| AC_MSG_ERROR([Cannot compile --with-bignums without the GMP library | It is highly recommended that you install a working GMP | library on your system as that will add features to Pike | required by many applications (such as Roxen). However, | if you know that you do not need GMP, you may re-run configure | with the option --without-bignums instead.])
Considering that Pike now needs Nettle to have at least a semi-usable make verify, maybe adding such a test to configure for Nettle too wouldn't be a bad idea?
Go ahead.
/ Martin Nilsson (saturator)
Previous text:
2004-02-04 13:07: Subject: nettle
| AC_MSG_ERROR([Cannot compile --with-bignums without the GMP library | It is highly recommended that you install a working GMP | library on your system as that will add features to Pike | required by many applications (such as Roxen). However, | if you know that you do not need GMP, you may re-run configure | with the option --without-bignums instead.])
Considering that Pike now needs Nettle to have at least a semi-usable make verify, maybe adding such a test to configure for Nettle too wouldn't be a bad idea?
/ Mirar
It's a fairly minimal src/post_modules/Nettle/configure.in.
How does make verify fail without Nettle? I thought crypto related tests should be disabled, and the rest still work fine. Is there lots of other stuff depending on randomness or md5?
/ Niels Möller (vässar rödpennan)
Previous text:
2004-02-04 13:10: Subject: nettle
If someone that actually do know how Nettle is configured and compiled doesn't volunteer, I'll do it... :)
/ Mirar
I get ~270 failed testsuite tests if I don't install Nettle. Then I get only 1 (what happened to the mktime test failures?),
/home/mirar/pike/src/modules/_Crypto/testsuite.in:601: Test 135 (shift 2) (CRNL) failed. 1: mixed a() { return 2: #pike 7.4 3: object aes_cbc = Crypto.cbc(Crypto.aes()); 4: aes_cbc->set_decrypt_key(String.hex2string("00000000000000000000000000000000")); 5: aes_cbc->set_iv(String.hex2string("00000000000000000000000000000000")); 6: string ct = String.hex2string("00000000000000000000000000000000"); 7: for(int i = 0; i < 10000; i++) { 8: ct = aes_cbc->crypt_block(ct); 9: } 10: return ct; 11: ; } 12: mixed b() { return "FACA37E0B0C85373DF706E73F7C9AF86" ; } 13:
o->a(): /home/mirar/pike/longlongint/test-install/pike/7.5.15/lib/7.4/modules/Crypto.pmod/cbc() o->b(): "FACA37E0B0C85373DF706E73F7C9AF86"
/ Mirar
Previous text:
2004-02-04 13:18: Subject: nettle
It's a fairly minimal src/post_modules/Nettle/configure.in.
How does make verify fail without Nettle? I thought crypto related tests should be disabled, and the rest still work fine. Is there lots of other stuff depending on randomness or md5?
/ Niels Möller (vässar rödpennan)
Grubba just added that one. I'm investigating that one now.
/ Martin Nilsson (saturator)
Previous text:
2004-02-04 13:21: Subject: nettle
I get ~270 failed testsuite tests if I don't install Nettle. Then I get only 1 (what happened to the mktime test failures?),
/home/mirar/pike/src/modules/_Crypto/testsuite.in:601: Test 135 (shift 2) (CRNL) failed. 1: mixed a() { return 2: #pike 7.4 3: object aes_cbc = Crypto.cbc(Crypto.aes()); 4: aes_cbc->set_decrypt_key(String.hex2string("00000000000000000000000000000000")); 5: aes_cbc->set_iv(String.hex2string("00000000000000000000000000000000")); 6: string ct = String.hex2string("00000000000000000000000000000000"); 7: for(int i = 0; i < 10000; i++) { 8: ct = aes_cbc->crypt_block(ct); 9: } 10: return ct; 11: ; } 12: mixed b() { return "FACA37E0B0C85373DF706E73F7C9AF86" ; } 13:
o->a(): /home/mirar/pike/longlongint/test-install/pike/7.5.15/lib/7.4/modules/Crypto.pmod/cbc() o->b(): "FACA37E0B0C85373DF706E73F7C9AF86"
/ Mirar
Many of the protocols use random functions and hashes. Though I think that e.g. Crypto.randomness.random_string has been a bit overused. I've replaced it with the much cheaper and totally insecure top level function random_string in many cases.
/ Martin Nilsson (saturator)
Previous text:
2004-02-04 13:18: Subject: nettle
It's a fairly minimal src/post_modules/Nettle/configure.in.
How does make verify fail without Nettle? I thought crypto related tests should be disabled, and the rest still work fine. Is there lots of other stuff depending on randomness or md5?
/ Niels Möller (vässar rödpennan)
Most protocols should use as secure random numbers as possible.
/ Henrik Grubbström (Lysator)
Previous text:
2004-02-04 13:24: Subject: nettle
Many of the protocols use random functions and hashes. Though I think that e.g. Crypto.randomness.random_string has been a bit overused. I've replaced it with the much cheaper and totally insecure top level function random_string in many cases.
/ Martin Nilsson (saturator)
But there is really no need for Tools.Shoot to use cryptografic randomness when it tries to test sscanf("%c").
/ Martin Nilsson (saturator)
Previous text:
2004-02-04 13:53: Subject: nettle
Most protocols should use as secure random numbers as possible.
/ Henrik Grubbström (Lysator)
And there isn't any need for it in for instance Bittorrent when it just needs a random string as identifyer, without any underlying security.
/ Mirar
Previous text:
2004-02-04 14:01: Subject: nettle
But there is really no need for Tools.Shoot to use cryptografic randomness when it tries to test sscanf("%c").
/ Martin Nilsson (saturator)
pike-devel@lists.lysator.liu.se
-
Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum -
Martin Nilsson (saturator) ＠ Pike (-) developers forum
-
Mirar ＠ Pike developers forum
-
Niels M�ller (v�ssar r�dpennan) ＠ Pike (-) developers forum