Re: [RFC PATCH 1/5] add lwip-external submodule

[Tom Rini <trini@konsulko.com>]

[-- Attachment #1: Type: text/plain, Size: 1838 bytes --]

On Wed, May 10, 2023 at 03:46:31PM +0100, Peter Robinson wrote:
> On Wed, May 10, 2023 at 8:40 AM Ilias Apalodimas
> <ilias.apalodimas@linaro.org> wrote:
> >
> > Hi Simon,
> >
> > On Mon, May 08, 2023 at 08:43:14AM -0600, Simon Glass wrote:
> > > Hi Maxim,
> > >
> > > On Fri, 5 May 2023 at 04:50, Maxim Uvarov <maxim.uvarov@linaro.org> wrote:
> > > >
> > > > Signed-off-by: Maxim Uvarov <maxim.uvarov@linaro.org>
> > > > ---
> > > >  .gitmodules            | 3 +++
> > > >  lib/lwip/lwip-external | 1 +
> > > >  2 files changed, 4 insertions(+)
> > > >  create mode 100644 .gitmodules
> > > >  create mode 160000 lib/lwip/lwip-external
> > > >
> > >
> > > Please no submodules. They are such a pain. If we want the code in
> > > U-Boot, let's put it in U-Boot and upstream our changes as needed.
> >
> > Can you explain a bit more the pain points you are seeing in u-boot with
> > submodules?  EDK2 does submodules for openSSL and it's quite convenient,
> > since you dont have to maintain any code, do backports etc.  Instead we can
> > just use upstream projects as-is.
> 
> I feel there's pros and cons for both, similarly different projects
> have different projects have different policies. Tom may have a more
> definite opinion.
> 
> > IMHO we should work on having it as an experimental feature in parallel
> > with the current TCP efforts for a while and have a Kconfig switch.  If we
> > are happy in the long run and the code increase isn't prohibitive, we can
> > consider switching permanently
> 
> At least for initial review of the prototype I don't see it as
> blocking for people to get a general idea what is going on.

Yeah, I also don't like submodules. But for the purpose of this RFC, at
this stage in the RFC, it's fine.  May or may not be easier to review.

-- 
Tom

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 659 bytes --]