RFR: 8347067: Load certificates without explicit trust settings in KeyChainStore [v2]

Tim Jacomb duke at openjdk.org
Sun Jan 26 23:01:36 UTC 2025


On Sat, 25 Jan 2025 01:10:41 GMT, Alexey Bakhtin <abakhtin at openjdk.org> wrote:

>> Tim Jacomb has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains 13 additional commits since the last revision:
>> 
>>  - Add non-trusted root CA cert
>>  - Merge branch 'master' into load-anchor-and-user-certificates-keychainstore
>>  - Executable files are not allowed...
>>  - Flag test as manual
>>  - Minor cleanups
>>  - Add new line
>>  - Add jtreg test
>>  - Release subjCerts
>>  - Revert unneeded changes
>>  - Merge branch 'master' into load-anchor-and-user-certificates-keychainstore
>>  - ... and 3 more: https://git.openjdk.org/jdk/compare/6ca5c5dd...d9605e12
>
> test/jdk/java/security/KeyStore/CheckMacOSKeyChainIntermediateCATrust.java line 166:
> 
>> 164:     private static void assertThat(boolean expected, String message, List<X509Certificate> certificates) {
>> 165:         if (!expected) {
>> 166:             throw new AssertionError(message + ", subjects: " + getSubjects(certificates));
> 
> I do not like printing all KeyChain certificates on the failure. It could be sensitive information.
> 
> If you do not collect all certificates, the test could be simplified - without Stream API

Without this its harder to debug what went wrong, this tells you what subjects are found, its not the certificate itself, just e.g. 


java.lang.AssertionError: Non trusted CA not found CN=Non Trusted Example CA,O=Example,C=US, subjects: [ CN=StartCom Certification Authority,OU=Secure Digital Certificate Signing,O=StartCom Ltd.,C=IL, CN=TIMJA-INTERMEDIATE,O=TIMJA,ST=ES,C=UK, CN=TIMJA-INTERMEDIATE-2,O=TIMJA,ST=ES,C=UK, CN=TIMJA-ROOT,O=TIMJA,ST=ES]

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/22911#discussion_r1929888919


More information about the security-dev mailing list