From: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
To: buildroot@busybox.net
Subject: [Buildroot] Buildroot hackathon day 1 highlights
Date: Sun, 1 Apr 2018 07:49:41 +0200 [thread overview]
Message-ID: <20180401074941.7a36ab0a@windsurf> (raw)
Hello,
Our first day of the hackathon yesterday was very productive, with
good progress made on a number of topics. The main highlights are:
- Maxime Hadjinlian revived his series implementing caching for Git
downloads, which avoids the need to constantly re-clone from
scratch Git repositories. His original series from July 2017 had
already seen a respin by Peter Seiderer, and Maxime has now another
iteration (version 3). His series starts at
http://lists.busybox.net/pipermail/buildroot/2018-March/217144.html,
but unfortunately doesn't have a cover letter.
- Peter Korsgaard reviewed, tweaked and applied the patch series from
Yann E. Morin that reworks how the filesystem images are generated.
Previously, all filesystem images were directly generated from the
contents of output/target, by running under a fakeroot session some
finalization steps (fixing permissions and ownerships, creating
users, etc.) and the tool actually creating the filesystem
image. This logic had two drawbacks: it was not compatible with
parallel creation of filesystem images (parallel in the sense of
top-level parallel build) and it re-ran the finalization steps for
every filesystem image, even though they are common between all
filesystem images. Another issue is that some filesystem types do
specific tweaks, and those tweaks therefore leak into the other
filesystem images that don't need those tweaks.
To solve this, Yann's patch series consists in generating in all
cases, under a fakeroot session, a tarball of the root filesystem,
in which all the common finalization steps have been done.
Then, each filesystem format will, under their own fakeroot
session, extract this tarball in a private location, do their
custom tweaks, and run their specific tool to create the filesystem
image.
This solves the issues mentioned above, but has the downside of
extracting/re-packing the filesystem for each filesystem image,
taking a bit more time.
- Thomas Petazzoni took over Angelo Compagnucci's patch series adding
the Golang package infrastructure. Thomas did a number of further
tweaks and improvements to the package infrastructure, further
reducing the amount of work that remains to be done in each
package, by making the infrastructure a bit more flexible. The
patch series was then reposted to the list, reviewed by Yann, and
finally applied by Arnout Vandecappelle, who did some final
adjustements.
- Peter Korsgaard worked on the 2018.02.x maintenance branch,
backporting a large number of fixes that had been made on the
master branch.
- Ricardo Martincoski, participating remotely to the hackathon,
posted a series that allows check-package to check more
directories/files, and a large number of associated fixes to remove
coding style problems in numerous Buildroot source files. Following
the comments, Ricardo even posted a second iteration of this patch
series, which will most likely be applied soon.
- Thomas Petazzoni took over the oldest patch in patchwork, which
added a <pkg>-show-rrdepends target that recursively displays the
reverse dependencies of a package. This single patch was turned
into a 5 patches patch series that reworks the graph-depends
script, extends it with --quiet and --flat-list options, and
finally adds <pkg>-show-recursive-depends and
<pkg>-show-recursive-rdepends targets.
- And of course, the main part of the hackathon is to look at all the
patches in patchwork that were sitting there for a while, and
taking a decision about them. A large number of patches have been
discussed, and either some comments were made that will require the
original contributor to rework the patch, or the patch was adjusted
and applied.
75 commits have been made, and even though Ricardo added a 36
patches patch series, the patchwork backlog has reduced from ~350
patches to 276 patches at the time of this writing.
We have two more days in this Buildroot hackathon, which will allow to
make even further progress on a number of topics.
If you're interested in seeing a picture of the atmosphere of this
hackathon, see our Twitter post at
https://twitter.com/buildrootorg/status/980081463730241536. You can
see that we are working in a very jungle-style atmosphere, in the very
nice environment kindly sponsored by La Maison from Scaleway
(https://careers.scaleway.com/la-maison/).
Best regards,
Thomas Petazzoni
--
Thomas Petazzoni, CTO, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
reply other threads:[~2018-04-01 5:49 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180401074941.7a36ab0a@windsurf \
--to=thomas.petazzoni@bootlin.com \
--cc=buildroot@busybox.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox