[OpenJDK 2D-Dev] Trying to understand changeset 1010:467e4f25965c for lcms

[Phil Race]

Andrew Haley wrote:
> Phil Race wrote:
>> Andrew Haley wrote:
>>> Phil Race wrote:
>>>> I recall that we refactored the patch to touch fewer of the littlecms
>>>> internals. This helped since just 4 days after that patch the littlecms
>>>> version was upgraded from 1.16 to 1.18. I'd be surprised if you have
>>>> only this patch and not the littlecms 1.18 patch. Maybe its 1.18
>>>> rather than this changeset that's the root of the problem.
>>>>
>>>> FYI, the code in 6-open and jdk7 should now, I think, be identical.
>>> I don't know where you might have got the idea that I don't have the
>>> littlecms 1.18 patch; I certainly do!
>>
>> Since you mentioned a changeset which was since completely blown
>> away by a subsequent one ..
> 
> Oh, I see.  Perhaps I misread the logs.

Nonetheless, looking at the patch you were right that in identifying the
changeset that caused this, since the 1.18 upgrade barely touched the file
in question since its not part of LittleCMS - its our glue code to LittleCMS,
it was indeed already broken by the previous changeset.

> 
>>> OK, so I'll open a bug for a JCK failure.  Is it sufficient simply
>>> to cite the test and the error message?  I have a webrev that I'll
>>> send as soon as there is a bugzilla entry to attach it to.
> 
> The bug is at https://bugs.openjdk.java.net/show_bug.cgi?id=100050, with
> a suggested patch attached to it.  Please have a look and let me know if
> I can push the patch to 6-open and jdk7.  If you need webrevs as well as
> the patch I can provide them.

I see it, although I was thrown at first. since it looks like
a patch to go from your fixed version back to the current state.
[Ah .. I think I just see that you uploaded a corrected version].
I'll give it a whirl in a bit.
Jennifer is coordinating getting webrevs archived and getting
reviews properly documented. She'll pick this up and you can
coordinate how to push with her.

-phil.

> 
> Andrew.