On Thu, Oct 22, 2015 at 02:43:09PM +0200, Andrea Bolognani wrote:
On Thu, 2015-10-22 at 13:27 +0200, Martin Kletzander wrote:
> > > As for patch 2, one would have to know they should use the
> > > copy-from-host.sh script. Perhaps what might be better and/or
> > > somewhat
> > > more interesting on this one is some make check rule that scans
> > > the
> > > nodeinfodata trees looking for files that shouldn't be there.
> > > That
> > > way
> > > if someone does use their own methodology to copy over the tree
> > > we'd
> > > know it (and could message to use the copy-from-host.sh script...
> >
> > I agree, as it stands it's not very discoverable, plus
> > adding the check you suggest would also prevent something
> > like e739d95 from ever being needed again.
> >
> > I'll work on that as soon as I have some time.
> >
>
> Maybe simple .gitignore entry would suffice.
Problem is, .gitignore works well if you have the list of
files you want to avoid, but we have the complementary
information: the list of files we want to keep.
You could do that as well, I took the liberty of trying that out ant
it looks like it works, but looking at it, it's very ugly. Decide
yourself whether you want to use it or not, I'll just paste the file
below. The .gitignore file goes into tests/nodeinfodata/.gitignore:
/*/*
!/*/node/
/*/node/*
!/*/node/online
!/*/node/possible
!/*/node/node*/
/*/node/node*/*
!/*/node/node*/cpu[0-9]*
!/*/node/node*/cpulist
!/*/node/node*/cpumap
!/*/node/node*/meminfo
!/*/cpu/
/*/cpu/*
!/*/cpu/kernel_max
!/*/cpu/offline
!/*/cpu/online
!/*/cpu/possible
!/*/cpu/present
!/*/cpu/cpu[0-9]*/
/*/cpu/cpu*/*
!/*/cpu/cpu*/online
!/*/cpu/cpu*/topology