Boost logo

Boost Users :

From: Gennadiy Rozental (gennadiy.rozental_at_[hidden])
Date: 2006-03-13 14:09:13


"Mateusz Loskot" <mateusz_at_[hidden]> wrote in message
news:4415BCE3.8000304_at_loskot.net...
> Gennadiy Rozental wrote:
>>>
>>> I have some more questions if you don't mind. Unfortunately, I'm a
>>> bit disappointed about how easy it's possible to break backward
>>> compatibility between close releases of Boost.
>>
>>
>> Boost.Test *is* backward compartible. It's not forward compartible
>> though. Meaning you couldn't use new examples with old library. Some
>> features are indeed getting deprecated. But usually it takes 2-3
>> releases at least (in time frame it's close to 1-2 years).
>>
>> Unless you have any particular issues using old example with new
>> library, do you?
>
> Hi Gennadiy,
>
> It seems I don't understand something or I've not expressed the problem
> clearly. Let me try again.
>
> I started with Boost.Test from Boost 1.34.
> After some time, I realized that many of my users will be able to use
> only Boost 1.33.
> Then, I noticed that my Unit Tests based on Boost's UTF are not working
> with Boost 1.33.
> So, that's what I understand as a "broken backward compatibility".
> Boost UTF interface seems to be changed in 1.34, so it's not possible to
> use Unit Tests written using Boost 1.34 with Boost 1.33.
> If Boost 1.34 was backward compatible, then those tests prepared with
> 1.34 could be run using 1.33. Am I right?

No. This is forward compartibility. You want older version handle code
written for newer version

> OK, I agree backward/forward understanding may be different, so please
> don't stick to those words, but to my explanation.
>
>
> Now, how should I move on with Boost.Test to ensure that:
> - users using Boost 1.33 will be able to run my tests
> - users using Boost 1.34 and higher will be able to run my tests

Don't use features/interfaces introduced in 1.34. Stick to 1.33 one.

Gennadiy


Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net