8225543: Jcmd fails to attach to the Java process on Linux using the main class name if whitespace options were used to launch the process

[David Holmes]

Hi Daniil,

These changes seem quite reasonable. In hindsight splitting the 
command-line based on spaces was bound to hit problems with arguments 
containing spaces. :(

Thanks for fixing.

David

On 13/06/2019 3:50 am, Daniil Titov wrote:
> Hi David and Serguei,
> 
> Could you please have a look at this change? This fix is related with the changes [3] that you reviewed back in February.
> 
> Mach5 sun/tools/jcmd, serviceability/dcmd/framework, jdk_svc, hotspot_serviceability, tier1, tier2 and tier3 tests passed.
> 
> Thanks a lot!
> --Daniil
> 
> On 6/10/19, 4:56 PM, "serviceability-dev on behalf of Daniil Titov" <serviceability-dev-bounces at openjdk.java.net on behalf of daniil.x.titov at oracle.com> wrote:
> 
>      Please review the change that fixes jcmd process name matching on Linux platform.
>      
>      After changes [3] , on Linux platform the proc filesystem is used to find a Java process, however, sun.tools.ProcessHelper  class,
>      introduced in that change, didn't take into account the presence of module whitespace options ( e.g. --patch-module) or the
>      fact that some values of Java options could contain whitespaces.
>      
>      The latter problem is fixed by keeping '\0' as a separator for command line arguments read from /proc/{pid}/cmdline rather than
>      replacing it with a whitespace. The former problem is fixed by making sun.tools.ProcessHelper  aware of these module whitespace
>       options (--add-opens, --add-exports,--add-reads, --add-modules, --patch-module,--limit-modules, or --upgrade-module-path) as well
>       as of the case when a source-file mode is used or when --module options is used in "--<name>=<value>" format.
>      
>      Testing: sun/tools/jcmd and serviceability/dcmd/framework tests succeeded in Mach5. jdk_svc, tier1, tier2, and tier3 tests are in progress.
>      
>      
>      [1] Webrev: http://cr.openjdk.java.net/~dtitov/8225543/webrev.01/
>      [2] Bug: https://bugs.openjdk.java.net/browse/JDK-8225543
>      [3]  https://bugs.openjdk.java.net/browse/JDK-8205654
>      
>      Thanks!
>      --Daniil
>      
>      
>      
>      
> 
>