[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: IMPORTANT REMINDER: no more binary incompatible changes after
From:       Nicolas Brodu <nicolas.brodu () free ! fr>
Date:       2000-05-31 12:42:20
[Download RAW message or body]

Stephan Kulow wrote:
> 
> Waldo Bastian wrote:
> >
> > On Tue, 30 May 2000, Kurt Granroth wrote:
> > > Alexei Dets wrote:
> > > > Kurt Granroth wrote:
> > > > > Tomorrow is the *LAST* time we will have to add binary incompatible
> > > > > changes to kdelibs.  This is because 1.91 (Kleopatra) will be coming
> > > > > out the following Wednesday.  And, from the release schedule:
> > > >
> > > > At least KURL needs to be fixed! And KProcess too. They don't work
> > > > correctly with non-ascii paths and arguments. In fact they don't
> > > > work at all in this case!!! I'm 99% sure that this will require
> > > > binary (and may be even source) incompatible changes. And such
> > > > important thing can't be done in one day :-(
> > >
> > > What binary incompatible API changes are necessary?
> >
> > The IO-slave interface needs to be changed from qstrings to KURL. KURL also
> > needs some fixes, I could use the d-pointer for that, but I rather not
> > introduce an extra 'new' in all the constructors if that's isn't strictly
> > necassery.
> >
> I have no problem with changing the IO-slave interface if it's put big
> into
> documentation that you should not check for any sanity of the URL and
> just
> think of the passed URL as container.

At present, I reconstruct an internal URL from what the smb slave 
receive. For example, one problem is that what KURL think is an host might
actually be a workgroup, so calling a library function is needed to decide on
this. If the slave receive the whole URL, then it's much simpler, since it
can just be passed on to the library (which will detect host/workgroup by
itself).

My question is, when you say 'no sanity check', does it include not checking for
it to be in decoded form?

Anyway, no objection here for receiving the whole URL in the slave.

Regards,
Nicolas
-- 
A shortcut is the longest distance between two points. (unknown author)

[prev in list] [next in list] [prev in thread] [next in thread]