Georg Baum wrote:
> It was easy to verify that owner_ is always set. I also made it more clear
> that it should never change by making it a const member.
I meant that the buffer to which owner_ points to must remain alive.
Quick peek at Buffer.cpp shows it should be fine during reload.
Pavel
Pavel Sanda wrote:
> Georg Baum wrote:
>> What do you prefer?
>
> No hard opinion, but if you are going to ditch it please make sure
> that owner_ object is living all the time and does not get invalid for
> a critical moment when reloading buffer or so (this kind of robustness
> against changes
Georg Baum wrote:
> What do you prefer?
No hard opinion, but if you are going to ditch it please make sure
that owner_ object is living all the time and does not get invalid for
a critical moment when reloading buffer or so (this kind of robustness
against changes in various bufferReload routines
I am currently cleaning up a patch that implements LFUN_VC_RENAME and
LFUN_VC_COPY for submission. I implemented it originally for my wife in 1.6,
and do not want to port it on each upgrade;-) During the cleanup I stumbled
over the file_ member of the CVS and SVN classes. RCS does not have it, a