You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openjpa.apache.org by Mark Struberg <st...@yahoo.de> on 2010/05/03 14:13:44 UTC
providing integration tests for known bugs?
Hi!
It seems that I suffer from the same problems which got reported in OPENJPA-525 [1] while trying to write null value to a not null field.
It seems that OpenJPA magically converts it to an empty string, thus I get NO Exception in MySQL and I DO get an exception in Oracle (because Oracle treats empty VARCHAR strings as null).
Is there any integration test or unit test which proves OPENJPA-525?
If I'd like to write such a test, where should I add it?
txs and LieGrue,
strub
[1] https://issues.apache.org/jira/browse/OPENJPA-525
Re: AW: providing integration tests for known bugs?
Posted by Donald Woods <dw...@apache.org>.
Since both of those issues have been closed and marked as fixed in
release notes, you should open a new JIRA for your scenario with the
details and testcase to recreate.
-Donald
On 5/5/10 6:44 AM, Mark Struberg wrote:
> Hoi!
>
> I was now able to reproduce the issue.
> I've created an own small project and attached it to OPENJPA-1387 [1]. This also contains a sample for the issue which is described in OPENJPA-525 [2]
>
> Can anyone test this and show me a workaround (or reopen those 2 bugs)?
>
> txs and LieGrue,
> strub
>
> [1] https://issues.apache.org/jira/browse/OPENJPA-1387
> [2] https://issues.apache.org/jira/browse/OPENJPA-525
>
>
> --- Mark Struberg <st...@yahoo.de> schrieb am Di, 4.5.2010:
>
>> Von: Mark Struberg <st...@yahoo.de>
>> Betreff: AW: providing integration tests for known bugs?
>> An: dev@openjpa.apache.org
>> Datum: Dienstag, 4. Mai, 2010 14:18 Uhr
>> Hi Mike!
>>
>> I did poke that Test on 2.1.0-SNAPSHOT and had 2 findings.
>>
>> a) adding a @Column(nullable=false) and setting this field
>> null causes an Exception. This is ok from the test
>> perspective. However I need to retest this with my real
>> world app.
>>
>> b) I removed the nullable=false but added a
>>> @Table(name = "ANNOTEST1",
>>> uniqueConstraints =
>> @UniqueConstraint(name = "MyStringConstraint", columnNames =
>> {"myString"}))
>>
>> which did lead to the old known problem that myString is
>> mandatory again. This is pretty much the same as reported in
>> OPENJPA-1387 [1]
>>
>>
>>
>> [1] https://issues.apache.org/jira/browse/OPENJPA-1387
>>
>>
>>
>> ----- Ursprüngliche Mail ----
>>> Von: Michael Dick <mi...@gmail.com>
>>> An: dev@openjpa.apache.org
>>> Gesendet: Montag, den 3. Mai 2010, 17:20:11 Uhr
>>> Betreff: Re: providing integration tests for known
>> bugs?
>>>
>>> Hi Mark,
>>
>> The unit test for OPENJPA-525 should validate that it's
>> working.
>>> Could you
>> try running TestSerializedLobs.java and see if it fails for
>>
>>> you?
>>
>> FWIW the tests are running with MySQL 5.1.12 in the 2.0.x
>> branch and
>>> the
>> test passed..
>>
>> Regards,
>> -mike
>>
>> On Mon, May 3, 2010 at 7:13
>>> AM, Mark Struberg <
>>> href="mailto:struberg@yahoo.de">struberg@yahoo.de>
>> wrote:
>>
>>>
>>> Hi!
>>>
>>> It seems that I suffer from the same problems which
>> got
>>> reported in
>>> OPENJPA-525 [1] while trying to write null value to a
>> not
>>> null field.
>>> It seems that OpenJPA magically converts it to an
>> empty
>>> string, thus I get
>>> NO Exception in MySQL and I DO get an exception in
>>> Oracle (because Oracle
>>> treats empty VARCHAR strings as
>>> null).
>>>
>>> Is there any integration test or unit test which
>> proves
>>> OPENJPA-525?
>>>
>>> If I'd like to write such a test, where should I
>>> add it?
>>>
>>> txs and LieGrue,
>>> strub
>>>
>>>
>>>
>>> [1]
>>>> https://issues.apache.org/jira/browse/OPENJPA-525
>>>
>>>
>>>
>>>
>>
>>
>>
>
>
>
AW: providing integration tests for known bugs?
Posted by Mark Struberg <st...@yahoo.de>.
Hoi!
I was now able to reproduce the issue.
I've created an own small project and attached it to OPENJPA-1387 [1]. This also contains a sample for the issue which is described in OPENJPA-525 [2]
Can anyone test this and show me a workaround (or reopen those 2 bugs)?
txs and LieGrue,
strub
[1] https://issues.apache.org/jira/browse/OPENJPA-1387
[2] https://issues.apache.org/jira/browse/OPENJPA-525
--- Mark Struberg <st...@yahoo.de> schrieb am Di, 4.5.2010:
> Von: Mark Struberg <st...@yahoo.de>
> Betreff: AW: providing integration tests for known bugs?
> An: dev@openjpa.apache.org
> Datum: Dienstag, 4. Mai, 2010 14:18 Uhr
> Hi Mike!
>
> I did poke that Test on 2.1.0-SNAPSHOT and had 2 findings.
>
> a) adding a @Column(nullable=false) and setting this field
> null causes an Exception. This is ok from the test
> perspective. However I need to retest this with my real
> world app.
>
> b) I removed the nullable=false but added a
> > @Table(name = "ANNOTEST1",
> > uniqueConstraints =
> @UniqueConstraint(name = "MyStringConstraint", columnNames =
> {"myString"}))
>
> which did lead to the old known problem that myString is
> mandatory again. This is pretty much the same as reported in
> OPENJPA-1387 [1]
>
>
>
> [1] https://issues.apache.org/jira/browse/OPENJPA-1387
>
>
>
> ----- Ursprüngliche Mail ----
> > Von: Michael Dick <mi...@gmail.com>
> > An: dev@openjpa.apache.org
> > Gesendet: Montag, den 3. Mai 2010, 17:20:11 Uhr
> > Betreff: Re: providing integration tests for known
> bugs?
> >
> > Hi Mark,
>
> The unit test for OPENJPA-525 should validate that it's
> working.
> > Could you
> try running TestSerializedLobs.java and see if it fails for
>
> > you?
>
> FWIW the tests are running with MySQL 5.1.12 in the 2.0.x
> branch and
> > the
> test passed..
>
> Regards,
> -mike
>
> On Mon, May 3, 2010 at 7:13
> > AM, Mark Struberg <
> > href="mailto:struberg@yahoo.de">struberg@yahoo.de>
> wrote:
>
> >
> > Hi!
> >
> > It seems that I suffer from the same problems which
> got
> > reported in
> > OPENJPA-525 [1] while trying to write null value to a
> not
> > null field.
> > It seems that OpenJPA magically converts it to an
> empty
> > string, thus I get
> > NO Exception in MySQL and I DO get an exception in
> > Oracle (because Oracle
> > treats empty VARCHAR strings as
> > null).
> >
> > Is there any integration test or unit test which
> proves
> > OPENJPA-525?
> >
> > If I'd like to write such a test, where should I
> > add it?
> >
> > txs and LieGrue,
> > strub
> >
> >
> >
> > [1]
> > >https://issues.apache.org/jira/browse/OPENJPA-525
> >
> >
> >
> >
>
>
>
AW: providing integration tests for known bugs?
Posted by Mark Struberg <st...@yahoo.de>.
Hi Mike!
I did poke that Test on 2.1.0-SNAPSHOT and had 2 findings.
a) adding a @Column(nullable=false) and setting this field null causes an Exception. This is ok from the test perspective. However I need to retest this with my real world app.
b) I removed the nullable=false but added a
> @Table(name = "ANNOTEST1",
> uniqueConstraints = @UniqueConstraint(name = "MyStringConstraint", columnNames = {"myString"}))
which did lead to the old known problem that myString is mandatory again. This is pretty much the same as reported in OPENJPA-1387 [1]
[1] https://issues.apache.org/jira/browse/OPENJPA-1387
----- Ursprüngliche Mail ----
> Von: Michael Dick <mi...@gmail.com>
> An: dev@openjpa.apache.org
> Gesendet: Montag, den 3. Mai 2010, 17:20:11 Uhr
> Betreff: Re: providing integration tests for known bugs?
>
> Hi Mark,
The unit test for OPENJPA-525 should validate that it's working.
> Could you
try running TestSerializedLobs.java and see if it fails for
> you?
FWIW the tests are running with MySQL 5.1.12 in the 2.0.x branch and
> the
test passed..
Regards,
-mike
On Mon, May 3, 2010 at 7:13
> AM, Mark Struberg <
> href="mailto:struberg@yahoo.de">struberg@yahoo.de> wrote:
>
> Hi!
>
> It seems that I suffer from the same problems which got
> reported in
> OPENJPA-525 [1] while trying to write null value to a not
> null field.
> It seems that OpenJPA magically converts it to an empty
> string, thus I get
> NO Exception in MySQL and I DO get an exception in
> Oracle (because Oracle
> treats empty VARCHAR strings as
> null).
>
> Is there any integration test or unit test which proves
> OPENJPA-525?
>
> If I'd like to write such a test, where should I
> add it?
>
> txs and LieGrue,
> strub
>
>
>
> [1]
> >https://issues.apache.org/jira/browse/OPENJPA-525
>
>
>
>
Re: providing integration tests for known bugs?
Posted by Michael Dick <mi...@gmail.com>.
Hi Mark,
The unit test for OPENJPA-525 should validate that it's working. Could you
try running TestSerializedLobs.java and see if it fails for you?
FWIW the tests are running with MySQL 5.1.12 in the 2.0.x branch and the
test passed..
Regards,
-mike
On Mon, May 3, 2010 at 7:13 AM, Mark Struberg <st...@yahoo.de> wrote:
> Hi!
>
> It seems that I suffer from the same problems which got reported in
> OPENJPA-525 [1] while trying to write null value to a not null field.
> It seems that OpenJPA magically converts it to an empty string, thus I get
> NO Exception in MySQL and I DO get an exception in Oracle (because Oracle
> treats empty VARCHAR strings as null).
>
> Is there any integration test or unit test which proves OPENJPA-525?
>
> If I'd like to write such a test, where should I add it?
>
> txs and LieGrue,
> strub
>
>
> [1] https://issues.apache.org/jira/browse/OPENJPA-525
>
>
>
>