RFC: Forward port icedtea6's debug-dir patch to icedtea7

[Omair Majid]

On 04/28/2011 08:55 AM, Dr Andrew John Hughes wrote:
> On 10:17 Wed 27 Apr     , Omair Majid wrote:
>> On 04/27/2011 08:53 AM, Dr Andrew John Hughes wrote:
>>> On 12:48 Thu 21 Apr     , Omair Majid wrote:
>>>> Hi,
>>>>
>>>> The debug-dir patch from icedtea6 needs to be added to icedtea7. Without
>>>> this patch, a 'make icedtea-debug' creates a openjdk.build-debug dir
>>>> which is essentially empty, but places the build in actually (on my
>>>> macine) linux-amd64-debug under the current directory.
>>>>
>>>> The attached patch is the trivial port. Okay to commit?
>>>>
>>>> ChangeLog:
>>>> 2011-04-21  Omair Majid<omajid at redhat.com>
>>>>
>>>>      * Makefile.am (ICEDTEA_PATCHES): Add debug-dir.patch
>>>>      * patches/debug-dir.patch: New file. Forward port from icedtea6.
>>>>
>>>> Cheers,
>>>> Omair
>>>
>>> No.  This needs to go in the IcedTea forest at a minimum, not IcedTea7.
>>>
>>> Has this not been pushed upstream?
>>
>> I think this was originally a backport from upstream. But since then,
>> upstream has reverted the change:
>> http://hg.openjdk.java.net/jdk7/tl/diff/47f6b7db1882/Makefile
>>
>> So I am not sure if they will be willing to accept this patch. Perhaps
>> if it was special cased for non-windows platforms, they might be more
>> forthcoming.
>>
>
> Can you explain what the point of the patch is?  I don't even remember this
> going in.
>

The icedtea6 patch (debug-dir.patch) causes 'make icedtea-debug' to 
build openjdk in openjdk.build-debug. Without this patch 'make 
icedtea-debug' builds in linux-amd64-debug (or another appropriate name 
based on your platform/arch).

Upstream actually had this fix for a while, but revered it [1]. It was, 
presumably, causing problems with paths on windows. What's weird is that 
upstream's change breaks ALT_OUTPUTDIR for debug/fastdebug builds.

Thanks,
Omair

[1] http://hg.openjdk.java.net/jdk7/tl/diff/47f6b7db1882/Makefile