System for manual regression testing

[Maarten Bosma]

Hi,

I think ReactOS lacks of testers since WaxDragon is gone. So I had the 
following idea: A page which lists all regressions and gives you the 
option to report a working / non-working revision and calculates the 
range in which it the regression was caused from that data.

I know that bugzilla comments are actually sufficient for this and of 
course this system would have to be linked with bugzilla, but I think 
that a page where the regressions are listed up would be more motivating 
to help testing.

What do you think about it ?


Best regards,
	Maarten Bosma

[WaxDragon]

On 3/24/06, Maarten Bosma <maarten.paul(a)bosma.de> wrote:
>   Hi,
>
>  I think ReactOS lacks of testers since WaxDragon is gone. 
Just because I don't have time to be TC doesn't mean I'm gone. ;0)

>   So I had the following idea: A page which lists all
> regressions and gives you the
>  option to report a working / non-working revision and calculates the
>  range in which it the regression was caused from that data. 
Honestly , Bugzilla is the best place.  Since testers should not only
concern themselves with regressions, but unimplemented features. 
Having a separate page will just be another thing that needs
maintained, and frankly getting good bug reports is hard enough.

>  
>  I know that bugzilla comments are actually sufficient for this and of
>  course this system would have to be linked with bugzilla, but I think
>  that a page where the regressions are listed up would be more motivating
>  to help testing.
>
>  What do you think about it ? 
I think a "status" page and a testing guide would be most helpful. 
The status page could be used to fend off the inevitable questions,
and the guide will highlight the things that *do* need testing, and if
everyone is testing the same things, then regressions will become
self-evident.

Here is the "guide", which is just my testing notes:
http://waxdragon.homeip.net/~ford/reactos/ros_regression_items.html

We really need a TC.

WD

--
<Bizzeh> it even has a panda that pops out of your case and mauls
people who ask stupid questions

[Aleksey Bragin]

I think the similar way - bugzilla can handle everything we need:  
bugs, regressions, etc. And it provides one "window" through which  
anyone can see all currently outstanding problems.

WaxDragon - what do you think about utilising some tool like scmbug?
Testing guide would be very cool too.


WBR,
Aleksey Bragin.

On Mar 25, 2006, at 2:49 AM, WaxDragon wrote:

>   On 3/24/06, Maarten Bosma
> <maarten.paul(a)bosma.de> wrote:
> >   Hi,
> >
> >  I think ReactOS lacks of testers since WaxDragon is gone. 
>  Just because I don't have time to be TC doesn't mean I'm gone. ;0)
>
> >   So I had the following idea: A page which lists
> > all regressions  
> >  and gives you the
> >  option to report a working / non-working revision and calculates the
> >  range in which it the regression was caused from that data. 
>  Honestly , Bugzilla is the best place.  Since testers should not only
>  concern themselves with regressions, but unimplemented features.
>  Having a separate page will just be another thing that needs
>  maintained, and frankly getting good bug reports is hard enough.
>
> >  
> >  I know that bugzilla comments are actually sufficient for this and of
> >  course this system would have to be linked with bugzilla, but I think
> >  that a page where the regressions are listed up would be more  
> >  motivating
> >  to help testing.
> >
> >  What do you think about it ? 
>  I think a "status" page and a testing guide would be most helpful.
>  The status page could be used to fend off the inevitable questions,
>  and the guide will highlight the things that *do* need testing, and if
>  everyone is testing the same things, then regressions will become
>  self-evident.
>
>  Here is the "guide", which is just my testing notes:
>  http://waxdragon.homeip.net/~ford/reactos/ros_regression_items.html
>
>  We really need a TC.
>
>  WD
>
>  --
>  <Bizzeh> it even has a panda that pops out of your case and mauls
>  people who ask stupid questions

[Maarten Bosma]

WaxDragon wrote:
>   Honestly , Bugzilla is the best place.  Since testers
> should not only
>  concern themselves with regressions, but unimplemented features. 
>  Having a separate page will just be another thing that needs
>  maintained, and frankly getting good bug reports is hard enough. 
You can simply query the information from the bugzilla database.


Maarten Bosma

[Maarten Bosma]

WaxDragon wrote:
>   I think a "status" page and a testing guide
> would be most helpful. 
>  The status page could be used to fend off the inevitable questions, 
How exactly would that status page look like ? Who would maintain it ?

>   and the guide will highlight the things that *do* need
> testing, and if
>  everyone is testing the same things, then regressions will become
>  self-evident. 
Dosn't everything need testing ?


Maarten Bosma

[WaxDragon]

>   How exactly would that status page look like ? Who
> would maintain it ? 
I haven't thought much about it, but there have been several requests for one.

>   Dosn't everything need testing ? 
Things that are not implemented or blocked by another feature do not
need to be tested, yet.  Having a guide lets testers know what is
known to be working and needs regression testing, and what features
are being worked on and need functional and regression testing.

WD
--
<Bizzeh> it even has a panda that pops out of your case and mauls
people who ask stupid questions

[Maarten Bosma]

WaxDragon wrote:
>   Things that are not implemented or blocked by another
> feature do not
>  need to be tested, yet.  Having a guide lets testers know what is
>  known to be working and needs regression testing, and what features
>  are being worked on and need functional and regression testing. 
Then the support database is actually what you want.


Maarten Bosma