Re: XPSSourceInformation.java

Christian Egli wrote:

>I just spent a few good hours of my life wondering why my publication
>did not work anymore.
>

Which publication? Your personal one or one of the distribution?

Why didn't you consult the mailing list?

> Turns out that I had to revert a change in
>XPSSourceInformation.java to make it work again.
>
>There's a few things that I'd like to say about this change:
>
>1. I don't particularly like <change author="foobar"> tag. This kind
>   of information is stored in CVS. In the source code it's just
>   noise.
>
I agree

>
>2. I suppose if we had some kind of regression test this wouldn't
>   have happened.
>

Well, I did some human regression test, and it seemed fine to me.That's 
why I checked it in.
But +100000 for writing regression tests.Can you start writing some?

>
>3. For people with productive servers that sync with us by CVS this
>   can be a nasty surprise. We should probably separate development
>   like that on a development branch and keep changes of the stable
>   branch to a minimum.
>
We don't have a stable branch yet, but we certainly will as we decided 
during our
brainstorming day.  (BTW: That reminds me that we should publish the 
outcome of our brainstorming day.)


Concerning the revert of the change:

I guess your publication is working now, but since you reverted it, it 
won't work
on Windows anymore. So, did it get any better? Regression tests 
certainly help a lot,
but in this case they would have failed in the case of Windows!

I think it would have been better to discuss it
first on the mailing list (although I am not a Windows User and why 
should I care).
Well, we are a community and that's why we should communicate ;-)


So, please tell us what didn't work,such that we can really fix this bug.

Thanks

Michael

>
>  
>