Current state of Jira and ".net-ification"

[Simon Svensson <si...@devhost.se>]

Hi,

I'm looking at some Jira issues, focusing on the bugs affecting 4.8.

LUCENENET-605 is asking to implement ArgumentNullException instead of
letting null values become NullReferenceException. The current behavior
matches the Java behavior.

LUCENENET-600 is about the number of first-chance exceptions. This
sounds like another .NET-ification.

About the .NET-ification; is this something we're doing for 4.8?

LUCENENET-598 is the same as LUCENE-6001, which was fixed in Lucene
4.10. Am I correct in thinking that we do not fix this issue since the
same problem exists in Lucene's 4.8?

There are three unreleased versions in Jira which I believe we can
archive; 3.6, 4.0 and 5.0 PCL. Archiving these means that they can no
longer be selected when new bugs are reported. Do we still have any
purpose for having these selectable when new issues are filed?

-- 
// Simon Svensson

Re: Current state of Jira and ".net-ification"

[Simon Svensson <si...@devhost.se>]

Hi,

LUCENENET-598 is reproducible with our current code. There's a failing
unit test provided in the issue, and it is fixed by the patch in
LUCENE-6001.

I'll write a comment in the issue and leave it open since it is a known
bug and have a known fixVersion. (However, our project doesn't have that
fixVersion yet, so we cannot properly tag it as such in Jira.)
Regarding the permissions; I'm not sure how they are configured. I have
permissions to resolve/close issues, but not much more.

// Simon Svensson

On 2019-01-19 19:12, Laimonas Simutis wrote:
> Simon,
> 
> Please see my thoughts below.
> 
> 
>> LUCENENET-605 is asking to implement ArgumentNullException instead of
>> letting null values become NullReferenceException. The current behavior
>> matches the Java behavior.
>>
> 
> We should close it as won't fix. I was actually going to do that earlier
> today but I do not have access to JIRA to close or modify in any way
> lucene.net issues. Our goal for 4.8 should be to stay as close to the Java
> version as possible and as you said, Lucene has the same behavior, it
> throws null exception. That should suffice as a good won't fix reason.
> 
> 
>>
>> LUCENENET-600 is about the number of first-chance exceptions. This
>> sounds like another .NET-ification.
> 
> 
>> About the .NET-ification; is this something we're doing for 4.8?
>>
> 
> Same as for LUCENENET-605, it should be left alone and not done for 4.8
> release. I know having gone through fixing hard to track down bugs, if we
> start stearing away from the original code base it will be hard to track
> down where the problem is. Let's avoid it. I know that there are
> performance issues due to exceptions, but again, my vote would be to
> address that separately once 4.8 is released.
> 
> 
>>
>> LUCENENET-598 is the same as LUCENE-6001, which was fixed in Lucene
>> 4.10. Am I correct in thinking that we do not fix this issue since the
>> same problem exists in Lucene's 4.8?
>>
> 
> I wonder if somehow we still have that issue in our current 4.8 lucene.net
> code base? Were you able to reproduce it based on the example the user
> provided?
> 
> 
>>
>> There are three unreleased versions in Jira which I believe we can
>> archive; 3.6, 4.0 and 5.0 PCL. Archiving these means that they can no
>> longer be selected when new bugs are reported. Do we still have any
>> purpose for having these selectable when new issues are filed?
>>
> 
> Have no opinion there.
> 
> Laimonas
> 

Re: Current state of Jira and ".net-ification"

[Laimonas Simutis <la...@gmail.com>]

Simon,

Please see my thoughts below.


> LUCENENET-605 is asking to implement ArgumentNullException instead of
> letting null values become NullReferenceException. The current behavior
> matches the Java behavior.
>

We should close it as won't fix. I was actually going to do that earlier
today but I do not have access to JIRA to close or modify in any way
lucene.net issues. Our goal for 4.8 should be to stay as close to the Java
version as possible and as you said, Lucene has the same behavior, it
throws null exception. That should suffice as a good won't fix reason.


>
> LUCENENET-600 is about the number of first-chance exceptions. This
> sounds like another .NET-ification.


> About the .NET-ification; is this something we're doing for 4.8?
>

Same as for LUCENENET-605, it should be left alone and not done for 4.8
release. I know having gone through fixing hard to track down bugs, if we
start stearing away from the original code base it will be hard to track
down where the problem is. Let's avoid it. I know that there are
performance issues due to exceptions, but again, my vote would be to
address that separately once 4.8 is released.


>
> LUCENENET-598 is the same as LUCENE-6001, which was fixed in Lucene
> 4.10. Am I correct in thinking that we do not fix this issue since the
> same problem exists in Lucene's 4.8?
>

I wonder if somehow we still have that issue in our current 4.8 lucene.net
code base? Were you able to reproduce it based on the example the user
provided?


>
> There are three unreleased versions in Jira which I believe we can
> archive; 3.6, 4.0 and 5.0 PCL. Archiving these means that they can no
> longer be selected when new bugs are reported. Do we still have any
> purpose for having these selectable when new issues are filed?
>

Have no opinion there.

Laimonas