Yes, I think if an implementation can throw an exception in this case,
we should add that as an @throws. For example, something like the following:
@throws ProviderException if this signature engine requires parameters
but does not support default or randomly generated parameter values
On 7/10/18 10:16 PM, Weijun Wang wrote:
About "it *may* return", do you mean it could also return null? My
understanding is no.
Is it better to clarify when the implementation "may also fail"? From the CSR,
it's this method. Can you add a @throws spec to this method then?
Also, I am a little confused by "default and randomly generated". Does this actually mean "default (might be randomly
generated)"? The "it may" sentence mentions "default and randomly generated" but the "if there" only says
"default", which sounds like the the "randomly generated" case could be different.
On Jul 11, 2018, at 5:12 AM, Valerie Peng <valerie.p...@oracle.com> wrote:
Would you have time to review the fix for JDK-8206171: Signature#getParameters
for RSASSA-PSS throws ProviderException when not initialized?
No source code changes, but just updating javadoc to mention the possible
Otherwise, JCK team expects a parameter object or null being returned.
I filed a CSR to track the javadoc clarification.