Re: [api-dev] Re: Missing interfaces in OOo 3.2: Why not a blocker?

2009-12-02 Thread Daniel B.
Hi,

On Fri, Nov 27, 2009 at 1:48 PM, Juergen Schmidt
juergen.schm...@sun.com wrote:
 Daniel B. wrote:
 Since our (and potentially a lot of other) extension breaks with this
 change this is a major issue that prevents us from rolling out OOo
 3.2. Reverting the change until OOo 4.0 would give us enough time to
 update our extension and roll out the updated version. So I support
 Michael's suggestion to re-implement the interfaces that are missing,
 at least temporary, and document the change so that people are aware
 of the issue and can adapt their macros/extensions.

 until now you are the only one with this problem. How many places using this
 code do you have in your extension? We gave you the necessary info to do it
 correct. It should be not really a big change i guess but anyway. I let it
 up to others to decide if it is a showstopper or not.

 But how can we prevent you from rolling out 3.2 when it is not already final
 and released? Either you change your broken or wrong extensions or we do the
 changes back. You have to do it anyway in the near future whereas we have
 double work. Sounds not really optimal ;-)

We really shouldn't argue this based on who has the most work because
of this change but based on what is the right thing to do in this
case. The fact of the matter is that multiple interfaces were removed
without advance notice. Sure, these interfaces may have been included
in error in the first place but since neither the wrong interfaces
nor the right way to do it were documented how could someone have
known that he used something that wasn't even supposed to be there? So
you can't really blame someone for using those interfaces. Okay, maybe
you CAN blame them for using _anything_ that isn't documented but in
this case that would have meant to just completely do without this
functionality which wasn't really a good option either.
So, removing the interfaces in general is fine, but not without giving
people ample time to react to this change and to learn how to do it
right. Just breaking their extensions/macros from one minor version to
the next isn't very nice.

But yes, I'm biased because our extension is affected. You are right
that it's not such a big change, and I may have exaggerated a bit. But
you must understand that it's a bit of a shock when you realize that
some interfaces suddenly just disappear from OOo 3.1 to 3.2 without
any warning or explanation. It also has the effect that you begin to
wonder what else could have gone missing that maybe you didn't notice
in your test cases.

Anyway, we now have an explanation what happened and we can do the
necessary changes. I still think it would be better to postpone the
removal of the interfaces to a later version, but if really noone else
has a problem with it I'm not going to argue any further.


Regards,
Daniel

-
To unsubscribe, e-mail: dev-unsubscr...@api.openoffice.org
For additional commands, e-mail: dev-h...@api.openoffice.org



Re: [api-dev] Re: Missing interfaces in OOo 3.2: Why not a blocker?

2009-12-02 Thread Juergen Schmidt

Daniel B. wrote:

Hi,

On Fri, Nov 27, 2009 at 1:48 PM, Juergen Schmidt
juergen.schm...@sun.com wrote:

Daniel B. wrote:

Since our (and potentially a lot of other) extension breaks with this
change this is a major issue that prevents us from rolling out OOo
3.2. Reverting the change until OOo 4.0 would give us enough time to
update our extension and roll out the updated version. So I support
Michael's suggestion to re-implement the interfaces that are missing,
at least temporary, and document the change so that people are aware
of the issue and can adapt their macros/extensions.


until now you are the only one with this problem. How many places using this
code do you have in your extension? We gave you the necessary info to do it
correct. It should be not really a big change i guess but anyway. I let it
up to others to decide if it is a showstopper or not.

But how can we prevent you from rolling out 3.2 when it is not already final
and released? Either you change your broken or wrong extensions or we do the
changes back. You have to do it anyway in the near future whereas we have
double work. Sounds not really optimal ;-)


We really shouldn't argue this based on who has the most work because
of this change but based on what is the right thing to do in this
case. The fact of the matter is that multiple interfaces were removed
without advance notice. Sure, these interfaces may have been included
in error in the first place but since neither the wrong interfaces
nor the right way to do it were documented how could someone have
known that he used something that wasn't even supposed to be there? So
you can't really blame someone for using those interfaces. Okay, maybe
you CAN blame them for using _anything_ that isn't documented but in
this case that would have meant to just completely do without this
functionality which wasn't really a good option either.
i don't want to blame anybody but i wanted to raise more awareness. If 
an API is not documented (which should be fixed always) ask before you 
use it. That would prevent situations like this. Nobody should rely on 
the given implementation when it is not documented. If yes than they 
have to live with changes.


The users of our API can help a lot to improve APIs and the 
documentation by communicating their problems, their ideas etc. Don't 
hesitate to submit an issue. It's always a valuable contribution and 
don't take it personal if an issue/feature request won't be fixed.




So, removing the interfaces in general is fine, but not without giving
people ample time to react to this change and to learn how to do it
right. Just breaking their extensions/macros from one minor version to
the next isn't very nice.
i agree at least a note on interface-announce should have taken place. 
But the developer forget it and didn't thought about the relevance i guess




But yes, I'm biased because our extension is affected. You are right
that it's not such a big change, and I may have exaggerated a bit. But
you must understand that it's a bit of a shock when you realize that
some interfaces suddenly just disappear from OOo 3.1 to 3.2 without
any warning or explanation. It also has the effect that you begin to
wonder what else could have gone missing that maybe you didn't notice
in your test cases.
i agree again and yes we should try to improve the overall situation. We 
need also better tests. When we allow even incompatible changes in the 
future it should never happen without a clear communication and a well 
documented workaround.




Anyway, we now have an explanation what happened and we can do the
necessary changes. I still think it would be better to postpone the
removal of the interfaces to a later version, but if really noone else
has a problem with it I'm not going to argue any further.
as i mentioned before it's up to you to put it on or suggest it for the 
showstopper list on the relea...@openoffice.org list


Regards

Juergen




Regards,
Daniel

-
To unsubscribe, e-mail: dev-unsubscr...@api.openoffice.org
For additional commands, e-mail: dev-h...@api.openoffice.org




-
To unsubscribe, e-mail: dev-unsubscr...@api.openoffice.org
For additional commands, e-mail: dev-h...@api.openoffice.org