Hi Rajini, Ron I've updated the KIP with the latest changes following our discussion. Please do give it a read. If you feel it is alright, I will follow up with a PR later.
Best, Stanislav On Thu, Aug 9, 2018 at 10:09 AM Rajini Sivaram <rajinisiva...@gmail.com> wrote: > Hi Ron/Stansilav, > > OK, let's just go with 2. I think it would be better to add a > OAuth-specific extensions handler OAuthBearerExtensionsValidatorCallback > that > provides OAuthBearerToken. > > To summarise, we chose option 2 out of these four options: > > 1. {OAuthBearerValidatorCallback, SaslExtensionsValidatorCallback} : We > don't want to use multiple ordered callbacks since we don't want the > context of one callback to come from another.callback, > 2. OAuthBearerExtensionsValidatorCallback(OAuthBearerToken token, > SaslExtensions ext): This allows extensions to be validated using > context from the token, we are ok with this. > 3. SaslExtensionsValidatorCallback(Map<String, Object> context, > SaslExtensions ext): This doesn't really offer any real advantage over > 2. > 4. OAuthBearerValidatorCallback(String token, SaslExtensions ext): We > don't want token validator to see extensions since these are insecure > but > token validation needs to be secure. So we prefer to use a second > callback > handler to validate extensions after securely validating token. > > > > On Wed, Aug 8, 2018 at 8:52 PM, Ron Dagostino <rndg...@gmail.com> wrote: > > > Hi Rajini. I think we are considering the following two options. Let me > > try to describe them along with their relative advantages/disadvantages. > > > > Option #1: Send two callbacks in a single array to the callback handler: > > ch.handle(new Callback[] {tokenCallback, extensionsCallback}); > > > > Option #2: Send two callbacks separately, in two separate arrays, to the > > callback handler: > > ch.handle(new Callback[] {tokenCallback}); > > ch.handle(new Callback[] {extensionsCallback}); > > > > I actually don't see any objective disadvantage with #1. If we don't get > > an exception then we know we have the information we need; if we do get > an > > exception then we can tell if the first callback was handled because > either > > its token() method or its errorStatus() method will return non-null; if > > both return null then we just send the token callback by itself and we > > don't publish any extension as negotiated properties. There is no > > possibility of partial results, and I don't think there is a performance > > penalty due to potential re-validation here, either. > > > > I see a subjective disadvantage with #1. It feels awkward to me to > > provide the token as context for extension validation via the first > > callback. > > > > Actually, it just occurred to me why it feels awkward, and I think this > is > > an objective disadvantage of this approach. It would be impossible to do > > extension validation in such a scenario without also doing token > validation > > first. We are using the first callback as a way to provide context, but > we > > are also using that first callback to request token validation. We are > > complecting two separate things -- context and a request for validation > -- > > into one thing, so this approach has an element of complexity to it. > > > > The second option has no such complexity. If we want to provide context > to > > the extension validation then we do that by adding a token to the > > extensionsCallback instance before we provide it to the callback handler. > > How we do that -- whether by Map<String, Object> or via a typed getter -- > > feels like a subjective decision, and assuming you agree with the > > complexity argument and choose option #2, I would defer to your > preference > > as to how to implement it. > > > > Ron > > > > > > > > > > > > > > > > > > > > > > > > > > On Wed, Aug 8, 2018 at 3:10 PM Rajini Sivaram <rajinisiva...@gmail.com> > > wrote: > > > > > Hi Ron, > > > > > > Yes, I was thinking of a SaslExtensionsValidatorCallback with > additional > > > context as well initially, but I didn't like the idea of name-value > pairs > > > and I didn't want generic objects passed around through the callback > So > > > providing context through other callbacks felt like a neater fit. There > > > are pros and cons for both approaches, so we could go with either. > > > > > > Callbacks are provided to callback handlers in an array and there is > > > implicit ordering in the callbacks provided to the callback handler. > > > In the typical example of {NameCallback, PasswordCallback}, we expect > > that > > > ordering so that password callback knows what the user name is. Kafka > > > guarantees ordering of server callbacks in each of its SASL mechanisms > > and > > > this is explicitly stated in > > > > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP- > > 86%3A+Configurable+SASL+callback+handlers > > > . > > > Until now, we didn't need to worry about ordering for OAuth. > > > > > > We currently do not have any optional callbacks - configured callback > > > handlers have to process all the callbacks for the mechanism or else we > > > fail authentication. We have to make SaslExtensionValidationCallback an > > > exception, at least for backward compatibility. We will only include > this > > > callback if the client provided some extensions. I think it is > reasonable > > > to expect that in general, custom callback handlers will handle this > > > callback if clients are likely to set extensions. In case it doesn't, > we > > > don't want to make any assumptions about which callbacks may have been > > > handled. Instead, it would be better to invoke the callback handler > again > > > without the extensions callback and not expose any extensions as > > negotiated > > > properties. Since we are doing this only for backward compatibility, > the > > > small performance hit would be reasonable, avoiding any assumptions > about > > > the callback handler implementation and partial results on hitting an > > > exception. > > > > > > Back to the other approach of providing a Map. For OAuth, we would like > > > extension validation to see the actual OAuthBearerToken object, for > > > instance to validate extensions based on scope. Having > > > these mechanism-specific objects in a Map doesn't feel ideal. It will > > > probably be better to define OAuthBearerExtensionsValidatorCallback > > with a > > > token getter that returns OAuthBearerToken in that case. > > > > > > Thoughts? > > > > > > > > > On Wed, Aug 8, 2018 at 6:09 PM, Ron Dagostino <rndg...@gmail.com> > wrote: > > > > > > > Hi Rajini. I also like that idea, but I think it might rely on one > or > > > > possibly two implicit assumptions that I'm not sure are guaranteed to > > be > > > > true. First, I'm not sure if the CallbackHandler interface > guarantees > > > that > > > > implementations must process callbacks in order. Second (and more > > > > plausibly than the first), I'm not sure CallbackHandler guarantees > that > > > > callbacks are to be processed in order until either there are no more > > > left > > > > in the array or one of the elements is an unsupported callback. The > > > > Javadoc simply says it throws UnsupportedCallbackException "if the > > > > implementation of this method does not support one or more of the > > > Callbacks > > > > specified in the callbacks parameter." This statement does not > preclude > > > the > > > > case that implementations might first check to make sure all of the > > > > provided callbacks are supported before processing any of them. > > > > > > > > We could update the Javadoc for AuthenticateCallbackHandler to make > it > > > > clear how implementations must work -- i.e. they must process the > > > callbacks > > > > in order, and they must process all recognized callbacks before > > throwing > > > > UnsupportedCallbackException due to an unrecognized one. > > > > > > > > Note that the above issue does not arise if we simply want the > ability > > to > > > > validate SASL extensions in isolation -- we could just give the > > callback > > > > handler an array containing a single instance of the proposed > > > > SaslExtensionsValidatorCallback.The issue only arises if we want to > > > > provide > > > > additional context (e.g. the token in the case of SASL/OATHBEARER) to > > the > > > > validation mechanism. > > > > > > > > If it is not just SASL Extension validation that we are interested in > > > > adding but in fact we want to be able to provide additional context > to > > > > SaslExtensionsValidatorCallback, then adding the ordering constraint > > > above > > > > is one way, but we could avoid the constraint by allowing > > > > SaslExtensionsValidatorCallback to accept not only the extensions to > > > > validate but also an arbitrary map of name/value pairs. Each SASL > > > > mechanism implementation could declare what additional context it > > > provides > > > > (if any) and at what key(s) the information is available. This > second > > > > approach feels more direct than the first one and would be my > > preference > > > > (assuming I',m not missing anything, which is certainly possible). > > > > Thoughts? > > > > > > > > Ron > > > > > > > > > > > > On Wed, Aug 8, 2018 at 12:39 PM Stanislav Kozlovski < > > > > stanis...@confluent.io> > > > > wrote: > > > > > > > > > Hi Rajini, > > > > > > > > > > That approach makes more sense to me. I like it > > > > > > > > > > On Wed, Aug 8, 2018 at 5:35 PM Rajini Sivaram < > > rajinisiva...@gmail.com > > > > > > > > > wrote: > > > > > > > > > > > Hi Ron/Stanislav, > > > > > > > > > > > > Do you think it makes sense to separate out > > > > OAuthBearerValidatorCallback > > > > > > and SaslExtensionsValidatorCallback so that it is clearer that > > these > > > > are > > > > > > two separate entities that need validation? When we add support > > > > > > for customisable extensions in other mechanisms, we could reuse > > > > > > SaslExtensionsValidatorCallback. We will invoke CallbackHandler > > with > > > { > > > > > > OAuthBearerValidatorCallback, SaslExtensionsValidatorCallback } > in > > > > that > > > > > > order like we do { NameCallback, PasswordCallback }. So typically > > we > > > > > expect > > > > > > to validate tokens with no reference to extensions, but we may > > refer > > > to > > > > > > token to validate extensions. Only validated extensions will be > > > > available > > > > > > as the server's negotiated properties. We will need to handle > > > > > > UnsupportedCallbackException for SaslExtensionsValidatorCallback > > for > > > > > > backwards compatibility, but that should be ok. > > > > > > > > > > > > What do you think? > > > > > > > > > > > > > > > > > > On Wed, Aug 8, 2018 at 5:06 PM, Stanislav Kozlovski < > > > > > > stanis...@confluent.io> > > > > > > wrote: > > > > > > > > > > > > > Hi Ron, > > > > > > > > > > > > > > Yes, I agree we should document it thoroughly > > > > > > > > > > > > > > On Wed, Aug 8, 2018 at 5:02 PM Ron Dagostino < > rndg...@gmail.com> > > > > > wrote: > > > > > > > > > > > > > > > Hi Stanislav. If the community agrees we should add it then > we > > > > > should > > > > > > > at a > > > > > > > > minimum add explicit warnings in the Javadoc making it very > > clear > > > > how > > > > > > > this > > > > > > > > should not be used. > > > > > > > > > > > > > > > > Ron > > > > > > > > > > > > > > > > On Wed, Aug 8, 2018 at 11:54 AM Stanislav Kozlovski < > > > > > > > > stanis...@confluent.io> > > > > > > > > wrote: > > > > > > > > > > > > > > > > > Hey Ron, > > > > > > > > > > > > > > > > > > I fully agree that token validation is a serious security > > > > > operation. > > > > > > > > > Although, I believe allowing the user to do more validation > > > with > > > > > the > > > > > > > > > extensions does not hurt - the user is fully responsible > for > > > his > > > > > > > security > > > > > > > > > once he starts implementing custom code for token > validation. > > > You > > > > > > would > > > > > > > > > expect people to take the appropriate considerations when > > > > > validating > > > > > > > > > unsecured extensions against the token. > > > > > > > > > I also think that using the extensions as a secondary > > > validation > > > > > > method > > > > > > > > > might be useful. You could do your normal validation using > > the > > > > > token > > > > > > > and > > > > > > > > > then have a second sanity-check validation on top (e.g > > validate > > > > > > > > > hostname/port is what client expected). Keep in mind that > the > > > > > server > > > > > > > > > exposes the properties via `getNegotiatedProperty` so it > > makes > > > > > sense > > > > > > to > > > > > > > > > allow the server to have custom validation on the > extensions. > > > > > > > > > > > > > > > > > > Best, > > > > > > > > > Stanislav > > > > > > > > > > > > > > > > > > On Wed, Aug 8, 2018 at 3:29 PM Ron Dagostino < > > > rndg...@gmail.com> > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > Hi Stanislav. If you wanted to do this a good way might > be > > > to > > > > > add > > > > > > a > > > > > > > > > > constructor to the org.apache.kafka.common. > > > > security.oauthbearer. > > > > > > > > > > OAuthBearerValidatorCallback class that accepts a > > > > SaslExtensions > > > > > > > > instance > > > > > > > > > > in addition to a token value. Then it would give the > > > callback > > > > > > > handler > > > > > > > > > the > > > > > > > > > > option to introspect the callback to see what extensions > > were > > > > > > > provided > > > > > > > > > with > > > > > > > > > > the > > > > > > > > > > token. > > > > > > > > > > > > > > > > > > > > That being said, token validation is a very > > > security-sensitive > > > > > > > > operation, > > > > > > > > > > and it would be a serious security issue if the result of > > > > > applying > > > > > > > the > > > > > > > > > > validation algorithm (which yields a valid vs. not valid > > > > > > > determination) > > > > > > > > > > depended on anything provided by the client other than > the > > > > actual > > > > > > > token > > > > > > > > > > value. Nobody should ever allow the client to specify a > > JWK > > > > Set > > > > > > URL, > > > > > > > > for > > > > > > > > > > example, or a whitelist of acceptable domains for > > retrieving > > > > JWK > > > > > > > Sets. > > > > > > > > > It > > > > > > > > > > feels to me that while a use case might exist (some kind > of > > > > trace > > > > > > ID, > > > > > > > > for > > > > > > > > > > example, to aid in debugging), someone might > inadvertently > > > hang > > > > > > > > > themselves > > > > > > > > > > if we give them the rope. The risk vs. reward value > > > > proposition > > > > > > here > > > > > > > > > > doesn't feel like a good one at first glance. Thoughts? > > > > > > > > > > > > > > > > > > > > Ron > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Wed, Aug 8, 2018 at 10:04 AM Stanislav Kozlovski < > > > > > > > > > > stanis...@confluent.io> > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > Hey everybody, > > > > > > > > > > > > > > > > > > > > > > Sorry for reviving this, but I neglected something the > > > first > > > > > time > > > > > > > > > around. > > > > > > > > > > > I believe it would be useful to provide the > > > > > > > > > > > `OAuthBearerUnsecuredValidatorCallbackHandler` with the > > > > OAuth > > > > > > > > > extensions > > > > > > > > > > > too. This would enable use cases where validators want > to > > > > > > reconcile > > > > > > > > > > > information from the extensions with the token (e.g if > > > users > > > > > have > > > > > > > > > > > implemented secured OAuth tokens). > > > > > > > > > > > The implementation would be to instantiate > > > > > > > > > > > `OAuthBearerUnsecuredValidatorCallback` with the > > extensions > > > > > (also > > > > > > > > leave > > > > > > > > > > the > > > > > > > > > > > current constructor, as its a public class). > > > > > > > > > > > > > > > > > > > > > > What are everybody's thoughts on this? If there are no > > > > > > objections, > > > > > > > > I'll > > > > > > > > > > > update the KIP in due time > > > > > > > > > > > > > > > > > > > > > > On Thu, Jul 26, 2018 at 11:14 AM Rajini Sivaram < > > > > > > > > > rajinisiva...@gmail.com > > > > > > > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > Looks good. Thanks, Stanislav. > > > > > > > > > > > > > > > > > > > > > > > > On Wed, Jul 25, 2018 at 7:46 PM, Stanislav Kozlovski > < > > > > > > > > > > > > stanis...@confluent.io > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > Hi Rajini, > > > > > > > > > > > > > > > > > > > > > > > > > > I updated the KIP. Please check if the > clarification > > is > > > > > okay > > > > > > > > > > > > > > > > > > > > > > > > > > On Wed, Jul 25, 2018 at 10:49 AM Rajini Sivaram < > > > > > > > > > > > rajinisiva...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Stanislav, > > > > > > > > > > > > > > > > > > > > > > > > > > > > 1. Can you clarify the following line in the KIP > in > > > the > > > > > > > 'Public > > > > > > > > > > > > > Interfaces' > > > > > > > > > > > > > > section? When you are reading the KIP for the > first > > > > time, > > > > > > it > > > > > > > > > sounds > > > > > > > > > > > > like > > > > > > > > > > > > > we > > > > > > > > > > > > > > adding a new Kafka config. But we are adding JAAS > > > > config > > > > > > > > options > > > > > > > > > > > with a > > > > > > > > > > > > > > prefix that can be used with the default > unsecured > > > > bearer > > > > > > > > tokens. > > > > > > > > > > We > > > > > > > > > > > > > could > > > > > > > > > > > > > > include the example in this section or at least > > link > > > to > > > > > the > > > > > > > > > > example. > > > > > > > > > > > > > > > > > > > > > > > > > > > > - New config option for default, unsecured > > bearer > > > > > > tokens - > > > > > > > > > > > > > > `unsecuredLoginExtension_<extensionname>`. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > 2. Can you add the package for > > SaslExtensionsCallback > > > > > > class? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Tue, Jul 24, 2018 at 10:03 PM, Stanislav > > > Kozlovski < > > > > > > > > > > > > > > stanis...@confluent.io> wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Ron, > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Thanks for the suggestions. I have applied them > > to > > > > the > > > > > > KIP. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Tue, Jul 24, 2018 at 1:39 PM Ron Dagostino < > > > > > > > > > rndg...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Stanislav. The statement "New config > option > > > for > > > > > > > > > > > > > > > OAuthBearerLoginModule" > > > > > > > > > > > > > > > > is technically incorrect; it should be "New > > > config > > > > > > option > > > > > > > > for > > > > > > > > > > > > > default, > > > > > > > > > > > > > > > > unsecured bearer tokens" since that is what > > > > provides > > > > > > the > > > > > > > > > > > > > functionality > > > > > > > > > > > > > > > (as > > > > > > > > > > > > > > > > opposed to the login module, which does not). > > > > Also, > > > > > > > please > > > > > > > > > > state > > > > > > > > > > > > > that > > > > > > > > > > > > > > > > "auth" is not supported as a custom extension > > > name > > > > > with > > > > > > > any > > > > > > > > > > > > > > > > SASL/OAUTHBEARER mechanism, including the > > > unsecured > > > > > > one, > > > > > > > > > since > > > > > > > > > > it > > > > > > > > > > > > is > > > > > > > > > > > > > > > > reserved by the spec for what is normally > sent > > in > > > > the > > > > > > > HTTP > > > > > > > > > > > > > > Authorization > > > > > > > > > > > > > > > > header an attempt to use it will result in a > > > > > > > configuration > > > > > > > > > > > > exception. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Finally, please also state that while the > > > > > > > > > > OAuthBearerLoginModule > > > > > > > > > > > > and > > > > > > > > > > > > > > the > > > > > > > > > > > > > > > > OAuthBearerSaslClient will be changed to > > request > > > > the > > > > > > > > > extensions > > > > > > > > > > > > from > > > > > > > > > > > > > > its > > > > > > > > > > > > > > > > callback handler, for backwards compatibility > > it > > > is > > > > > not > > > > > > > > > > necessary > > > > > > > > > > > > for > > > > > > > > > > > > > > the > > > > > > > > > > > > > > > > callback handler to support > > > SaslExtensionsCallback > > > > -- > > > > > > any > > > > > > > > > > > > > > > > UnsupportedCallbackException that is thrown > > will > > > be > > > > > > > ignored > > > > > > > > > and > > > > > > > > > > > no > > > > > > > > > > > > > > > > extensions will be added. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Ron > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Tue, Jul 24, 2018 at 11:20 AM Stanislav > > > > Kozlovski > > > > > < > > > > > > > > > > > > > > > > stanis...@confluent.io> > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hey everybody, > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > I have updated the KIP to reflect the > latest > > > > > changes > > > > > > as > > > > > > > > > best > > > > > > > > > > > as I > > > > > > > > > > > > > > > could. > > > > > > > > > > > > > > > > If > > > > > > > > > > > > > > > > > there aren't more suggestions, I intent to > > > start > > > > > the > > > > > > > > [VOTE] > > > > > > > > > > > > thread > > > > > > > > > > > > > > > > > tomorrow. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Best, > > > > > > > > > > > > > > > > > Stanislav > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Tue, Jul 24, 2018 at 6:34 AM Ron > > Dagostino < > > > > > > > > > > > rndg...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Stanislav. Could you update the KIP > to > > > > > reflect > > > > > > > the > > > > > > > > > > latest > > > > > > > > > > > > > > > > definition > > > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > > SaslExtensions and confirm or correct the > > > > impact > > > > > it > > > > > > > has > > > > > > > > > to > > > > > > > > > > > the > > > > > > > > > > > > > > > > > > SCRAM-related classes? I'm not sure if > the > > > > > > > > > > > currently-described > > > > > > > > > > > > > > > impact > > > > > > > > > > > > > > > > is > > > > > > > > > > > > > > > > > > still accurate. Also, could you mention > > the > > > > > > changes > > > > > > > to > > > > > > > > > > > > > > > > > > OAuthBearerUnsecuredLoginCallbackHandler > > in > > > > the > > > > > > > text in > > > > > > > > > > > > > addition to > > > > > > > > > > > > > > > > > giving > > > > > > > > > > > > > > > > > > the examples? The examples show the new > > > > > > > > > > > > > > > > > > unsecuredLoginExtension_<extensionName> > > > > feature, > > > > > > but > > > > > > > > that > > > > > > > > > > > > > feature > > > > > > > > > > > > > > is > > > > > > > > > > > > > > > > not > > > > > > > > > > > > > > > > > > described anywhere prior to it appearing > > > there. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Ron > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Mon, Jul 23, 2018 at 1:42 PM Ron > > > Dagostino < > > > > > > > > > > > > rndg...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Rajini. I think a class is fine as > > long > > > > as > > > > > we > > > > > > > > make > > > > > > > > > > sure > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > semantics > > > > > > > > > > > > > > > > > > > of immutability are clear -- it would > > have > > > to > > > > > be > > > > > > a > > > > > > > > > value > > > > > > > > > > > > class, > > > > > > > > > > > > > > and > > > > > > > > > > > > > > > > any > > > > > > > > > > > > > > > > > > > constructor that accepts a Map as input > > > would > > > > > > have > > > > > > > to > > > > > > > > > > copy > > > > > > > > > > > > that > > > > > > > > > > > > > > Map > > > > > > > > > > > > > > > > > > rather > > > > > > > > > > > > > > > > > > > than store it in a member variable. > > > > Similarly, > > > > > > any > > > > > > > > Map > > > > > > > > > > > that > > > > > > > > > > > > it > > > > > > > > > > > > > > > might > > > > > > > > > > > > > > > > > > > return would have to be unmodifiable. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Ron > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Mon, Jul 23, 2018 at 12:24 PM Rajini > > > > > Sivaram < > > > > > > > > > > > > > > > > > rajinisiva...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >> Hi Ron, Stanislav, > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > >> I agree with Stanislav that it would > be > > > > better > > > > > > to > > > > > > > > > leave > > > > > > > > > > > > > > > > > `SaslExtensions` > > > > > > > > > > > > > > > > > > >> as > > > > > > > > > > > > > > > > > > >> a class rather than make it an > > interface. > > > We > > > > > > > don''t > > > > > > > > > > really > > > > > > > > > > > > > > expect > > > > > > > > > > > > > > > > > users > > > > > > > > > > > > > > > > > > to > > > > > > > > > > > > > > > > > > >> extends this class, so it is > convenient > > to > > > > > have > > > > > > an > > > > > > > > > > > > > > implementation > > > > > > > > > > > > > > > > > since > > > > > > > > > > > > > > > > > > >> users need to create an instance. The > > > class > > > > > > > provided > > > > > > > > > by > > > > > > > > > > > the > > > > > > > > > > > > > > public > > > > > > > > > > > > > > > > API > > > > > > > > > > > > > > > > > > >> should be sufficient in the vast > > majority > > > of > > > > > the > > > > > > > > > cases. > > > > > > > > > > > Ron, > > > > > > > > > > > > > do > > > > > > > > > > > > > > > you > > > > > > > > > > > > > > > > > > agree? > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > >> On Mon, Jul 23, 2018 at 11:35 AM, Ron > > > > > Dagostino > > > > > > < > > > > > > > > > > > > > > > rndg...@gmail.com> > > > > > > > > > > > > > > > > > > >> wrote: > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > >> > Hi Stanislav. See > > > > > > > > > > > > > > > > https://tools.ietf.org/html/ > > rfc7628#section-3.1, > > > > > > > > > > > > > > > > > > and > > > > > > > > > > > > > > > > > > >> > that section refers to the core ABNF > > > > > > productions > > > > > > > > > > defined > > > > > > > > > > > > in > > > > > > > > > > > > > > > > > > >> > > > > > > > https://tools.ietf.org/html/rfc5234#appendix-B. > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > >> > Ron > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > >> > > On Jul 23, 2018, at 1:30 AM, > > Stanislav > > > > > > > > Kozlovski < > > > > > > > > > > > > > > > > > > >> stanis...@confluent.io> > > > > > > > > > > > > > > > > > > >> > wrote: > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > > >> > > Hey Ron and Rajini, > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > > >> > > Here are my thoughts: > > > > > > > > > > > > > > > > > > >> > > Regarding separators in > > > SaslExtensions - > > > > > > > Agreed, > > > > > > > > > > that > > > > > > > > > > > > was > > > > > > > > > > > > > a > > > > > > > > > > > > > > > bad > > > > > > > > > > > > > > > > > > move. > > > > > > > > > > > > > > > > > > >> > > Should definitely not be a concern > > of > > > > > > > > > > CallbackHandler > > > > > > > > > > > > and > > > > > > > > > > > > > > > > > > LoginModule > > > > > > > > > > > > > > > > > > >> > > implementors. > > > > > > > > > > > > > > > > > > >> > > SaslExtensions interface - > Wouldn't > > > > > > > implementing > > > > > > > > > it > > > > > > > > > > as > > > > > > > > > > > > an > > > > > > > > > > > > > > > > > interface > > > > > > > > > > > > > > > > > > >> mean > > > > > > > > > > > > > > > > > > >> > > that users will have to make sure > > > > they're > > > > > > > > passing > > > > > > > > > in > > > > > > > > > > > an > > > > > > > > > > > > > > > > > unmodifiable > > > > > > > > > > > > > > > > > > >> map > > > > > > > > > > > > > > > > > > >> > > themselves. I believe it would be > > > better > > > > > if > > > > > > we > > > > > > > > > > > enforced > > > > > > > > > > > > > that > > > > > > > > > > > > > > > > > through > > > > > > > > > > > > > > > > > > >> > class > > > > > > > > > > > > > > > > > > >> > > constructors instead. > > > > > > > > > > > > > > > > > > >> > > SaslExtensions#map() - I'd also > > prefer > > > > > this. > > > > > > > The > > > > > > > > > > > reason > > > > > > > > > > > > I > > > > > > > > > > > > > > went > > > > > > > > > > > > > > > > > with > > > > > > > > > > > > > > > > > > >> > > `extensionValue` and > > `extensionNames` > > > > was > > > > > > > > because > > > > > > > > > I > > > > > > > > > > > > > figured > > > > > > > > > > > > > > it > > > > > > > > > > > > > > > > > made > > > > > > > > > > > > > > > > > > >> sense > > > > > > > > > > > > > > > > > > >> > > to have `ScramExtensions` extend > > > > > > > > `SaslExtensions` > > > > > > > > > > and > > > > > > > > > > > > > > > therefore > > > > > > > > > > > > > > > > > have > > > > > > > > > > > > > > > > > > >> > their > > > > > > > > > > > > > > > > > > >> > > API be similar. In the end, do you > > > think > > > > > > that > > > > > > > it > > > > > > > > > is > > > > > > > > > > > > worth > > > > > > > > > > > > > it > > > > > > > > > > > > > > > to > > > > > > > > > > > > > > > > > have > > > > > > > > > > > > > > > > > > >> > > `ScramExtensions` extend > > > > `SaslExtensions`? > > > > > > > > > > > > > > > > > > >> > > @Ron, could you point me to the > SASL > > > > OAuth > > > > > > > > > mechanism > > > > > > > > > > > > > > specific > > > > > > > > > > > > > > > > > > regular > > > > > > > > > > > > > > > > > > >> > > expressions for keys/values you > > > > mentioned > > > > > > are > > > > > > > in > > > > > > > > > RFC > > > > > > > > > > > > 7628 > > > > > > > > > > > > > ( > > > > > > > > > > > > > > > > > > >> > > > https://tools.ietf.org/html/rfc7628 > > ) > > > ? > > > > I > > > > > > > could > > > > > > > > > not > > > > > > > > > > > find > > > > > > > > > > > > > any > > > > > > > > > > > > > > > > while > > > > > > > > > > > > > > > > > > >> > > originally implementing this. > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > > >> > > Best, > > > > > > > > > > > > > > > > > > >> > > Stanislav > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > > >> > >> On Sun, Jul 22, 2018 at 6:46 PM > Ron > > > > > > > Dagostino < > > > > > > > > > > > > > > > > rndg...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >> > wrote: > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> Hi again, Rajini and Stanislav. > I > > > > wonder > > > > > > if > > > > > > > > > making > > > > > > > > > > > > > > > > > SaslExtensions > > > > > > > > > > > > > > > > > > an > > > > > > > > > > > > > > > > > > >> > >> interface rather than a class > might > > > be > > > > a > > > > > > good > > > > > > > > > > > solution. > > > > > > > > > > > > > > For > > > > > > > > > > > > > > > > > > example: > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> public interface SaslExtensions { > > > > > > > > > > > > > > > > > > >> > >> /** > > > > > > > > > > > > > > > > > > >> > >> * @return an immutable map > view > > of > > > > the > > > > > > > SASL > > > > > > > > > > > > extensions > > > > > > > > > > > > > > > > > > >> > >> */ > > > > > > > > > > > > > > > > > > >> > >> Map<String, String> map(); > > > > > > > > > > > > > > > > > > >> > >> } > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> This solves the issue of lack of > > > > clarity > > > > > on > > > > > > > > > > > > immutability, > > > > > > > > > > > > > > and > > > > > > > > > > > > > > > > it > > > > > > > > > > > > > > > > > > also > > > > > > > > > > > > > > > > > > >> > >> eliminates copying, like this: > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> SaslExtensions myMethod() { > > > > > > > > > > > > > > > > > > >> > >> Map<String, String> myRetval = > > > > > > > > > > > > > > > > > > getUnmodifiableSaslExtensionsMap(); > > > > > > > > > > > > > > > > > > >> > >> return new SaslExtensions() { > > > > > > > > > > > > > > > > > > >> > >> public Map<String, String> > > > > map() { > > > > > > > > > > > > > > > > > > >> > >> return myRetval; > > > > > > > > > > > > > > > > > > >> > >> } > > > > > > > > > > > > > > > > > > >> > >> } > > > > > > > > > > > > > > > > > > >> > >> } > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> Alternatively, we could do it > like > > > > this: > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> /** > > > > > > > > > > > > > > > > > > >> > >> * Supplier that returns immutable > > map > > > > > view > > > > > > of > > > > > > > > > SASL > > > > > > > > > > > > > > Extensions > > > > > > > > > > > > > > > > > > >> > >> */ > > > > > > > > > > > > > > > > > > >> > >> public interface SaslExtensions > > > extends > > > > > > > > > > > > > > Supplier<Map<String, > > > > > > > > > > > > > > > > > > >> String>> { > > > > > > > > > > > > > > > > > > >> > >> // empty > > > > > > > > > > > > > > > > > > >> > >> } > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> The we could simply return the > > > instance > > > > > > like > > > > > > > > > this, > > > > > > > > > > > > again > > > > > > > > > > > > > > > > without > > > > > > > > > > > > > > > > > > >> > copying: > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> SaslExtensions myMethod() { > > > > > > > > > > > > > > > > > > >> > >> Map<String, String> myRetval = > > > > > > > > > > > > > > > > > > getUnmodifiableSaslExtensionsMap(); > > > > > > > > > > > > > > > > > > >> > >> return () -> myRetval; > > > > > > > > > > > > > > > > > > >> > >> } > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> I think the main reason for > making > > > > > > > > SaslExtensions > > > > > > > > > > > part > > > > > > > > > > > > of > > > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > > public > > > > > > > > > > > > > > > > > > >> > >> interface is to avoid adding a > Map > > to > > > > the > > > > > > > > > Subject's > > > > > > > > > > > > > public > > > > > > > > > > > > > > > > > > >> credentials. > > > > > > > > > > > > > > > > > > >> > >> Making SaslExtensions an > interface > > > > meets > > > > > > that > > > > > > > > > > > > requirement > > > > > > > > > > > > > > and > > > > > > > > > > > > > > > > > then > > > > > > > > > > > > > > > > > > >> > allows > > > > > > > > > > > > > > > > > > >> > >> us to be free to implement > whatever > > > we > > > > > want > > > > > > > > > > > internally. > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> Thoughts? > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >> Ron > > > > > > > > > > > > > > > > > > >> > >> > > > > > > > > > > > > > > > > > > >> > >>> On Sun, Jul 22, 2018 at 12:45 PM > > Ron > > > > > > > > Dagostino < > > > > > > > > > > > > > > > > > rndg...@gmail.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >> > wrote: > > > > > > > > > > > > > > > > > > >> > >>> > > > > > > > > > > > > > > > > > > >> > >>> Hi Rajini. The SaslServer is > > going > > > to > > > > > > have > > > > > > > to > > > > > > > > > > > > validate > > > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > > >> extensions, > > > > > > > > > > > > > > > > > > >> > >>> too, but I’m okay with keeping > the > > > > > > > validation > > > > > > > > > > logic > > > > > > > > > > > > > > > elsewhere > > > > > > > > > > > > > > > > as > > > > > > > > > > > > > > > > > > >> long > > > > > > > > > > > > > > > > > > >> > as > > > > > > > > > > > > > > > > > > >> > >> it > > > > > > > > > > > > > > > > > > >> > >>> can be reused in both the client > > and > > > > the > > > > > > > > secret. > > > > > > > > > > > > > > > > > > >> > >>> > > > > > > > > > > > > > > > > > > >> > >>> I strongly prefer exposing a > map() > > > > > method > > > > > > as > > > > > > > > > > opposed > > > > > > > > > > > > to > > > > > > > > > > > > > > > > > > >> > extensionNames() > > > > > > > > > > > > > > > > > > >> > >>> and extensionValue(String) > > methods. > > > It > > > > > is > > > > > > a > > > > > > > > > > smaller > > > > > > > > > > > > API > > > > > > > > > > > > > (2 > > > > > > > > > > > > > > > > > methods > > > > > > > > > > > > > > > > > > >> > >> instead > > > > > > > > > > > > > > > > > > >> > >>> of 1), and it gives clients of > the > > > API > > > > > > full > > > > > > > > > > > > map-related > > > > > > > > > > > > > > > > > > >> functionality > > > > > > > > > > > > > > > > > > >> > >>> (there’s a lot of support for > > > dealing > > > > > with > > > > > > > > maps > > > > > > > > > > in a > > > > > > > > > > > > > > variety > > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > > >> ways). > > > > > > > > > > > > > > > > > > >> > >>> > > > > > > > > > > > > > > > > > > >> > >>> Regardless of whether we go > with a > > > > map() > > > > > > > > method > > > > > > > > > or > > > > > > > > > > > > > > > > > > extensionNames() > > > > > > > > > > > > > > > > > > >> and > > > > > > > > > > > > > > > > > > >> > >>> extensionValue(String) methods, > > the > > > > > > > semantics > > > > > > > > of > > > > > > > > > > > > > > mutability > > > > > > > > > > > > > > > > need > > > > > > > > > > > > > > > > > > to > > > > > > > > > > > > > > > > > > >> be > > > > > > > > > > > > > > > > > > >> > >>> clear. I think either way we > > should > > > > > never > > > > > > > > > share a > > > > > > > > > > > map > > > > > > > > > > > > > > that > > > > > > > > > > > > > > > > > anyone > > > > > > > > > > > > > > > > > > >> else > > > > > > > > > > > > > > > > > > >> > >>> could possibly mutate — either a > > map > > > > > that > > > > > > > > > someone > > > > > > > > > > > > gives > > > > > > > > > > > > > us > > > > > > > > > > > > > > > or > > > > > > > > > > > > > > > > a > > > > > > > > > > > > > > > > > > map > > > > > > > > > > > > > > > > > > >> > that > > > > > > > > > > > > > > > > > > >> > >> we > > > > > > > > > > > > > > > > > > >> > >>> might expose. > > > > > > > > > > > > > > > > > > >> > >>> > > > > > > > > > > > > > > > > > > >> > >>> Thoughts? > > > > > > > > > > > > > > > > > > >> > >>> > > > > > > > > > > > > > > > > > > >> > >>> Ron > > > > > > > > > > > > > > > > > > >> > >>> > > > > > > > > > > > > > > > > > > >> > >>>> On Jul 22, 2018, at 11:23 AM, > > > Rajini > > > > > > > Sivaram > > > > > > > > < > > > > > > > > > > > > > > > > > > >> rajinisiva...@gmail.com > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > > >> > >>> wrote: > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> Hmm.... I think we need a much > > > > simpler > > > > > > > > > > > SaslExtensions > > > > > > > > > > > > > > class > > > > > > > > > > > > > > > > if > > > > > > > > > > > > > > > > > we > > > > > > > > > > > > > > > > > > >> are > > > > > > > > > > > > > > > > > > >> > >>>> making it part of the public > API. > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> 1. I don't see the point of > > > including > > > > > > > > separator > > > > > > > > > > > > > anywhere > > > > > > > > > > > > > > in > > > > > > > > > > > > > > > > > > >> > >>> SaslExtensions. > > > > > > > > > > > > > > > > > > >> > >>>> Extensions provide a map and we > > > > > propagate > > > > > > > the > > > > > > > > > map > > > > > > > > > > > > from > > > > > > > > > > > > > > > client > > > > > > > > > > > > > > > > > to > > > > > > > > > > > > > > > > > > >> > server > > > > > > > > > > > > > > > > > > >> > >>>> using the protocol associated > > with > > > > the > > > > > > > > > mechanism > > > > > > > > > > in > > > > > > > > > > > > > use. > > > > > > > > > > > > > > > The > > > > > > > > > > > > > > > > > > >> separator > > > > > > > > > > > > > > > > > > >> > >> is > > > > > > > > > > > > > > > > > > >> > >>>> not configurable and should not > > be > > > a > > > > > > > concern > > > > > > > > of > > > > > > > > > > the > > > > > > > > > > > > > > > > implementor > > > > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > > >> > >>>> SaslExtensionsCallback > interface > > > that > > > > > > > > provides > > > > > > > > > an > > > > > > > > > > > > > > instance > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > > >> > >>> SaslExtensions > > > > > > > > > > > > > > > > > > >> > >>>> . > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> 2. I agree with Ron that we > need > > > > > > > > > > mechanism-specific > > > > > > > > > > > > > > > > validation > > > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > > >> the > > > > > > > > > > > > > > > > > > >> > >>>> values from SaslExtensions. > But I > > > > think > > > > > > we > > > > > > > > > could > > > > > > > > > > do > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > > validation > > > > > > > > > > > > > > > > > > >> in > > > > > > > > > > > > > > > > > > >> > >> the > > > > > > > > > > > > > > > > > > >> > >>>> appropriate `SaslClient` > > > > implementation > > > > > > of > > > > > > > > that > > > > > > > > > > > > > > mechanism. > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> I think we could just have a > very > > > > > simple > > > > > > > > > > extensions > > > > > > > > > > > > > class > > > > > > > > > > > > > > > and > > > > > > > > > > > > > > > > > > move > > > > > > > > > > > > > > > > > > >> > >>>> everything else to appropriate > > > > internal > > > > > > > > classes > > > > > > > > > > of > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > mechanisms > > > > > > > > > > > > > > > > > > >> > using > > > > > > > > > > > > > > > > > > >> > >>>> extensions. What do you think? > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> public class SaslExtensions { > > > > > > > > > > > > > > > > > > >> > >>>> private final Map<String, > > String> > > > > > > > > > extensionMap; > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> public > > SaslExtensions(Map<String, > > > > > > String> > > > > > > > > > > > > > > extensionMap) { > > > > > > > > > > > > > > > > > > >> > >>>> this.extensionMap = > > > > extensionMap; > > > > > > > > > > > > > > > > > > >> > >>>> } > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> public String > > > extensionValue(String > > > > > > > name) { > > > > > > > > > > > > > > > > > > >> > >>>> return > > > extensionMap.get(name); > > > > > > > > > > > > > > > > > > >> > >>>> } > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> public Set<String> > > > > extensionNames() { > > > > > > > > > > > > > > > > > > >> > >>>> return > > extensionMap.keySet(); > > > > > > > > > > > > > > > > > > >> > >>>> } > > > > > > > > > > > > > > > > > > >> > >>>> } > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>> > > > > > > > > > > > > > > > > > > >> > >>>>> On Sat, Jul 21, 2018 at 9:01 > PM, > > > Ron > > > > > > > > > Dagostino < > > > > > > > > > > > > > > > > > > rndg...@gmail.com > > > > > > > > > > > > > > > > > > >> > > > > > > > > > > > > > > > > > > > >> > >>> wrote: > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> Hi Stanislav and Rajini. If > > > > > > > SaslExtensions > > > > > > > > is > > > > > > > > > > > going > > > > > > > > > > > > > to > > > > > > > > > > > > > > > part > > > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > > >> the > > > > > > > > > > > > > > > > > > >> > >>> public > > > > > > > > > > > > > > > > > > >> > >>>>> API, then it occurred to me > that > > > one > > > > > of > > > > > > > the > > > > > > > > > > > > > requirements > > > > > > > > > > > > > > > of > > > > > > > > > > > > > > > > > all > > > > > > > > > > > > > > > > > > >> SASL > > > > > > > > > > > > > > > > > > >> > >>>>> extensions is that the keys > and > > > > values > > > > > > > need > > > > > > > > to > > > > > > > > > > > match > > > > > > > > > > > > > > > > > > >> > >> mechanism-specific > > > > > > > > > > > > > > > > > > >> > >>>>> regular expressions. For > > example, > > > > RFC > > > > > > > 5802 > > > > > > > > ( > > > > > > > > > > > > > > > > > > >> > >>>>> > > > https://tools.ietf.org/html/rfc5802 > > > > ) > > > > > > > > > specifies > > > > > > > > > > > the > > > > > > > > > > > > > > > regular > > > > > > > > > > > > > > > > > > >> > >> expressions > > > > > > > > > > > > > > > > > > >> > >>> for > > > > > > > > > > > > > > > > > > >> > >>>>> the SCRAM-specific SASL > > > mechanisms, > > > > > and > > > > > > > RFC > > > > > > > > > > 7628 ( > > > > > > > > > > > > > > > > > > >> > >>>>> > > > https://tools.ietf.org/html/rfc7628 > > > > ) > > > > > > > > > specifies > > > > > > > > > > > > > > different > > > > > > > > > > > > > > > > > > regular > > > > > > > > > > > > > > > > > > >> > >>>>> expressions for the > OAUTHBEARER > > > SASL > > > > > > > > > > mechanism. I > > > > > > > > > > > > am > > > > > > > > > > > > > > > > thinking > > > > > > > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > > >> > >>>>> SaslExtensions class should > > > probably > > > > > > > > provide a > > > > > > > > > > way > > > > > > > > > > > > to > > > > > > > > > > > > > > make > > > > > > > > > > > > > > > > > sure > > > > > > > > > > > > > > > > > > >> the > > > > > > > > > > > > > > > > > > >> > >> keys > > > > > > > > > > > > > > > > > > >> > >>>>> and values match the > appropriate > > > > > regular > > > > > > > > > > > > expressions. > > > > > > > > > > > > > > > What > > > > > > > > > > > > > > > > do > > > > > > > > > > > > > > > > > > you > > > > > > > > > > > > > > > > > > >> > >>> think of > > > > > > > > > > > > > > > > > > >> > >>>>> something along the lines of > the > > > > below > > > > > > > > > > definition > > > > > > > > > > > > for > > > > > > > > > > > > > > the > > > > > > > > > > > > > > > > > > >> > >> SaslExtensions > > > > > > > > > > > > > > > > > > >> > >>>>> class? It is missing Javadoc > > and > > > > > > > > > > > > > > > > > toString()/hashCode()/equals() > > > > > > > > > > > > > > > > > > >> > >>> methods, > > > > > > > > > > > > > > > > > > >> > >>>>> of course, but aside from > that, > > do > > > > you > > > > > > > think > > > > > > > > > > this > > > > > > > > > > > is > > > > > > > > > > > > > > > > > sufficient > > > > > > > > > > > > > > > > > > >> and > > > > > > > > > > > > > > > > > > >> > >>>>> appropriate? > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> Ron > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> public class SaslExtensions { > > > > > > > > > > > > > > > > > > >> > >>>>> private final Map<String, > > > String> > > > > > > > > > > extensionsMap; > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> public SaslExtensions(String > > > > mapStr, > > > > > > > > String > > > > > > > > > > > > > > > > > keyValueSeparator, > > > > > > > > > > > > > > > > > > >> > >> String > > > > > > > > > > > > > > > > > > >> > >>>>> elementSeparator, > > > > > > > > > > > > > > > > > > >> > >>>>> Pattern > > > > > saslNameRegexPattern, > > > > > > > > > Pattern > > > > > > > > > > > > > > > > > > >> > >> saslValueRegexPattern) > > > > > > > > > > > > > > > > > > >> > >>> { > > > > > > > > > > > > > > > > > > >> > >>>>> > > this(Utils.parseMap(mapStr, > > > > > > > > > > > keyValueSeparator, > > > > > > > > > > > > > > > > > > >> > >> elementSeparator), > > > > > > > > > > > > > > > > > > >> > >>>>> saslNameRegexPattern, > > > > > > > > saslValueRegexPattern); > > > > > > > > > > > > > > > > > > >> > >>>>> } > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> public > > > SaslExtensions(Map<String, > > > > > > > String> > > > > > > > > > > > > > > extensionsMap, > > > > > > > > > > > > > > > > > > Pattern > > > > > > > > > > > > > > > > > > >> > >>>>> saslNameRegexPattern, > > > > > > > > > > > > > > > > > > >> > >>>>> Pattern > > > > > > saslValueRegexPattern) { > > > > > > > > > > > > > > > > > > >> > >>>>> Map<String, String> > > > > > sanitizedCopy > > > > > > = > > > > > > > > new > > > > > > > > > > > > > > > > > > >> > >>>>> HashMap<>(extensionsMap.size() > > ); > > > > > > > > > > > > > > > > > > >> > >>>>> for (Entry<String, > String> > > > > > entry : > > > > > > > > > > > > > > > > > > >> extensionsMap.entrySet()) { > > > > > > > > > > > > > > > > > > >> > >>>>> if > > > (!saslNameRegexPattern. > > > > > > > > > > > > > > > matcher(entry.getKey()). > > > > > > > > > > > > > > > > > > >> > matches() > > > > > > > > > > > > > > > > > > >> > >>>>> || > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > !saslValueRegexPattern.matcher(entry.getValue()). > > > > > > > > > > > > > > > matches()) > > > > > > > > > > > > > > > > > > >> > >>>>> throw new > > > > > > > > > > > > > > IllegalArgumentException("Invalid > > > > > > > > > > > > > > > > key > > > > > > > > > > > > > > > > > or > > > > > > > > > > > > > > > > > > >> > >>>>> value"); > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > sanitizedCopy.put(entry.getKe > > > > > > > y(), > > > > > > > > > > > > > > > > entry.getValue()); > > > > > > > > > > > > > > > > > > >> > >>>>> } > > > > > > > > > > > > > > > > > > >> > >>>>> this.extensionsMap = > > > > > > > > > > > > > > > > > > >> > >> > > > > > Collections.unmodifiableMap(sanitizedCopy); > > > > > > > > > > > > > > > > > > >> > >>>>> } > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> public Map<String, String> > > > map() { > > > > > > > > > > > > > > > > > > >> > >>>>> return extensionsMap; > > > > > > > > > > > > > > > > > > >> > >>>>> } > > > > > > > > > > > > > > > > > > >> > >>>>> } > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>> On Fri, Jul 20, 2018 at 12:49 > PM > > > > > > Stanislav > > > > > > > > > > > > Kozlovski < > > > > > > > > > > > > > > > > > > >> > >>>>> stanis...@confluent.io> > > > > > > > > > > > > > > > > > > >> > >>>>> wrote: > > > > > > > > > > > > > > > > > > >> > >>>>> > > > > > > > > > > > > > > > > > > >> > >>>>>> Hi Ron, > > > > > > > > > > > > > > > > > > >> > >>>>>> > > > > > > > > > > > > > > > > > > >> > >>>>>> I saw that and decided that > > would > > > > be > > > > > > the > > > > > > > > best > > > > > > > > > > > > > approach. > > > > > > > > > > > > > > > The > > > > > > > > > > > > > > > > > > >> current > > > > > > > > > > > > > > > > > > >> > >>>>>> ScramExtensions > implementation > > > > uses a > > > > > > Map > > > > > > > > in > > > > > > > > > > the > > > > > > > > > > > > > public > > > > > > > > > > > > > > > > > > >> credentials > > > > > > > > > > > > > > > > > > >> > >>> and I > > > > > > > > > > > > > > > > > > >> > >>>>>> thought I would follow > > convention > > > > > > rather > > > > > > > > than > > > > > > > > > > > > > introduce > > > > > > > > -- Best, Stanislav