[Mono-dev] Sys.SM.Configuration changes caused too much regressions

[Noam Lampert]

Hi Atsushi,

Thank you for these constructive comments. I personally was not aware
that the build/test status was available.

I have reviewed with Igor the recent changes.
The current failures are new tests. These failures are not related to
the original problem you reported below. However, it would have been
better to commit these tests with [NotWorking] attribute.
Igor will commit by tomorrow a clean status in which either the tests
are working, or they are marked with [NotWorking].

We will also try to maintain a cleaner test status of the committed
code, and will pay closer attention to the ChangeLog comments.

Noam

-----Original Message-----
From: Igor Zelmanovich 
Sent: Monday, March 24, 2008 7:46 PM
To: Atsushi Eno
Cc: mono-devel-list at lists.ximian.com; Noam Lampert; Eyal Alaluf
Subject: RE: Sys.SM.Configuration changes caused too much regressions

You refer to tests I added today.
It is not regression, just new tests,
And I am working on it right now.

Please let me finish my work.

I make a lot changes in configuration stuff, but it not suppose to
affect other aspects of current implementation.


Igor.




-----Original Message-----
From: Atsushi Eno [mailto:atsushi at ximian.com] 
Sent: Monday, March 24, 2008 7:33 PM
To: Igor Zelmanovich
Cc: mono-devel-list at lists.ximian.com; Noam Lampert; Eyal Alaluf
Subject: Re: Sys.SM.Configuration changes caused too much regressions

No matter how it works in your machine, the tests are indeed broken
as monobuild shows. See System.ServiceModel test results:
http://mono.ximian.com/monobuild/python/monobuild.py
http://mono.ximian.com/monobuild/builds/HEAD/suse-101-i586/olive/98880/l
ogs/test.log


BTW you always have meaningless ChangeLog entry such as:

> Log:
> fix implementation, refactoring

but it is not true. For example for r98854 which you put *new file*
which is neither implementation fix nor refactoring at all.

Atsushi Eno


Igor Zelmanovich wrote:
> Hello Atsushi,
> 
> There was some not implemented API in System.Configuration
> I've implemented it in r98680
> 
> You have rebuild (and install) System.Configuration to have the tests
> pass.
> 
> I have no such regressions in my Ubuntu.
> 
> If you still have any such regression, please send me exception's
stack
> trace.
> Be sure I run nunit test suite each time I have commit.
> 
> Igor Zelmanovich.
> 
> -----Original Message-----
> From: Atsushi Eno [mailto:atsushi at ximian.com] 
> Sent: Friday, March 21, 2008 1:50 PM
> To: Igor Zelmanovich
> Cc: mono-devel-list at lists.ximian.com; mono-olive at googlegroups.com
> Subject: Sys.SM.Configuration changes caused too much regressions
> 
> Hello Igor,
> 
> I usually don't care too much about NUnit test regressions in the
> face of significant development in olive, but lately you broke too
> much tests due to System.ServiceModel.Configuration improvements.
> More than 10 tests now fails with NotImplementedException which
> is raised inside configuration stuff.
> 
> Please remove any configuration hook from existing working stuff
> such as ChannelFactory. If you do not, I'll revert several things
> instead next week, which could mess your work. Thanks in advance.
> 
> Atsushi Eno
> 

 

__________ Information from ESET NOD32 Antivirus, version of virus
signature database 2740 (20071221) __________

The message was checked by ESET NOD32 Antivirus.

http://www.eset.com