Cut down on usage of deprecated APIs in //crypto.

Cut down on usage of deprecated APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653
Committed: https://crrev.com/7dad2a3ec1c5eada75bdd6af49a17c85113814e8
Cr-Commit-Position: refs/heads/master@{#378610}

[davidben, 2016-02-26 21:51:34]

davidben@chromium.org changed reviewers:
+ rsleevi@chromium.org

[davidben, 2016-02-26 21:51:34]

https://codereview.chromium.org/1739403002/diff/1/crypto/auto_cbb.h
File crypto/auto_cbb.h (right):

https://codereview.chromium.org/1739403002/diff/1/crypto/auto_cbb.h#newcode16
crypto/auto_cbb.h:16: class AutoCBB {
I still very strongly think this is a bad name and it should be ScopedCBB, but
I'm tired of arguing over this.

https://codereview.chromium.org/1739403002/diff/1/crypto/ec_private_key_nss.cc
File crypto/ec_private_key_nss.cc (right):

https://codereview.chromium.org/1739403002/diff/1/crypto/ec_private_key_nss.c...
crypto/ec_private_key_nss.cc:313: bool
ECPrivateKey::ExportValueForTesting(std::vector<uint8_t>* output) {
(See comment on the openssl side.)

https://codereview.chromium.org/1739403002/diff/1/crypto/ec_private_key_opens...
File crypto/ec_private_key_openssl.cc (left):

https://codereview.chromium.org/1739403002/diff/1/crypto/ec_private_key_opens...
crypto/ec_private_key_openssl.cc:227: bool
ECPrivateKey::ExportECParams(std::vector<uint8_t>* output) {
I didn't want to bother with this since I anticipate changing the API again
later to decouple from the giant OID table. It also is only used by tests,
documented as such, and the ExportValue version is an ECPrivateKey which
includes the parameters.

On the NSS side, it's a different matter. ExportValue gives back the raw private
key, which is why you needed to split in two. I've just done a dumb
concatenation (ExportValue is always going to be the same size) because, again,
this is test-only.

It didn't seem worth the trouble to do anything more interesting and this seems
less ridiculous than having ExportECParams return the empty string.

[davidben, 2016-02-26 23:08:22]

Description was changed from

==========
Cut down on usage of deprecate APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653
==========

to

==========
Cut down on usage of deprecated APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653
==========

[Ryan Sleevi, 2016-02-26 23:16:08]

lgtm

https://codereview.chromium.org/1739403002/diff/20001/crypto/auto_cbb.h
File crypto/auto_cbb.h (right):

https://codereview.chromium.org/1739403002/diff/20001/crypto/auto_cbb.h#newco...
crypto/auto_cbb.h:29: CBB cbb_;
Grumble grumble on over-abbreviated type names but that's OpenSSL

https://codereview.chromium.org/1739403002/diff/20001/crypto/ec_private_key_o...
File crypto/ec_private_key_openssl.cc (right):

https://codereview.chromium.org/1739403002/diff/20001/crypto/ec_private_key_o...
crypto/ec_private_key_openssl.cc:206: OpenSSLErrStackTracer
err_tracer(FROM_HERE);
Did we also need this for line 186?

[davidben, 2016-02-26 23:45:35]

https://codereview.chromium.org/1739403002/diff/20001/crypto/auto_cbb.h
File crypto/auto_cbb.h (right):

https://codereview.chromium.org/1739403002/diff/20001/crypto/auto_cbb.h#newco...
crypto/auto_cbb.h:29: CBB cbb_;
On 2016/02/26 23:16:08, Ryan Sleevi wrote:
> Grumble grumble on over-abbreviated type names but that's OpenSSL

Actually, this type's completely our fault. :-) I dunno, if I had to write long
chains of CRYPTO_BYTE_BUILDER_add_asn1, I'd probably go crazy. I think it's a C
thing. In C++, it's just be builder.AddAsn1(...) and so it's okay for things to
be long.

https://codereview.chromium.org/1739403002/diff/20001/crypto/ec_private_key_o...
File crypto/ec_private_key_openssl.cc (right):

https://codereview.chromium.org/1739403002/diff/20001/crypto/ec_private_key_o...
crypto/ec_private_key_openssl.cc:206: OpenSSLErrStackTracer
err_tracer(FROM_HERE);
On 2016/02/26 23:16:08, Ryan Sleevi wrote:
> Did we also need this for line 186?

Done. (One of these days, we'll figure out how to make that thing
unnecessary...)

[davidben, 2016-03-01 21:39:43]

Apparently I had a typo that was only caught by higher-level tests. In penance,
I've now added some more tests to ec_private_key_unittest.cc that would have
noticed this. PTAL.

[Ryan Sleevi, 2016-03-01 22:05:27]

LGTM % bug

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
File crypto/ec_private_key_unittest.cc (right):

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
crypto/ec_private_key_unittest.cc:189: static const char kOpenSSLRawPublicKey[]
= {
uint8_t

[davidben, 2016-03-01 22:10:08]

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
File crypto/ec_private_key_unittest.cc (right):

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
crypto/ec_private_key_unittest.cc:189: static const char kOpenSSLRawPublicKey[]
= {
On 2016/03/01 22:05:27, Ryan Sleevi wrote:
> uint8_t

I think this actually does want to be the signed one. ExportRawPrivateKey gives
back a std::string (whyyyyy....).

[davidben, 2016-03-01 22:10:48]

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
File crypto/ec_private_key_unittest.cc (right):

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
crypto/ec_private_key_unittest.cc:189: static const char kOpenSSLRawPublicKey[]
= {
On 2016/03/01 22:10:08, davidben wrote:
> On 2016/03/01 22:05:27, Ryan Sleevi wrote:
> > uint8_t
> 
> I think this actually does want to be the signed one. ExportRawPrivateKey
gives
> back a std::string (whyyyyy....).

Oh, but MSVC hates it. Fine. I'll add the annoying casts... :-)

[Ryan Sleevi, 2016-03-01 22:10:59]

On 2016/03/01 22:10:08, davidben wrote:
>
https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
> File crypto/ec_private_key_unittest.cc (right):
> 
>
https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
> crypto/ec_private_key_unittest.cc:189: static const char
kOpenSSLRawPublicKey[]
> = {
> On 2016/03/01 22:05:27, Ryan Sleevi wrote:
> > uint8_t
> 
> I think this actually does want to be the signed one. ExportRawPrivateKey
gives
> back a std::string (whyyyyy....).

You can't use the constants like you are, then. That's why the win bots are
yelling - truncating the char constants.

[davidben, 2016-03-01 22:18:43]

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
File crypto/ec_private_key_unittest.cc (right):

https://codereview.chromium.org/1739403002/diff/60001/crypto/ec_private_key_u...
crypto/ec_private_key_unittest.cc:189: static const char kOpenSSLRawPublicKey[]
= {
On 2016/03/01 22:10:47, davidben wrote:
> On 2016/03/01 22:10:08, davidben wrote:
> > On 2016/03/01 22:05:27, Ryan Sleevi wrote:
> > > uint8_t
> > 
> > I think this actually does want to be the signed one. ExportRawPrivateKey
> gives
> > back a std::string (whyyyyy....).
> 
> Oh, but MSVC hates it. Fine. I'll add the annoying casts... :-)

Done.

[davidben, 2016-03-01 22:18:59]

The CQ bit was checked by davidben@chromium.org

[davidben, 2016-03-01 22:19:00]

The patchset sent to the CQ was uploaded after l-g-t-m from rsleevi@chromium.org
Link to the patchset: https://codereview.chromium.org/1739403002/#ps80001
(title: "grumble grumble string vector char uint8_t grumble")

[commit-bot: I haz the power, 2016-03-01 22:20:08]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1739403002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1739403002/80001

[commit-bot: I haz the power, 2016-03-01 23:47:56]

Description was changed from

==========
Cut down on usage of deprecated APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653
==========

to

==========
Cut down on usage of deprecated APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653
==========

[commit-bot: I haz the power, 2016-03-01 23:47:57]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-03-01 23:48:46]

Description was changed from

==========
Cut down on usage of deprecated APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653
==========

to

==========
Cut down on usage of deprecated APIs in //crypto.

SSL_library_init is deprecated. It's CRYPTO_library_init. Switch from the
legacy ASN.1 APIs to the new parsers where feasible.
ECPrivateKey::CreateFromEncryptedPrivateKeyInfo is left alone for now as we
still need a new version of those APIs.

This also adds a scoper for CBB for use in later CLs.

BUG=499653

Committed: https://crrev.com/7dad2a3ec1c5eada75bdd6af49a17c85113814e8
Cr-Commit-Position: refs/heads/master@{#378610}
==========

[commit-bot: I haz the power, 2016-03-01 23:48:47]

Patchset 5 (id:??) landed as
https://crrev.com/7dad2a3ec1c5eada75bdd6af49a17c85113814e8
Cr-Commit-Position: refs/heads/master@{#378610}