Age | Commit message (Collapse) | Author |
|
This method is untested, and calls the same method in RegistryImpl -- which doesn't exist. Hence, if there would be any user of this method their compilation would already fail, and we can remove the method.
PiperOrigin-RevId: 525074582
|
|
PiperOrigin-RevId: 514689012
|
|
PiperOrigin-RevId: 514652127
|
|
PiperOrigin-RevId: 513704158
|
|
PiperOrigin-RevId: 474238613
|
|
PiperOrigin-RevId: 365011843
|
|
As it turns out, in C++ it isn't really sufficient to put things into a directory internal, because users who include the normal registry will also get the internal directory included transitively automatically. Hence, in order to make sure that users notice when they include an internal target, we should always mark things with a namespace internal.
PiperOrigin-RevId: 363643315
|
|
PiperOrigin-RevId: 363400889
|
|
PiperOrigin-RevId: 294175877
|
|
get_key_manager() remain valid, even in the presence of concurrent registering calls.
This was previously already the case, except for calls to RegisterAsymmetricKeyManagers after a non-asymmetric register call. Hence, there's a bit of a change of behavior, but if someone does this already I think they will prefer the new failure to the previous potential one.
I also remove the reference_wrapper TODO, since I think it's infeasible to change this at this point (see e.g. [1]). Let me know if you want to keep it.
PiperOrigin-RevId: 279289623
|
|
PiperOrigin-RevId: 276745125
|
|
We previously used passed an owned raw pointer from registry.h to registry_impl.h, because this makes it possible to infer template parameters (which doesn't work with unique pointers). However, it is better to instead specify the parameters explicitly, which is easiest if we add typedefs to the KeyTypeManager class.
This will also be useful elsewhere (to create [1] automatically).
PiperOrigin-RevId: 262295924
|
|
PiperOrigin-RevId: 260722972
|
|
PiperOrigin-RevId: 260483519
|
|
registering a pair of an internal key manager and a corresponding private key manager.
PiperOrigin-RevId: 257372013
|
|
PiperOrigin-RevId: 253555275
|
|
PiperOrigin-RevId: 235517508
GitOrigin-RevId: d8c88fc5b3102c699ecab77fc431d854293cc032
|
|
Previously, the keyset_handle and registry would depend on each other, but in a way which could be handled. If we want to introduce the method keyset_handle.GetPrimitive<P>, doing this becomes at the very least very hard and confusing (I would have to split the header into two parts, essentially).
Instead, Bartosz pointed out that an option is to make the keyset_handle depend on the registry, and move the method GetPrimitives from the registry to the keyset_handle. I believe that this is the better option. However, it breaks clients who access GetPrimitives in the registry directly. There are no such clients within google; and I hope that there are non in the outside world either.
In any case, what needs to be done by any user is to replace all calls
Registry::GetPrimitives(keyset_handle, keymanager);
with
keyset_handle.GetPrimitives(keymanager);
PiperOrigin-RevId: 219093515
GitOrigin-RevId: efcd8ce66a94459a073b30f9818008a85904dc52
|
|
We create a class PrimitiveWrapper with a method Wrap which essentially replaces the static function [Primitive]SetWrapper::New[Primitive] for each Primitive. We then allow registering this object in the registry and add a method Wrap to the registry.
This allows us to replace the function AeadFactory::GetPrimitive with a completely templatized one, which means we will be able to remove the factories. As Thai suggested, I think later we should add a function GetPrimitive<P> to the KeysetHandle.
I added a call to register to the AeadFactory; this ensures that it works the same way as before.
PiperOrigin-RevId: 218659629
GitOrigin-RevId: 4b99c6356f158c94d7e6d1497f483ea32062940f
|
|
This is the same fix as for the registry before.
PiperOrigin-RevId: 215716532
GitOrigin-RevId: 04fb7d1b20fa73921403d034516a86983ee713d7
|
|
inferred.
Also, add tests of the new, non-deprecated functionality.
PiperOrigin-RevId: 215707924
GitOrigin-RevId: 957c628080f38bbc8a36992cf43ace903b2f402b
|
|
unique_ptr instead of raw pointers.
PiperOrigin-RevId: 215701167
GitOrigin-RevId: 7bbbbce6db88dbae15614016f229b6e282e4e50d
|
|
PiperOrigin-RevId: 215570590
GitOrigin-RevId: 7c887e67cb957d1fd7d608dd0910f3739d648c0b
|
|
This is fairly straightforward -- the only change is that the functions get_new_key_allowed and get_key_factory are internal anyhow, and hence should *not* be guarded by the mutex.
Also, there's a pre-existing clang-tidy warning because of all the global variables which are illegal. We should make the registry a class and properly "static-initialize" it.
PiperOrigin-RevId: 214780540
GitOrigin-RevId: ae98b3f64affd4ca770fb31fa2c1c4d46086228c
|
|
PiperOrigin-RevId: 209768143
GitOrigin-RevId: c28dbf80a4df7be228868273081713733653171d
|
|
PiperOrigin-RevId: 206279840
GitOrigin-RevId: d77309c992f7328cb4ef4ad193e82d8d175003a4
|
|
Registry.registerKeyManager()-methods.
PiperOrigin-RevId: 203816425
GitOrigin-RevId: c9b34b7ee5eeb761d96baa0c8976722f9d2125ba
|
|
restrictive
PiperOrigin-RevId: 203283070
GitOrigin-RevId: 058f0b9a347ec73ba0ae1ee38679c0739a8eec89
|
|
(based on cl/196889713 and cl/197018177)
PiperOrigin-RevId: 202529544
GitOrigin-RevId: f290535f5a57642982d30c43b113cdae6fd7a385
|
|
Changes:
- Renamed all references of something_cc_proto to something_portable_proto.
- Added //cc/util/protobuf_helper which creates a namespace portable_port that maps to the correct namespace.
- Replaced all deps to //net/proto2/ with //cc/util:protobuf_helper.
- Replaced all instances of the proto2:: namespace with portable_proto::
- Deleted the line 'namespace util = crypto::tink::util' because it was conflicting with //util/task/status.h (a dependency of portable_proto_lib). Also, it wasn't really used in any of the files.
Notes:
- When adding a new proto you will need to whitelist it in proto/portable_tink_filter.asciipb
- For google3 (BUILD) we should use portable_proto_lib, for opensource (BUILD.bazel) we should use cc_proto_lib.
PiperOrigin-RevId: 193084414
GitOrigin-RevId: 266a0dbb7f66549655e521350b70b10d4a872286
|
|
PiperOrigin-RevId: 189379168
GitOrigin-RevId: 2524f3b25fbaa08011b512a2730265435f48db37
|
|
Change-Id: If71775ddc160f4ba3b3e6c194b589ebd903f7ef4
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 434de034db10383329ec23ef7b872b8a0586073e
|
|
This makes sense if the StringPiece is converted to std::string
with StringPiece.ToString() which does copy the string, thus should
be avoided.
This also makes the migration to absl::string_view easier. Somehow
one cannot pass a string_view to methods that take const std::string&
(it's possible to do that with StringPiece so I guess it's because
string_view does not provide an implicit conversion to const std::string&,
but I'm not sure).
Change-Id: Ic905a78752c63ac60f25d73ba7898f723e41d9df
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: fbd1fb69494ac4b9b982804c86fcae1f2902aef7
|
|
Deleting AesCtrHmacAeadParams proto as it's not used anywhere.
Change-Id: I069b6ee749ce3522c4a5935fb7b8628921bc6f0b
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: e958abc862f4f4fce1b6d55ce1bb0dc3e12bf4dc
|
|
Change-Id: I33c2a09d2983cd5aee28a1eec886eb22e31bf9f9
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: e6763b4edc0b874dec4e7056bd124271eeca9f2d
|
|
Change-Id: I0c4af7ce40698561e1d4c98c66628bdc6319990b
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: 02289b9631aca3c5aeec907e45a97d76c825b152
|
|
Change-Id: I0b1d48d5a29e7d12df805f1cc28aab07d2d9affb
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 44205bea27caded86c765b480a47fb78b8a72eca
|
|
Change-Id: Ifc49082cfcbcf04335a69e047cf6e7589463a18a
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: aea52716172a1da2cb250f9afe81a71bfc5cd353
|
|
Change-Id: I4b17e65859db914458a32d3359f1a44b5b1b6bde
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: b67033740fe4d4f49410c397aeab503da6365950
|
|
Change-Id: Ic497392503b9e6ffb129b039895fc5a2fc5a02ac
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 19648fdb4d66c4d93db5e85125cacc87ea3b9485
|
|
Change-Id: I2e628298c31e26da0067196634b5dc4d3bd5ed84
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: 9b88cb7f7aad68ff16094ab7ff6a3a1620f58cd5
|
|
Change-Id: I15136873c1b4a00b27169084f354ffbd9a0c9a53
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 47c154adf3389b287362fd4a2bde41acf5b53682
|
|
Change-Id: I115b64a1d4a3e4eadb330ed707c9206efd42ae31
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: 58c9c7744df883bbbe906d3928027607208a7d1a
|
|
It seems that blaze (not bazel) doesn't allow implicit conversion from StringPiece to string.
Change-Id: I76142bcb2eec0ef2754e6a9a48a3d7ea4356ad49
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: d98cb3ead7e8342d8dd3820d727025c492959e2a
|
|
Change-Id: I0a56f34273c6d4d1db45f70a77c44521a2b96e42
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 4ad25f03e09cde7eb7ab72799483f94a06425e08
|
|
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 64c290e4198e0bf305a446ee71c26b7a4d20c69d
|
|
Change-Id: I246a1d31cb02adc3def4404ff694892e76151c13
ORIGINAL_AUTHOR=Thai Duong <thaidn@google.com>
GitOrigin-RevId: 2d6ceb9bd716d1ece5a9a953b27ce98f7089685b
|
|
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 974a6e60e7fe670500d01317d724203aed87e15d
|
|
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: 8c6f92901ae6be91fdc842be9db3915953fc36a6
|
|
Change-Id: I9d4d1199c36b859b5870951c85e8f0bb6c035dd0
ORIGINAL_AUTHOR=Bartosz Przydatek <przydatek@google.com>
GitOrigin-RevId: baaa74af86d9883d826b14bda18ed313868fa092
|