Re: [Xen-devel] [XEN PATCH 3/8] xen: Update Kconfig to Linux v5.4

[Anthony PERARD <anthony.perard@xxxxxxxxxx>]

On Mon, Dec 16, 2019 at 04:55:52PM +0100, Jan Beulich wrote:
> On 16.12.2019 15:01, Anthony PERARD wrote:
> > On Mon, Dec 16, 2019 at 11:16:52AM +0100, Jan Beulich wrote:
> >> What headers are you taking about? My question was about the placement
> >> of .gitignore entries only. I'm pretty sure I had previously expressed
> >> that I'm not overly happy to see needless scattering around of them.
> >> I'm merely trying to understand if here we have a case of "needless".
> >>
> >>> I find your lack of consistency in review requests impossible to predict.
> >>
> >> What consistency? What's wrong with putting the two lines in the top
> >> level .gitignore, or keeping them where they are? I anyway doubt
> >> there's a need to keep our .gitignore in sync with Linux'es.
> > 
> > I think scattering the .gitignore around make them easier to manage,
> > especially when entries aren't useful anymore. Otherwise, we end up
> > with a .gitignore at the root with entries that aren't needed anymore
> > because they can be hard to figure out if they are useful or not.
> > Or, when globing is involve, we could end-up ignoring files that
> > shouldn't.
> 
> Right, especially this last point makes scattering entries have some
> value. But then let's truly scatter them, not accumulate all sorts
> of stuff in xen/.gitignore.
> 
> > I'm planning to import more of Kbuild, which Makefile.host comes from,
> > to build only the hypervisor, so there's going to be other artefact that
> > will be generated only in xen/. So I've added those two new entries in
> > xen/.gitignore in anticipation, and to avoid getting xen/tools/kconfig
> > out-of-sync with Linux's one (Well I didn't know those two entries were
> > there in the first place).
> > 
> > I could add "/xen/**/*.tab.[ch]" in the root .gitignore if you prefer
> > even though I don't like this approch.
> 
> But this then still is too wide a pattern, as per above. A
> sufficiently specific one would be as good as one without path
> in a .gitignore deep enough into the tree, if we wanted to go
> the "truly scatter" route.

Fine, we can have those two entries into kconfig/.gitignore, with a note
in the commit message saying that the only thing different from
upstream. That's where they belong for now.

-- 
Anthony PERARD

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel