Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1360340 > unrolled thread

Re: [GIT PULL] Staging driver patches for 4.6-rc1

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [GIT PULL] Staging driver patches for 4.6-rc1 Linus Torvalds <torvalds@linux-foundation.org> - 2016-03-18 06:20 +0100
    Re: [GIT PULL] Staging driver patches for 4.6-rc1 Greg KH <gregkh@linuxfoundation.org> - 2016-03-18 16:30 +0100

#1360340 — Re: [GIT PULL] Staging driver patches for 4.6-rc1

On Thu, Mar 17, 2016 at 8:25 PM, Greg KH <gregkh@linuxfoundation.org> wrote:
> Ok, the diffstat below seems "odd" in that I had done a merge with my
> char-misc tree to resolve some merge issues a while ago, and that tree
> is now in your tree, so the diffstat shouldn't be showing it (I updated
> my master branch), but somehow it is.

When there isn't a single well-defined common base for the merge
(which happens when you cross-merge, but also if you have multiple
different bases internally and then merge inside one branch), the "git
diff" output will pick one of the (multiple) common bases, and use
that as the base of the diff.

That usually works fine, but it can also give wildly wrong diffstat
results, simply because the only way to get the final real diff would
be to just do the merge.

> Also, there are a number of conflicts here when merging with your tree.
>         both modified:   drivers/media/dvb-frontends/mn88473.c
>         deleted by them: drivers/staging/dgap/dgap.c
>         both modified:   drivers/staging/lustre/lnet/libcfs/linux/linux-crypto.c
>         both modified:   drivers/staging/media/davinci_vpfe/davinci_vpfe_user.h
>         both modified:   include/linux/irqdomain.h
>
> I can fix these up and resend this with a merged tree pull request if
> you want me to, it's your call.

No, all good. I've already merged it, I'm just waiting for the build
test to finish before pushing out the end result.

You might want to double-check the end result, but nothing looked
particularly odd.

                 Linus

[toc] | [next] | [standalone]

#1360730

On Thu, Mar 17, 2016 at 10:19:04PM -0700, Linus Torvalds wrote:
> On Thu, Mar 17, 2016 at 8:25 PM, Greg KH <gregkh@linuxfoundation.org> wrote:
> > Ok, the diffstat below seems "odd" in that I had done a merge with my
> > char-misc tree to resolve some merge issues a while ago, and that tree
> > is now in your tree, so the diffstat shouldn't be showing it (I updated
> > my master branch), but somehow it is.
> 
> When there isn't a single well-defined common base for the merge
> (which happens when you cross-merge, but also if you have multiple
> different bases internally and then merge inside one branch), the "git
> diff" output will pick one of the (multiple) common bases, and use
> that as the base of the diff.
> 
> That usually works fine, but it can also give wildly wrong diffstat
> results, simply because the only way to get the final real diff would
> be to just do the merge.

Ah, ok, that makes more sense, thanks.

> > Also, there are a number of conflicts here when merging with your tree.
> >         both modified:   drivers/media/dvb-frontends/mn88473.c
> >         deleted by them: drivers/staging/dgap/dgap.c
> >         both modified:   drivers/staging/lustre/lnet/libcfs/linux/linux-crypto.c
> >         both modified:   drivers/staging/media/davinci_vpfe/davinci_vpfe_user.h
> >         both modified:   include/linux/irqdomain.h
> >
> > I can fix these up and resend this with a merged tree pull request if
> > you want me to, it's your call.
> 
> No, all good. I've already merged it, I'm just waiting for the build
> test to finish before pushing out the end result.
> 
> You might want to double-check the end result, but nothing looked
> particularly odd.

All looks good, thanks for doing the merge.

greg k-h

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web