Progress on 94

[Lewis John Mcgibbney <le...@gmail.com>]

Hi,
Just a quick note to say I've been hacking away on GORA-94.
Thanks to the patch uploaded by Yasin Tamer + some other cleaning up and
rejigging of testing classes due to new persistency model, we are nearly in
a position to commit the fix for this long overdue issue.
If we keep this thread alive for on list discussion then that would be
great.
I think, that before we push an RC for 0.4, we'll need to look into the
diff's between the 94 branch and trunk as some patches may need to be
ported prior to the merge.
I'll update this thread once I get all tests passing for my local 94 branch.
Thanks
Lewis

-- 
*Lewis*

Re: Progress on 94

[Renato Marroquín Mogrovejo <re...@gmail.com>]

Great work as usual Lewis! Let's try to clear this one out ;)


Renato M.


2013/10/24 Lewis John Mcgibbney <le...@gmail.com>

> Hi Team,
> OK, as things stands the 94 build is now building fine [0] with the
> exception of one test, namely
>
>
> org.apache.gora.mapreduce.TestPersistentSerialization.testSerdeMultipleWebPages
> in gora-core
>
> Having debugged this locally, I see that things go to the wall within
> TestIOUtils.testSerializeDeserialize(T... objects) where we pass three
> WebPage objects with different fields set to dirty. The code within this
> method which seems to be the source of the problem is
>
> IOUtils.serialize(conf, dos , before, (Class<T>)before.getClass());
>
> The output from the test states "null of map in field outlinks of
> org.apache.gora.examples.generated.WebPage" and I can confirm that this
> happens when we try to serialize the first WebPage object passed to the
> method.
>
> So is there a way we can work around null values for certain fields before
> (de)serializing?
>
> Once we get this last test sorted it's a free for all to get the other
> modules stable. It's important to note that the overwhleming degree of work
> is now behind us if we get the gora-core module stable.
>
> Thanks
>
> [0] https://builds.apache.org/job/GORA_94_BRANCH/114/
>
>
> On Wed, Oct 23, 2013 at 10:26 PM, Lewis John Mcgibbney <
> lewis.mcgibbney@gmail.com> wrote:
>
> > Hi,
> > Just a quick note to say I've been hacking away on GORA-94.
> > Thanks to the patch uploaded by Yasin Tamer + some other cleaning up and
> > rejigging of testing classes due to new persistency model, we are nearly
> in
> > a position to commit the fix for this long overdue issue.
> > If we keep this thread alive for on list discussion then that would be
> > great.
> > I think, that before we push an RC for 0.4, we'll need to look into the
> > diff's between the 94 branch and trunk as some patches may need to be
> > ported prior to the merge.
> > I'll update this thread once I get all tests passing for my local 94
> > branch.
> > Thanks
> > Lewis
> >
> > --
> > *Lewis*
> >
>
>
>
> --
> *Lewis*
>

Re: Progress on 94

[Lewis John Mcgibbney <le...@gmail.com>]

Hi Team,
OK, as things stands the 94 build is now building fine [0] with the
exception of one test, namely

org.apache.gora.mapreduce.TestPersistentSerialization.testSerdeMultipleWebPages
in gora-core

Having debugged this locally, I see that things go to the wall within
TestIOUtils.testSerializeDeserialize(T... objects) where we pass three
WebPage objects with different fields set to dirty. The code within this
method which seems to be the source of the problem is

IOUtils.serialize(conf, dos , before, (Class<T>)before.getClass());

The output from the test states "null of map in field outlinks of
org.apache.gora.examples.generated.WebPage" and I can confirm that this
happens when we try to serialize the first WebPage object passed to the
method.

So is there a way we can work around null values for certain fields before
(de)serializing?

Once we get this last test sorted it's a free for all to get the other
modules stable. It's important to note that the overwhleming degree of work
is now behind us if we get the gora-core module stable.

Thanks

[0] https://builds.apache.org/job/GORA_94_BRANCH/114/


On Wed, Oct 23, 2013 at 10:26 PM, Lewis John Mcgibbney <
lewis.mcgibbney@gmail.com> wrote:

> Hi,
> Just a quick note to say I've been hacking away on GORA-94.
> Thanks to the patch uploaded by Yasin Tamer + some other cleaning up and
> rejigging of testing classes due to new persistency model, we are nearly in
> a position to commit the fix for this long overdue issue.
> If we keep this thread alive for on list discussion then that would be
> great.
> I think, that before we push an RC for 0.4, we'll need to look into the
> diff's between the 94 branch and trunk as some patches may need to be
> ported prior to the merge.
> I'll update this thread once I get all tests passing for my local 94
> branch.
> Thanks
> Lewis
>
> --
> *Lewis*
>



-- 
*Lewis*