RFR: JDK-8319516 - Native library suffix impact on the library loading in AIX- Java Class Loader [v5]

Suchismith Roy sroy at openjdk.org
Wed Mar 27 17:23:50 UTC 2024


On Wed, 27 Mar 2024 17:19:31 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>>> We are not supporting that. Are you referring to the comment in the code ? Yeah it should be resconstruction of libname(member_name).so , which is the first filename the classLoader constructs.
>> 
>> Note that `System.mapLibraryName` and `mapAlternativeName` are called for `System.loadLibrary`  (i.e. prepending `lib` and appending `.so`).    "libname(member_name)" is not a valid name and no reason for `System.loadLibrary`  to support it.
>
>> I think we both mean that the if (name.contains("(")) block should get removed.
> 
> Because of the VM support, we can remove `mapAlternativeName` completely.

Oh ok.Yes i have removed that now. You mean ,load library wont support it, but load will be able to support it since the loadLibraryIfPresent is returned false now. 

Now the loadLibrary can only handle .so to .a mapping due to mapAlternativeName

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17945#discussion_r1541534243


More information about the core-libs-dev mailing list