> could have basic tests for us, and a separate set of standard  
>>> tests.  It
>>> would be ok with me to move this test to a standard test class.
>>>
>>> Ron
>>>
>>>> -----Original Message-----
>>>> From: [hidden email]
>>>> [mailto:[hidden email]] On  
>>>> Behalf Of
>>>> Robert Withers
>>>> Sent: Wednesday, October 04, 2006 10:51 PM
>>>> To: Cryptography Team Development List
>>>> Subject: [Cryptography Team] Test cases
>>>>
>>>> I was thinking of splitting the Test Case up into multiple test  
>>>> cases
>>>> that can be run independently.  There are a few tests that take a
>>>> very long time, including the ElGamal test cases and one of the  
>>>> tests
>>>> for SHA256.  I was to create the following test cases:
>>>>
>>>> DSATestCase
>>>> RSATestCase
>>>> DESTestCase
>>>> HashFunctionTestCase
>>>> ElGamalTestCase
>>>> RijndaelTestCase
>>>> CipherBlockModeTestCase
>>>>
>>>>
>>>> Then ElGamal can take a long time.  Is it ok to remove the SHA256
>>>> test case of  hashing a 1000000 $a string?  The function is already
>>>> tested with the other two tests.
>>>>
>>>> Does anyone have an issue with my doing this?
>>>>
>>>> cheers,
>>>> Robert
>>>> _______________________________________________
>>>> Cryptography mailing list
>>>> [hidden email]
>>>> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
>>>> cryptography
>>>
>>>
>>> _______________________________________________
>>> Cryptography mailing list
>>> [hidden email]
>>> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
>>> cryptography
>>
>> _______________________________________________
>> Cryptography mailing list
>> [hidden email]
>> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
>> cryptography
>
> _______________________________________________
> Cryptography mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/ 
> cryptography

_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography
Reply | Threaded
Open this post in threaded view
|

Re: RE: Re: Test cases

Lukas Renggli
In reply to this post by Ron Teitelbaum
> Agreed!  Does SUnit support that?

No I just wanted to say that we are lacking at least this feature in
SUnit. Other xUnit frameworks support categories for years ...

Lukas

--
Lukas Renggli
http://www.lukas-renggli.ch
_______________________________________________
Cryptography mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/cryptography