[interfacekit] Re: A patch for BString

On 2003-11-12 at 08:53:07 [+0100], burton666@xxxxxxxxx wrote:
> 
> Ok, I applied the patch to my local tree. Before committing it, I'm going 
> to refactor a bit the code (braces weren't used consistently, my fault). 
> Another thing: Oliver, you changed the return value from B_BAD_VALUE to 
> B_ERROR when the user supply a NULL string to FindFirst()/Last(), etc.. But 
> IMHO, B_BAD_VALUE is better in that case (and I tought we decided that way 
> long ago). No problem, though. I'll "fix" it.

Sorry, I should have explained that change with more detail. My thinking was 
that with my patch I tried to enhance BString while strictly following R5 
API, behaviour and documentation (unless there are clear bugs). 
Since the R5-documentation states that the (I)Find...-methods only ever 
return B_ERROR as a negative return-value, I thought it'd be best to do just 
that. Otherwise R5-applications could misbehave/crash if they get B_BAD_VALUE 
when they expect B_ERROR. 
But I can see that you can argue both ways, so I don't really mind using 
B_BAD_VALUE...

> Expect the commit in the next days.

Great!

> P.S: nice work on the replace methods, Oliver. I like this version even 
> more than the former.

Thanks >:o) I prefer the new one, too! The first one was just a backport from 
the B(m)String implementation used by Beam, which was taking along a helper 
class that didn't quite fit-in with BString. That's why I decided to do it 
more properly.

cheers,
        Oliver

Other related posts: