Re: svn_wc_conflict_result_choose_repos found dangerous?

On 7/18/07, Ben Collins-Sussman <sussman@red-bean.com> wrote:
> On 7/17/07, Mark Phippard <markphip@gmail.com> wrote:
> > On 7/17/07, Ben Collins-Sussman <sussman@red-bean.com> wrote:
> > > Aw, crud, you guys are right. I need to fix that. When perforce
> > > prompts you whether you want to "choose mine or theirs", it's
> > > referring to conflicting hunks, not the entire file. Gah.
> > >
> > > Thanks for pointing this out.
> >
> > Just to be clear, I do not think it needs to prompt hunk by hunk, it
> > would be fine if it simply used your choice for the hunks where it
> > encounters conflicts.
> >
> > I'd imagine binary files would have to work the way you have it right now.
> >
>
> I agree 100%. In the case where the 'merged' file doesn't exist (i.e.
> binary files), selecting mine/theirs would choose the entire file.
> But if a 'merged' file does exist, we need to scan the file, detect
> conflict markers, and select one side of the hunk.
>
> That said, doesn't Jeremy's 'svn resolved --accept {left|right}' code
> have the exact same problem? It shouldn't be accepting the entire
> file, but selecting certain sides of conflicted hunks. If so, we're
> finally at a place where Jeremy and I can share some common code to do
> this. :-)
>

Here is the link to the beginning of the thread where I raised the
same question:

http://svn.haxx.se/dev/archive-2007-06/0014.shtml

I need to re-read it to. I cannot remember if the consensus was that
it was just not needed, too much work, or reasons to not want it at
all. svn resolved has the advantage that the files are already
created so you can at least always say the user had the option to
examine the file before they made their decision.

-- 
Thanks
Mark Phippard
http://markphip.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org