On Wed, 15 Apr 2026 08:45:29 GMT, Severin Gehwolf <[email protected]> wrote:

>>> In a jmod-less JDK, we can propagate the `java.security` changes by passing 
>>> `--ignore-modified-runtime` to `jlink`, getting a warning (expand to see an 
>>> example).
>> 
>> Interesting - this option seems to be undocumented and not mentioned in JEP 
>> 493. Is there a reason? 
>> 
>> Your proposal seems more prone to make mistakes where you inadvertently 
>> include modifications from the configuration file in the JDK binary -- for 
>> example perhaps you were testing and made some temporary changes to the 
>> `java.security` file and forgot to change it back, unless you do a diff 
>> against the original you may not notice them. Using the `jlink 
>> --security-properties <filename>` option is a more specific action where 
>> nothing is happening behind the scenes w/o your knowledge, and there is a 
>> file that you can specifically audit to see the exact properties that are 
>> being modified. If you really want to do this, I think you need to modify 
>> the original source file and do your own build of the JDK.
>
>> > In a jmod-less JDK, we can propagate the `java.security` changes by 
>> > passing `--ignore-modified-runtime` to `jlink`, getting a warning (expand 
>> > to see an example).
>> 
>> Interesting - this option seems to be undocumented and not mentioned in JEP 
>> 493. Is there a reason?
> 
> Yes, as you say it's not recommended to use this option. It should stay that 
> way. Since it was not clear what kind of post-processing JDK vendors do ahead 
> of the JEP it's been a useful option.

So far the proposal to support includes seems a reasonable compromise:


--security-properties 
props=<filename>:include=/etc/crypto-policies/back-ends/java.config

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

PR Review Comment: https://git.openjdk.org/jdk/pull/30635#discussion_r3085184031

Reply via email to