Re: [Buildroot] [PATCH v2 00/18] apply patches with fuzz factor 0

From: "Yann E. MORIN" <yann.morin.1998@free.fr>
To: Dario Binacchi <dario.binacchi@amarulasolutions.com>
Cc: "Mahyar Koshkouei" <mahyar.koshkouei@gmail.com>,
	"Sergey Matyukevich" <geomatsi@gmail.com>,
	"Romain Naour" <romain.naour@gmail.com>,
	"Bernd Kuhls" <bernd@kuhls.net>,
	"Jagan Teki" <jagan@amarulasolutions.com>,
	"Michał Łyszczek" <michal.lyszczek@bofc.pl>,
	buildroot@buildroot.org,
	"Giulio Benetti" <giulio.benetti@benettiengineering.com>,
	michael@amarulasolutions.com, linux-amarula@amarulasolutions.com,
	"Fabrice Fontaine" <fontaine.fabrice@gmail.com>,
	"Julien Corjon" <corjon.j@ecagroup.com>
Subject: Re: [Buildroot] [PATCH v2 00/18] apply patches with fuzz factor 0
Date: Tue, 25 Jun 2024 22:58:41 +0200	[thread overview]
Message-ID: <ZnsvgdItWs8iXni4@landeda> (raw)
In-Reply-To: <20240522070238.3282121-1-dario.binacchi@amarulasolutions.com>

Dario, All,

On 2024-05-22 09:02 +0200, Dario Binacchi spake thusly:
> The series will allow applying package patches with a fuzz factor of 0
> instead of 2 (the patch command default value).
> By setting the maximum fuzz factor to 0, we avoid that patches which
> cannot be applied are incorrectly reported as valid, with positive
> side-effects on version bumps.
> 
> To avoid regressions, it was necessary to fix all those patches that
> were applicable with a fuzz factor of 1 or 2 before implementing this
> change. The series was tested on commit 47078cc11862 ("package/zxing-cpp:
> add options for enabling readers and/or writers") with a script that
> executed the make legal-info command for all defconfigs contained in
> Buildroot. Then, for all the affected packages, the compilation was
> executed.

Thanks a lot for all those efforts! 👍

As Arnout said, it is not unexpected that we missed some, and that new
patches added since then would not be fuzz-zero-ready.

So, I came up with a little bit of shell trickedry to start and test all
our patches (or at least the vast majority of them):

    $ cat foo.lst

    # Ignore boot/ and linux/ as they are too special...
    find package/ -type f -name '*.mk' -printf '%h\n' \
    |sort -u \
    |while read dir; do
        [ -e "${dir}/Config.in" ] || continue
        ( find "${dir}" -type f -name '*.patch' -print -quit |grep -q . ) || continue
        pkg="$( basename "${dir}" )"
        [ -e "${dir}/${pkg}.mk" ] || continue
        sed -r -e '/^\$\(eval \$\((host-)?[^-]+-package\)\)$/!d; s//\1'"${pkg}"'-patch/' "${dir}/${pkg}.mk"
    done \
    |grep -v -E 'am335x-pru-package|asterisk|aufs-util|aumix|cgic'

    $ make defconfig

    $ ./utils/config --set-str BR2_BACKUP_SITE ""

    $ make $(. ./foo.lst)

I stopped adding exceptions when I reached cgic...

Exceptions:
  - Patch fuzz:
      - asterisk

  - Others:
      - am335x-pru-package: hash error, github-generated
      - aufs-util: needs a kernel, patches unknwon
      - aumix: 403 forbidden, patches OK, tested with s.b.o.
      - bsdiff: 403 forbidden, patches OK, tested with s.b.o.
      - cgic: hash error, upstream is dead (parking site)

So, I found more download issues than I found patch fuzz issues... I'm
not sure if that's a good or a bad thing? ;-]

Then linux/ and boot/ will have to be checked "manually"...

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