how to pre-screen patch submissions from a newbie janitor?

Robert P. J. Day rpjday at crashcourse.ca
Wed May 19 07:40:16 EDT 2010


On Wed, 19 May 2010, Wolfram Sang wrote:

> >   however, since he's new to patch creation and submission, i
> > offered to "vet" his patches first.  i realize that my vetting
> > carries no weight on the MTD list, but given that i've submitted
> > lots and lots of patches elsewhere, i can still at least sanity
> > check what he comes up with before it hits the list.
> >
> >   is there a canonical signage for that nowadays?  is it
> > "Reviewed-By" or something like that?  so that by the time it gets
> > to the MTD list, he would have done the "Signed-Off-By" and i'd
> > sign off as a reviewer. is that the way it works?  or is there an
> > alternative?  thanks.
>
> Yes, that sounds like a good case for Reviewed-by. It will surely
> help these patches knowing there is somebody with experience who
> already had a look at (and in that case probably build-tested) them.
>
> Looking forward to the series,

  what i'm going to have this trainee(?) do is just simplify the
config/Kconfig/Makefile combos, which i openly admit is not
earth-shaking in its complexity.

  for example. in a number of places elsewhere in the kernel tree,
i've submitted patches to change things like this in the config menus:

<M>   OneNAND Device Support  --->
      LPDDR flash memory drivers  --->
      UBI - Unsorted block images  --->

  note how, to select either LPDDR or UBI, you need to *go* to those
submenus, where you find a top-level selector.  it's simpler to just
put that selector right on that menu item, as it's done for "OneNAND"
just above it.  no wasting time.

  and another simplification is to cut down the conditional inclusion
in the makefiles.  for instance, here's tests/Makefile:

obj-$(CONFIG_MTD_TESTS) += mtd_oobtest.o
obj-$(CONFIG_MTD_TESTS) += mtd_pagetest.o
obj-$(CONFIG_MTD_TESTS) += mtd_readtest.o
... etc etc ...

  there's little point having every single line in the Makefile test
CONFIG_MTD_TESTS when that can all be done with a single test one
level up that checks whether or not to process the tests/ directory at
all.

  in short, nothing spectacular, just tidying up.

rday

-- 

========================================================================
Robert P. J. Day                               Waterloo, Ontario, CANADA

            Linux Consulting, Training and Kernel Pedantry.

Web page:                                          http://crashcourse.ca
Twitter:                                       http://twitter.com/rpjday
========================================================================



More information about the linux-mtd mailing list