Re: [Zephyr-users] Introducing west, Zephyr's meta-tool

Carles Cufi

Hi Luiz,

-----Original Message-----
From: Luiz Augusto von Dentz <>
Sent: 29 January 2019 12:25
To: Cufi, Carles <>
Cc: Jamie Mccrae <>; zephyr-devel <zephyr->;
Subject: Re: [Zephyr-devel] [Zephyr-users] Introducing west, Zephyr's

Hi Carles,
On Tue, Jan 29, 2019 at 11:40 AM Cufi, Carles
<> wrote:

Hi Jamie, Luiz,

-----Original Message-----
From: Jamie Mccrae <>
Sent: 29 January 2019 10:35
To: Luiz Augusto von Dentz <>; Cufi, Carles
Cc: zephyr-devel <>; zephyr-
Subject: RE: [Zephyr-devel] [Zephyr-users] Introducing west,
Zephyr's meta-tool

Hi Carles,
Note that we have not at any point considered moving away from
GitHub. The plan is to adapt to what the service offers and try to
find a solution that works well with it. This is also one of the
reasons we are first introducing west without actually splitting the
ext/ folder into multiple external repos: we want to expose users to
west before we start doing more radical changes.

Well good, though Im not sure what problem does west solves then?
For pulling it sounds like submodules is the preferable solution
(to linux folks at least):
I agree with Luiz here, what does west provide that submodules
doesn't provide? Why do I and others now need yet another piece of
software to build Zephyr code (python)?
There are a number of reasons why we don't use Git submodules, all of
which are described in the GitHub issue describing multi-repo:
Well I didn't have to go as far since someone already commented similar

So attached vs detached approach, there seems to be quite many on the
camp that we need to have the repositories attached while some
considered the detached approach a must have. As individual contributor
I think using the attached approach is a must, specially if one wants to
actually run sanity checks which sometimes is necessary e.g. when you
are actually writing a test case which depends on a feature from a
different repository.
If by "attached" you mean that projects in the manifest are locked at a particular SHA, then that is exactly what we will do upstream ( That is not to say that some downstreams may choose to free float on the tip of a branch, but this is *not* going to be the case for upstream Zephyr. I will edit the body of the issue to make that clear.



Join to automatically receive all group messages.