Collision using hash() about spyder HOT 7 CLOSED

Thanks, Alexander. I take your point in principle, but it is a question of what one expects of a hash – and not just for me. Ideally, a hash gives not only a very low expectation of collision, closely related to the chance of two random numbers being the same in a particular range of numbers, but also an expectation (harder to quantify) that combinations of numbers should be derived independently. What I showed was that four hashes had a very simple relationship a+b=c+d, suggesting that the hash was created in a mathematically simplistic way, and by ill luck this caused a rather hard-to-nail-down bug in my software are (experimental group-theory). The simple work around is to hash a derived string, and this is OK, just a bit less efficient in computing-heavy experimental programs. The simplicity of this work-around (even though it is by no means perfect) suggests that a better algorithm could easily have been chosen for hash(). I’m a bit sensitive about this issue, having created a new database system (for a commercial X.500 system) way back in the 80’s that used a relatively simple empirically derived hashing algorithm that eventually caused a customer problem. Fortunately, in those days, there were a number of good hashing algorithms that had grown up with public key cryptography (then a nascent technology), that I had avoided for the first design for efficiency reasons, but was then obliged to adopt because of the impact on the database. Fortunately, developing technology absorbed the performance problem. Anthony Anthony Hodson ***@***.******@***.***> Spring Lanes House, Holly Spring Lane, Bracknell, Berks RG12 2JL Tel: 01344 483347 Mob: 0771 360 7086 Web: www.xdotd.com<http://www.xdotd.com/> www.fosmw.com<http://www.fosmw.com/> From: Alexander Bessman ***@***.***> Sent: Monday, June 3, 2024 2:28 PM To: spyder-ide/spyder ***@***.***> Cc: Anthony Hodson ***@***.***>; Author ***@***.***> Subject: Re: [spyder-ide/spyder] Collision using hash() (Issue #22139) 1. This is obviously not a bug in Spyder. 2. It's probably not a bug at all. hash makes no guarantee that different objects have different hashes, though excessive collisions may be considered erroneous. 3. Use hashlib if your program cannot handle collisions. — Reply to this email directly, view it on GitHub<#22139 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BE6KYN7KWL6GBH5K472KKT3ZFRVNLAVCNFSM6AAAAABIUHUSK2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBVGIYDKNRRHA>. You are receiving this because you authored the thread.Message ID: ***@***.***>

Thank you, Carlos, happy to leave it at that, but perhaps useful for people to note the attitude taken by @bessman<https://github.com/bessman> – “don’t expect too much!!” – which I did, naively, as an ancient professional (working with real computers since 1960), expecting great things 64 years later. Best wishes Anthony Anthony Hodson ***@***.******@***.***> Spring Lanes House, Holly Spring Lane, Bracknell, Berks RG12 2JL Tel: 01344 483347 Mob: 0771 360 7086 Web: www.xdotd.com<http://www.xdotd.com/> www.fosmw.com<http://www.fosmw.com/> From: Carlos Cordoba ***@***.***> Sent: Thursday, June 6, 2024 4:25 PM To: spyder-ide/spyder ***@***.***> Cc: Anthony Hodson ***@***.***>; Mention ***@***.***> Subject: Re: [spyder-ide/spyder] Collision using hash() (Issue #22139) Hey @aehatxdotd<https://github.com/aehatxdotd>, thanks for reporting. Unfortunately, as @bessman<https://github.com/bessman> said, this is not a Spyder bug so we can't help you with it here. However, just as a quick advice, you could take a look at the hash functions from the cryptography module: https://cryptography.io/en/latest/hazmat/primitives/cryptographic-hashes/ — Reply to this email directly, view it on GitHub<#22139 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BE6KYNYHUIE7DIZRJZ5SV5LZGB5LFAVCNFSM6AAAAABIUHUSK2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNJSHAYDSNRXGM>. You are receiving this because you were mentioned.Message ID: ***@***.***>

Dear Carlos, Further to my last, @bessman <https://github.com/bessman> made a point that would be valid if collisions with pairs of hashes were rare. This is not the case. To prove this, please see the results of a little test on the permutations of the tiny symmetric group S4, where collisions occurred with no less than16 pairs, using the attached test program. If I change the program on the 10th line by changing ‘tuple’ to ‘str’, all the collisions disappear. I don’t need help, but perhaps Python does? Best wishes Anthony From: Carlos Cordoba ***@***.***> Sent: Thursday, June 6, 2024 4:25 PM To: spyder-ide/spyder ***@***.***> Cc: Anthony Hodson ***@***.***>; Mention ***@***.***> Subject: Re: [spyder-ide/spyder] Collision using hash() (Issue #22139) Hey @aehatxdotd <https://github.com/aehatxdotd> , thanks for reporting. Unfortunately, as @bessman <https://github.com/bessman> said, this is not a Spyder bug so we can't help you with it here. However, just as a quick advice, you could take a look at the hash functions from the cryptography module: https://cryptography.io/en/latest/hazmat/primitives/cryptographic-hashes/ — Reply to this email directly, view it on GitHub <#22139 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/BE6KYNYHUIE7DIZRJZ5SV5LZGB5LFAVCNFSM6AAAAABIUHUSK2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNJSHAYDSNRXGM> . You are receiving this because you were mentioned.Message ID: ***@***.***> perm-pair collision analysis in symmetric group 4 perms: [[0, 1, 2, 3], [0, 1, 3, 2], [0, 2, 1, 3], [0, 2, 3, 1], [0, 3, 1, 2], [0, 3, 2, 1], [1, 0, 2, 3], [1, 0, 3, 2], [1, 2, 0, 3], [1, 2, 3, 0], [1, 3, 0, 2], [1, 3, 2, 0], [2, 0, 1, 3], [2, 0, 3, 1], [2, 1, 0, 3], [2, 1, 3, 0], [2, 3, 0, 1], [2, 3, 1, 0], [3, 0, 1, 2], [3, 0, 2, 1], [3, 1, 0, 2], [3, 1, 2, 0], [3, 2, 0, 1], [3, 2, 1, 0]] collision found: -1273677301289743406 collision found: -1273677301289743406 collision found: -1273677301289743406 collision found: 5411105356768632435 collision found: 9957446576626508425 COLLISION: -1273677301289743406==-8185860710695822707(SYM[7])6912183409406079301(SYM[16]) COLLISION: -1273677301289743406==-5222634740135955210(SYM[8])3948957438846211804(SYM[15]) COLLISION: -1273677301289743406==1760123315908831802(SYM[9])-3033800617198575208(SYM[14]) COLLISION: -1273677301289743406==-3750713600196636906(SYM[11])2477036298906893500(SYM[12]) COLLISION: -1273677301289743406==-8185860710695822707(SYM[7])6912183409406079301(SYM[16]) COLLISION: -1273677301289743406==-5222634740135955210(SYM[8])3948957438846211804(SYM[15]) COLLISION: -1273677301289743406==1760123315908831802(SYM[9])-3033800617198575208(SYM[14]) COLLISION: -1273677301289743406==-3750713600196636906(SYM[11])2477036298906893500(SYM[12]) COLLISION: -1273677301289743406==-8185860710695822707(SYM[7])6912183409406079301(SYM[16]) COLLISION: -1273677301289743406==-5222634740135955210(SYM[8])3948957438846211804(SYM[15]) COLLISION: -1273677301289743406==1760123315908831802(SYM[9])-3033800617198575208(SYM[14]) COLLISION: -1273677301289743406==-3750713600196636906(SYM[11])2477036298906893500(SYM[12]) COLLISION: 5411105356768632435==3948957438846211804(SYM[15])1462147917922420631(SYM[22]) COLLISION: 5411105356768632435==2365842189548203311(SYM[17])3045263167220429124(SYM[19]) COLLISION: 9957446576626508425==6912183409406079301(SYM[16])3045263167220429124(SYM[19]) COLLISION: 9957446576626508425==2365842189548203311(SYM[17])7591604387078305114(SYM[18])

Dear Alexander, I wrote a little program to exhaustively detect collisions in the sum of two permutations using the hash value hash_value=hash(tuple(perm)) where perm is a list containing ( in the example I took for demonstration) such as [1,3,2,4], taken from the 24 such permutations in the very simple symmetric group S4. Collisions in the sum of the hash-value gave 8 pairs that generated them. I attach the summary. SYM[7] is (of course) [[1, 0, 3, 2]. The first three collisions were three-way. Not an impressive record for hash(), but substituting str for tuple does the trick adequately for me. My loss was the time spent trying to work out why I was not having the results that I expected, and collisions as the problem seemed at first less likely than the possibility of a bug in my software. Best wishes Anthony Anthony Hodson ***@***.******@***.***> Spring Lanes House, Holly Spring Lane, Bracknell, Berks RG12 2JL Tel: 01344 483347 Mob: 0771 360 7086 Web: www.xdotd.com<http://www.xdotd.com/> www.fosmw.com<http://www.fosmw.com/> From: Alexander Bessman ***@***.***> Sent: Monday, June 3, 2024 2:28 PM To: spyder-ide/spyder ***@***.***> Cc: Anthony Hodson ***@***.***>; Author ***@***.***> Subject: Re: [spyder-ide/spyder] Collision using hash() (Issue #22139) 1. This is obviously not a bug in Spyder. 2. It's probably not a bug at all. hash makes no guarantee that different objects have different hashes, though excessive collisions may be considered erroneous. 3. Use hashlib if your program cannot handle collisions. — Reply to this email directly, view it on GitHub<#22139 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BE6KYN7KWL6GBH5K472KKT3ZFRVNLAVCNFSM6AAAAABIUHUSK2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBVGIYDKNRRHA>. You are receiving this because you authored the thread.Message ID: ***@***.***> perm-pair collision analysis in symmetric group 4 perms: [[0, 1, 2, 3], [0, 1, 3, 2], [0, 2, 1, 3], [0, 2, 3, 1], [0, 3, 1, 2], [0, 3, 2, 1], [1, 0, 2, 3], [1, 0, 3, 2], [1, 2, 0, 3], [1, 2, 3, 0], [1, 3, 0, 2], [1, 3, 2, 0], [2, 0, 1, 3], [2, 0, 3, 1], [2, 1, 0, 3], [2, 1, 3, 0], [2, 3, 0, 1], [2, 3, 1, 0], [3, 0, 1, 2], [3, 0, 2, 1], [3, 1, 0, 2], [3, 1, 2, 0], [3, 2, 0, 1], [3, 2, 1, 0]] collision found: -1273677301289743406 collision found: -1273677301289743406 collision found: -1273677301289743406 collision found: 5411105356768632435 collision found: 9957446576626508425 COLLISION: -1273677301289743406==-8185860710695822707(SYM[7])6912183409406079301(SYM[16]) COLLISION: -1273677301289743406==-5222634740135955210(SYM[8])3948957438846211804(SYM[15]) COLLISION: -1273677301289743406==1760123315908831802(SYM[9])-3033800617198575208(SYM[14]) COLLISION: -1273677301289743406==-3750713600196636906(SYM[11])2477036298906893500(SYM[12]) COLLISION: 5411105356768632435==3948957438846211804(SYM[15])1462147917922420631(SYM[22]) COLLISION: 5411105356768632435==2365842189548203311(SYM[17])3045263167220429124(SYM[19]) COLLISION: 9957446576626508425==6912183409406079301(SYM[16])3045263167220429124(SYM[19]) COLLISION: 9957446576626508425==2365842189548203311(SYM[17])7591604387078305114(SYM[18])