hg: jigsaw/jake/jdk: 8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module

2017-04-08 Thread mandy . chung
Changeset: b2160414fe8c Author:mchung Date: 2017-04-08 10:21 -0700 URL: http://hg.openjdk.java.net/jigsaw/jake/jdk/rev/b2160414fe8c 8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module Reviewed-by: alanb ! src/java.base/share/classes/java/lang/module

Re: Review Request JDK-8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module

2017-04-08 Thread Remi Forax
Ok, thanks. Rémi On April 8, 2017 4:49:13 PM GMT+02:00, Mandy Chung wrote: >No change in that regards. The updated test [1] in the webrev shows it >patches java.base/jdk.internal.modules.SystemModules > >Before this fix, you can do —patch-module jdk.internal.vm.compiler=.jar >—upgrade-module-

Re: Review Request JDK-8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module

2017-04-08 Thread Mandy Chung
> On Apr 8, 2017, at 12:21 AM, Alan Bateman wrote: > > On 08/04/2017 06:50, Mandy Chung wrote: > >> http://cr.openjdk.java.net/~mchung/jdk9/webrevs/8177844/webrev.01/ >> >> This fixes -—patch-module to do hash checking on the module being patched >> so that it will ensure that a non-upgradeabl

Re: Review Request JDK-8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module

2017-04-08 Thread Mandy Chung
No change in that regards. The updated test [1] in the webrev shows it patches java.base/jdk.internal.modules.SystemModules Before this fix, you can do —patch-module jdk.internal.vm.compiler=.jar —upgrade-module-path graal.jar. The reason is that —patch-module disables the hash checking that m

Re: Review Request JDK-8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module

2017-04-08 Thread Remi Forax
Hi Mandy, how can I test a change in java.lang after that patch ? Rémi On April 8, 2017 7:50:50 AM GMT+02:00, Mandy Chung wrote: >http://cr.openjdk.java.net/~mchung/jdk9/webrevs/8177844/webrev.01/ > >This fixes -—patch-module to do hash checking on the module being >patched >so that it will en

Aliasing (automatic) modules

2017-04-08 Thread Nicolai Parlog
Hi again, so far most (all?) proposals for handling automatic module names tackled the "what is the module's real name" aspect of this question but I have seen none that would push the likelihood of the same dependency being referenced by different names reasonably close to zero. As argued elsewh

Naming unmodularized modules

2017-04-08 Thread Nicolai Parlog
Hi again, one underlying problem of automatic modules is that maintainers have no way to express their desired module name without fully modularizing their artifacts. This will make "proper names" appear rather late in the migration. Brian's and Robert's idea to read Maven meta information addre

The case for dependencies on automatic modules

2017-04-08 Thread Nicolai Parlog
Hi! Automatic modules are a great feature because they allow decentralized and uncoordinated modularization. In an ecosystem as gigantic and diverse as Java's that is _the only_ tractable approach! Compared to that ability and in the big scheme of things, how an automatic module's name is chosen

Re: Review Request JDK-8177844: Ensure non-upgradeable modules cannot be upgraded even with --patch-module

2017-04-08 Thread Alan Bateman
On 08/04/2017 06:50, Mandy Chung wrote: http://cr.openjdk.java.net/~mchung/jdk9/webrevs/8177844/webrev.01/ This fixes -—patch-module to do hash checking on the module being patched so that it will ensure that a non-upgradeable module remains not upgradeable. The comment at L523 needs updatin