Please review: 7194005: (launcher) needs to be enhanced for 64-bit jar file handling (original) (raw)

Kumar Srinivasan kumar.x.srinivasan at oracle.com
Tue Sep 4 18:34:10 UTC 2012

Hi Sherman,

Thanks for the review, I will make that test conditional.

As for using the entry points I looked at it, the issue is that java_util_zip reading code is dependent on jni and the jvm loaded, but here in the launcher we are still early in the launch process and we do not yet have a functional jvm running.

So definitely your suggestion is a good future project and we may have to refactor the zip code to be jvm neutral.

Thanks

Kumar

----- xueming.shen at oracle.com wrote:

From: xueming.shen at oracle.com To: kumar.x.srinivasan at oracle.com Cc: joe.darcy at oracle.com, core-libs-dev at openjdk.java.net, james.holmlund at oracle.com Sent: Tuesday, September 4, 2012 10:41:44 AM GMT -08:00 US/Canada Pacific Subject: Re: Please review: 7194005: (launcher) needs to be enhanced for 64-bit jar file handling

Kumar, The change looks fine with me. Agreed that you definitely don't want to run the auto test for a > 4G file. The >4G test in zip/jar area was checked in as a manual test. Just wonder you might want to consider to use (statically include, like you do for those zlib code) the ziputil/ZIPOpen/FineEntry/ReadEntry instead of having your own zip format handling code in launcher (source). Not necessary for this patch. -Sherman On 09/04/2012 10:05 AM, Kumar Srinivasan wrote:

Hi Joe,

I was thinking about that too, yes I will make it conditional on a system property or env variable. Kumar ----- joe.darcy at oracle.com wrote: > From: joe.darcy at oracle.com > To: kumar.x.srinivasan at oracle.com > Cc: core-libs-dev at openjdk.java.net , xueming.shen at oracle.com , james.holmlund at oracle.com > Sent: Tuesday, September 4, 2012 9:50:03 AM GMT -08:00 US/Canada Pacific > Subject: Re: Please review: 7194005: (launcher) needs to be enhanced for 64-bit jar file handling > > Hi Kumar, > > From a quick look I didn't seen any problems, but defaulting to running a test that writes multiple gibabytes to disk may have unwanted performance costs for general testing. Perhaps that scenario could be included in the test code, but only run on an opt-in basis? > > -Joe > > > On 9/4/2012 6:47 AM, Kumar Srinivasan wrote: >

> Hi Sherman, Joe, et. al. > > Please review the launcher changes to read the manifest of > a ZIP64 archive. > > http://cr.openjdk.java.net/~ksrini/7194005/webrev.0/ > > Thanks > Kumar > > > >

More information about the core-libs-dev mailing list