Migrate to new cert parsing code

Migrate to new cert parsing code

[ekasper, 2013-11-08 18:43:56]

With this change, the migration to the new parser is essentially complete - I'll
remove old code in a follow-up CL.

All weird cert tests pass, and I've run it through the first 100k certs in the
log, in strict mode, without errors, in about the time that it took me to fetch
a cup of coffee :) I'll let it run through the rest of it before this gets
submitted.

[Eran, 2013-11-11 11:35:58]

LGTM, please review the comments first though.

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/cert.py
File src/python/ct/crypto/cert.py (right):

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/cert....
src/python/ct/crypto/cert.py:225: bc =
self._get_decoded_extension_value(oid.ID_CE_BASIC_CONSTRAINTS)
nit: clearer name

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/my_as...
File src/python/ct/crypto/my_asn1/types.py (right):

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/my_as...
src/python/ct/crypto/my_asn1/types.py:1049: def component_key(self):
Is it valid to call this method and the following one if there's more than one
key / value?

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/name.py
File src/python/ct/crypto/name.py (right):

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/name....
src/python/ct/crypto/name.py:22: if self._type in ["dNSName", "rfc822Name",
Move the names to constants - can be local to this module for now, but not doing
so now would mean client code will need to duplicate these strings and it'll be
hard cleaning it up afterwards.

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/name....
src/python/ct/crypto/name.py:32: # FIXME(ekasper): this appears wrong as it's
only considering
Nit: Create an issue for this in the issue tracker or comment in one of the
issues mentioned above.

[ekasper, 2013-11-19 11:10:53]

Thanks!

Running this through the log now, and committing when done, unless any issues
appear.

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/cert.py
File src/python/ct/crypto/cert.py (right):

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/cert....
src/python/ct/crypto/cert.py:225: bc =
self._get_decoded_extension_value(oid.ID_CE_BASIC_CONSTRAINTS)
On 2013/11/11 11:35:58, Eran wrote:
> nit: clearer name

Normally I agree but basic_constraints causes ugly line wraps, and I think this
block is short enough to be clear.

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/my_as...
File src/python/ct/crypto/my_asn1/types.py (right):

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/my_as...
src/python/ct/crypto/my_asn1/types.py:1049: def component_key(self):
On 2013/11/11 11:35:58, Eran wrote:
> Is it valid to call this method and the following one if there's more than one
> key / value?

Choice setters forbid having more than one key/value.

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/name.py
File src/python/ct/crypto/name.py (right):

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/name....
src/python/ct/crypto/name.py:22: if self._type in ["dNSName", "rfc822Name",
On 2013/11/11 11:35:58, Eran wrote:
> Move the names to constants - can be local to this module for now, but not
doing
> so now would mean client code will need to duplicate these strings and it'll
be
> hard cleaning it up afterwards.

Done.

https://codereview.appspot.com/23720043/diff/40001/src/python/ct/crypto/name....
src/python/ct/crypto/name.py:32: # FIXME(ekasper): this appears wrong as it's
only considering
On 2013/11/11 11:35:58, Eran wrote:
> Nit: Create an issue for this in the issue tracker or comment in one of the
> issues mentioned above.

Done.

[Eran, 2013-11-19 18:44:51]

Still LGTM.