Re: [PROPOSAL] close GORA-53 and track 3 bugs in subtasks of a new issue

2012-07-18 Thread Lewis John Mcgibbney
Hi Kaz,

On Mon, Jul 16, 2012 at 2:55 PM, Kazuomi Kashii kazu...@kashii.net wrote:
 I think that gora-cassandra test is working now, even though there are
 several issues as attached.
 Considering that the objective of GORA-53 is to add (or implement)
 gora-cassandra test, and
 that it is difficult to track many issues in a single issue,
 I'd propose to:
 1) commit GORA-53 (and GORA-149) patch(es);

+1 for GORA-53, GORA-149 is pretty meaty and I've only had limited
oppertunity to try it out so far. I'd like to run through some mock up
situations then run with Nutchgora a couple of times to see that all
is well. AFAICT the new serializers look great in gora-cassandra and a
good improvement. It will be great to eventually integrate into
codebase.

 2) close GORA-53;

+1. Based on the above

 3) open a new issue that gora-cassandra test fails; and

Yes we will need to override these tests in the core module. As I've
been away for the weekend I wonder if you managed to get a grasp on
what is wrong e.g. whether there are bugs in gora-core, gora-cassandra
or that the tests simply need to be overridden in gora-cassandra? I do
not know this just now and would need to look into them before I have
an idea?

 4) open 3 subtasks of a new bug;

Well this seems to have partly answered my question in 3 above. I
remember seeing somewhere that you think there is a bug in core?  gain
I need to look further into the log output and debug before I will
know any more.

In summary... are you able to commit 53  149? I've had a look at 149
and it looks A OK to me. Although I have no overpowering concerns
about committing into codebase and getting others using it I think it
would be excellent if you could please provide some clarification on
the above comments? Thank you in advance Kaz.

Best
Lewis


Re: [PROPOSAL] close GORA-53 and track 3 bugs in subtasks of a new issue

2012-07-18 Thread Kazuomi Kashii
Hi Lewis,

Thanks for your comments, and
I will be committing GORA-149 and GORA-53 patches shortly.

For the current failures of gora-cassandra test, I know 3 as described at:
https://issues.apache.org/jira/browse/GORA-53?focusedCommentId=13414694#comment-13414694

1) AssertionFailedError is simple to fix at CassandraStore.

3) ComparisonFailure is a bug of gora-cassandra, and to be investigated
further.

2) StackOverflowError is caused between ListGenericArray.java of
gora-core and GenericData.java of Avro as follows.

testDeleteByQuery(org.apache.gora.cassandra.store.TestCassandraStore)
Time elapsed: 0.62 sec  ERROR! 
java.lang.StackOverflowError 
at
org.apache.gora.persistency.ListGenericArray.compareTo(ListGenericArray.java:32)
 
at
org.apache.avro.generic.GenericData.compare(GenericData.java:487) 
at
org.apache.gora.persistency.ListGenericArray.compareTo(ListGenericArray.java:102)
 
at
org.apache.gora.persistency.ListGenericArray.compareTo(ListGenericArray.java:32)
 

Currently, I have no clue.
It looks a bug of ListGenericArray, but another gora-cassandra bug might
cause this.
I will investigate it further.

Thanks,
-Kaz


On 7/18/12 3:35 AM, Lewis John Mcgibbney wrote:
 Hi Kaz,

 On Mon, Jul 16, 2012 at 2:55 PM, Kazuomi Kashii kazu...@kashii.net wrote:
 I think that gora-cassandra test is working now, even though there are
 several issues as attached.
 Considering that the objective of GORA-53 is to add (or implement)
 gora-cassandra test, and
 that it is difficult to track many issues in a single issue,
 I'd propose to:
 1) commit GORA-53 (and GORA-149) patch(es);
 +1 for GORA-53, GORA-149 is pretty meaty and I've only had limited
 oppertunity to try it out so far. I'd like to run through some mock up
 situations then run with Nutchgora a couple of times to see that all
 is well. AFAICT the new serializers look great in gora-cassandra and a
 good improvement. It will be great to eventually integrate into
 codebase.

 2) close GORA-53;
 +1. Based on the above

 3) open a new issue that gora-cassandra test fails; and
 Yes we will need to override these tests in the core module. As I've
 been away for the weekend I wonder if you managed to get a grasp on
 what is wrong e.g. whether there are bugs in gora-core, gora-cassandra
 or that the tests simply need to be overridden in gora-cassandra? I do
 not know this just now and would need to look into them before I have
 an idea?

 4) open 3 subtasks of a new bug;
 Well this seems to have partly answered my question in 3 above. I
 remember seeing somewhere that you think there is a bug in core?  gain
 I need to look further into the log output and debug before I will
 know any more.

 In summary... are you able to commit 53  149? I've had a look at 149
 and it looks A OK to me. Although I have no overpowering concerns
 about committing into codebase and getting others using it I think it
 would be excellent if you could please provide some clarification on
 the above comments? Thank you in advance Kaz.

 Best
 Lewis