2 issues with hs25-b37

[Lindenmaier, Goetz]

Hi Alan,

you were right, with a later jdk it works again.

Yes, we are using a jdk8 as boot jdk. But there won't
be a problem, I checked in the fix I proposed to 
the jdk8 repo (not the staging, naturally).

Thanks for the help,
  Goetz.



-----Original Message-----
From: Alan Bateman [mailto:Alan.Bateman at oracle.com] 
Sent: Monday, June 24, 2013 8:13 PM
To: Volker Simonis
Cc: Lindenmaier, Goetz; hotspot-dev at openjdk.java.net
Subject: Re: 2 issues with hs25-b37

On 24/06/2013 16:56, Volker Simonis wrote:
> :
> Do you mean "8015630 : Remove default restriction settings of jaxp 1.5
> properties in JDK8"
> (http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=8015630)?
>
> I think that will indeed fix the problem. It's just the chicken/egg
> problem when bootstrapping a port. We'll first have to use an older
> JDK to build a new version which can then be used as bootstrap-JDK
> again - but I think we can cope with that:)
>
Yes, 8015630 sorted out the compatibility issue in JAXP and the only 
remaining (and deliberate) difference to long standing behavior will be 
when feature secure processing is enabled (which you have to opt-in).

I'm not sure that I understand the chicken & egg situation that you are 
running into. Normally the boot JDK for jdk8 is a jdk7 build. It sounds 
like you might be using a jdk8 build as the boot JDK but that is only 
guaranteed to work if the matches exactly the jdk8 that you are 
building. When you do a bootcycle-images build then it will of course be 
guarantee to match because the newly built JDK will be used to build it 
(so they will match).

-Alan.