RFR: 8020779 & 8026988 : Improvements to JTReg executable location handling

[Mike Duigou]

On Nov 7 2013, at 01:24 , Erik Joelsson <erik.joelsson at oracle.com> wrote:

> Hello Mike,
> 
> I think this looks good. A couple of comments that don't necessarily call for any changes.
> 
> The version check is commented out, when is it planned to be activated?

I was working on this piece at the same time as a the "make version 4.0" discussion was happening and I wanted to wait for the solution used there and reapply it. I can complete this code before commit.

> 
> I see you use 'which' as a fallback in the makefiles. Just be aware that we have had some trouble using the which command in the past. Sometimes on Solaris it gives unexpected output depending on your environment and the behavior on fail is different on various platforms and implementations. It looks like you are trying to handle all the weird cases though.

The pattern I used is copied from hgforest.sh where the problems were indeed previously encountered. I would prefer to have no fallback at all and plan to eventually remove it in a future changeset. The current fallback behaviour is intended to help with the transition from the old build system to the new build infrastructure and the variety of usage patterns for test/Makefile which are still being used.

It will be a week or two before I try to finalize this changeset.

Thanks,

Mike

> /Erik
> 
> On 2013-11-06 21:40, Mike Duigou wrote:
>> Hello all;
>> 
>> With JDK-8015068 only very recently pushed it's time for the next set of jtreg test/Makfile changes. These changes improve the configure script detection of jtreg and improve the handling of the specification variables in the jdk and langtools test/Makefiles. This changeset is also the first step in syncing the two test/Makefiles. More will follow which will further simplify the integration of testing into the overall build/test process.
>> 
>> http://cr.openjdk.java.net/~mduigou/JDK-8020779/0/webrev/
>> 
>> Mike
>