Illegal reflection access denial in which future release

[Remi Forax]

----- Mail original -----
> De: "Claes Redestad" <claes.redestad at oracle.com>
> À: "S A" <simeon.danailov.andreev at gmail.com>, "jigsaw-dev" <jigsaw-dev at openjdk.java.net>
> Cc: "Andrey Loskutov" <loskutov at gmx.de>
> Envoyé: Lundi 8 Juin 2020 16:09:28
> Objet: Re: Illegal reflection access denial in which future release

> Hi Simeon,
> 
> it's not documented or discussed much anywhere because it's as-of-yet
> undecided.
> 
> *Expectation* is that there'll be a JEP proposed to make
> --illegal-access=deny the *default* in JDK 16. I expect the ability to
> explicitly permit illegal accesses with --illegal-access=permit will
> stick around for a good while longer, though.

As a member of the expert group on JPMS, i think we should move to deny by default for 16.
A lot of people are now migrating their applications to 11, i think it's cheaper to do the migration to a world with --illegal-access=deny than to do first a migration to 11 and then a migration to 17.

> 
> Best regards
> /Claes

regards,
Rémi

> 
> On 2020-06-08 15:40, S A wrote:
>> Hi all,
>> 
>> TL;DR is a future JDK release known, when illegal reflection access is no
>> longer possible? If not, are there estimations / discussions / FAQs /
>> documentation?
>> 
>> The product I work on (Eclipse+xtext IDE + backend) has nearly completed
>> the move to OpenJDK 11 (from OpenJDK 8) and we are working on supposedly
>> not pressing issues. One of those issues is illegal access via reflection.
>> 
>> We would obviously like to be prepared and so are estimating effort and
>> time constraints of fixing illegal reflection access. As this includes
>> multiple libraries, knowing *when* we need to finish our efforts is quite
>> important.
>> 
>> Sorry if this has been asked before (I imagine yes), but searching for the
>> topic yields lots of how to work around or fix the issue. Thank you in
>> advance.
>> 
>> Best regards,
>> Simeon