On Fri, Sep 15, 2006 at 04:58:43PM +0400, Ivan Zhakov wrote:
> On 9/11/06, Malcolm Rowe <malcolm-svn-dev@farside.org.uk> wrote:
> >I notice that I've got some entries in my newly-upgraded 1.4 WC's that
> >have non-blank ->copyfrom_url and ->copyfrom_rev values, but also have
> >->copied set to FALSE.
> >
> >
> >Would it be appropriate to add some code to read_entry() to force
> >copyfrom_* to empty if copied is FALSE? (and error if copied is TRUE
> >and copyfrom_* is empty?)
> Yes, you are right. It's a bug and it should be fixed. But
> read_entry() is not appropriate place for do this, because this
> copyfrom_url/copyfrom_rev should be fixed before writing entries to
> the disk. So for me right place for fix this bug is fold_entry().
>
I was thinking that we'd want to filter out the bogus entries on read,
so that we could rely on invariants like the above being true everywhere,
but you're right that fold_entry() is already responsible for cleaning
up meaningless combinations, so it makes sense to add more checks there.
Regards,
Malcolm
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Sep 15 15:45:16 2006