Re: [Buildroot] [PATCH v2 1/4] Makefile: only check _FINAL_DEPENDENCIES are selected

From: "Yann E. MORIN" <yann.morin.1998@free.fr>
To: James Hilliard <james.hilliard1@gmail.com>
Cc: "Naumann Andreas" <ANaumann@ultratronik.de>,
	"Joshua Henderson" <joshua.henderson@microchip.com>,
	"Angelo Compagnucci" <angelo.compagnucci@gmail.com>,
	"Bartosz Bilas" <b.bilas@grinn-global.com>,
	"Peter Seiderer" <ps.report@gmx.net>,
	"Julien Corjon" <corjon.j@ecagroup.com>,
	"Thomas Petazzoni" <thomas.petazzoni@bootlin.com>,
	buildroot <buildroot@buildroot.org>,
	"Gaël Portay" <gael.portay@collabora.com>
Subject: Re: [Buildroot] [PATCH v2 1/4] Makefile: only check _FINAL_DEPENDENCIES are selected
Date: Thu, 6 Jan 2022 21:19:25 +0100	[thread overview]
Message-ID: <20220106201925.GP614810@scaer> (raw)
In-Reply-To: <CADvTj4otRbTpruWEZF=FcYH20x72_wi98A2_WTDYzSKH0Q0mrA@mail.gmail.com>

James, All,

On 2022-01-06 13:14 -0600, James Hilliard spake thusly:
> On Thu, Jan 6, 2022 at 1:07 PM Arnout Vandecappelle <arnout@mind.be> wrote:
> > On 06/01/2022 19:51, James Hilliard wrote:
> > > On Thu, Jan 6, 2022 at 11:59 AM Thomas Petazzoni
> > > <thomas.petazzoni@bootlin.com> wrote:
> > >> On Wed,  5 Jan 2022 23:24:04 -0700
> > >> James Hilliard <james.hilliard1@gmail.com> wrote:
> > >>> _FINAL_ALL_DEPENDENCIES includes _FINAL_PATCH_DEPENDENCIES which may
> > >>> not need to be selected in kconfig in cases where building is not
> > >>> desired, so only check that _FINAL_DEPENDENCIES are selected.
[--SNIP--]
> >   I think the patch is wrong though. A patch dependency is still a dependency,
> > so the package should be selected.
> Well I thought selection was to enable a package build, I thought selection
> was only for packages that need their build executed.

No, they are needed for all target packages.

> >   The problem is that you add two target packages which have no Config.in. That
> > is simply a no go: target packages should always have a Config.in symbol.
> Well we don't want to actually build the package, only extract+patch
> it, I wasn't
> sure how else to do that.

From what we discussed on IRC the other day, I thought I had udnerstood
that qt5webengine-chromium-catapult was a hot package, that's why I
suggested it to not have a Config.in entry.

However, I might have misunderstood, and it really is a target package,
so it needs to have a Config.in entry, even one that does not have a
prompt

we already have for a few packages in that situation, that are not
user-selectable but are still target packages: glibc, mesa3d-headers,
bluez5_utils-headers and a few others... Most interestingly, though, is
that we already have a similar situation with linux extensions: they are
packages that are present only to provide files that are copied into the
kernel tree, like ev3dev-linux-drivers or fbtft.

So, in your case, I believe that qt5webengine-chromium-catapult.mk is
basically correct, but a simple Config.in is missing:

    # Not user-selectable, it is only needed to be copied
    # into the chromium source tree.
    config BR2_PACKAGE_QT5WEBENGINE_CHROMIUM_CATAPULT
        bool

And then you should not have any issue anymore with patch-depednencies.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot