[interfacekit] Re: Reproducing... bug ?

>By example, if we try to set the size of the buffer in BMemoryIO and 
there's
>not enough room for that, the BeBook said the function return 
B_NO_MEMORY.
>But in reality when I made my testings, the function actually return a
>straight B_ERROR.
>
>Ok it's not a big deal. But I want to know if we reproduce the exact
>behavior from the original libbe.so  or we path the problems by 
following
>blindly the BeBook ?

I would still #ifdef with the current behaviour as the default.  After 
we've had a chance to test with real apps, we can move to the "correct" 
behaviour if we're not breaking anything.

In general, the actual behaviour is what we want to model.  The BeBook 
frequently fails to match reality, and reality is what apps are dealing 
with, so we need to as well. =)  Having said that, existing behaviour is 
less than optimal, and it would be nice to slip in any fixes we can.

e

>- Steve
>
>
>----- Original Message -----
>From: "Erik Jakowatz" <erik@xxxxxxxxxxxxxx>
>To: <svallee@xxxxxxxxxxxxxx>
>Cc: "interface kit team" <interfacekit@xxxxxxxxxxxxx>
>Sent: Friday, January 25, 2002 10:15 AM
>Subject: Re: Reproducing... bug ?
>
>
>> Do both and #ifdef them.  I think we'll need to do some "field tests" 
to
>> determine whether or not we can make the improvement.  I would like 
to
>> see it work correctly if we aren't going to break a bunch of apps by
>> doing it.  I have a hunch that BMemoryIO is not a widely used class.
>>
>> Any other thoughts, folks?
>>
>> e
>>
>> >I'm currently doing full testing on my code, and realized there's a 
bug
>> (or
>> >non-implemented) method. It's the Seek() of BMemoryIO. It *always*
>> return 0,
>> >and do absolutely nothing, whatever the parameter we give. So as the
>> >function is supposed to move the current pointer in the Memory IO, 
do I
>> ..
>> >
>> >1) Emulate the *exact* behavior of the original Seek ...
>> >
>> >2) Implement it as it's supposed to be, because anyway no 
applications
>> sould
>> >have used it before especially because of it's bug...
>> >
>> >What do I choose in that kind of situation ?
>> >
>> >Thanks a lot !
>> >Steve

Data is not information, and information is not knowledge: knowledge is 
not understanding, and understanding is not wisdom.
        - Philip Adams

Other related posts: