[PATCH] Morph: "except:" bad
by Lars Wirzenius
repo: git://git.baserock.org/baserock/baserock/morph
branch: liw/kill-bare-except
commit: 6375554363c111fda70b7540c3c56a0e9cf1497b
This patch makes ./check output a number of log messages to stderr.
These will go away when we update CoverageTestRunner to its current
upstream version, which automatically hides log messages by unit
tests.
Lars Wirzenius (1):
Stop using bare except: statements
check | 5 +++++
morphlib/app.py | 4 ++--
morphlib/builder2.py | 9 +++++++--
morphlib/cachedrepo_tests.py | 9 +++++----
morphlib/extractedtarball.py | 11 ++++++++---
morphlib/mountableimage.py | 19 ++++++++++++-------
morphlib/plugins/branch_and_merge_plugin.py | 18 ++++++++++--------
morphlib/remoterepocache.py | 10 +++++++---
8 files changed, 56 insertions(+), 29 deletions(-)
--
1.7.10.4
10 years, 6 months
[PATCH 0/2] Support environment variables in ssh
by Tiago Gomes
repo: git://git.baserock.org/delta/openssh
branch: baserock/tiagogomes/openssh-branch
commit: b7fe05106e07c3615fcbd022cf7a3aa5ddcf8143
Tiago Gomes (2):
Make SSH pass over (and make SSHd accept) locale environment
variables.
Fix the environment variable names
ssh_config | 2 ++
sshd_config | 6 ++++--
2 files changed, 6 insertions(+), 2 deletions(-)
--
1.7.10.4
10 years, 7 months
System-branches
by Sam Thursfield
Good morning, baserock-dev readers
There's a discussion on the GNOME desktop-level list at the minute which
is (very tentatively) exploring the idea that major changes in UX / UI
should be prototyped in a branch independent of the 6-month release
cycle:
https://mail.gnome.org/archives/desktop-devel-list/2013-April/msg00211.html
Without going too much into the social aspects, I think a big problem
GNOME has at the minute is that the pipeline from designers coming up
with a great new way of doing things, developers implementing it,
distributions packaging it, users upgrading their machines, testing it
and finding out that the new design fails in the face of a certain edge
case and causes a regression, is way too long.
Baserock and OSTree are the only two pieces of software I'm aware of
which can do system system branching in a useful way. I've not explored
OSTree much but I think it so far lacks the merging between
system-branches that is possible in Baserock. I haven't jumped into the
desktop-devel thread yet because we've not built GNOME in Baserock for a
long time so nobody can try it out.
I think this is a great way of demonstrating system-branches, though.
They're quite tricky to explain and our demos so far have been things
like upgrading Linux from 3.7 to 3.8, but this is looking at them from a
code point of view, and the code in a system is already highly modular.
The modular structure of the code doesn't help people who look at things
from a user experience design point of view, though, and for them being
able to branch the whole system and demonstrate a change as a whole
makes a lot more sense. As the GNOME thread demonstrates it's not just
convenience -- there's a structural problem of not being able to test
and back out big design changes in a way that fits with 6 month release
cycles, and at the moment end-users often become unwilling testers as a
result.
I wrote this email mainly to make sense of what's going on in my head,
but hopefully it's useful to others to relate system-branching in
Baserock to an existing open source project.
Sam
10 years, 7 months
[PATCH 0/3] Morph: Change test suite to use rootfs-tarball only
by Lars Wirzenius
repo: git://git.baserock.org/baserock/baserock/morph
branch: liw/rootfs-tarballs-only
commit: e5e89fcb5f098c39f358f650fec4b9f329024ad4
The Morph test suite was failing for me, and while investigating, I found
that we're still using syslinux-disk system-kinds in tests. This is not
acceptable. We're going to be dropping support for that soon, hopefully
for Baserock 8, so it's time to convert the tests to use rootfs-tarball
only. This patch series does that.
./check --full now passes for me, with these patches.
Lars Wirzenius (3):
Fix tests to use rootfs-tarball system-kinds, instead of disks
Disable make-patch test, since make-patch is broken
Fix run-in-artifact test
scripts/setup-3rd-party-strata | 2 +-
tests.as-root/make-patch.script | 7 ++++++-
tests.as-root/rootfs-tarball-builds-rootfs-and-kernel.script | 6 ------
tests.as-root/run-in-artifact-with-different-artifacts.script | 2 +-
tests.as-root/setup | 4 ++--
tests.as-root/system-overlap.script | 2 +-
tests.branching/edit-updates-stratum-build-depends.stdout | 2 +-
tests.branching/edit-updates-stratum.stdout | 2 +-
tests.branching/setup | 2 +-
tests.branching/tag-creates-commit-and-tag.stdout | 8 ++++----
tests.branching/tag-tag-works-as-expected.stdout | 8 ++++----
tests.branching/tag-works-with-multiple-morphs-repos.script | 2 +-
tests.branching/tag-works-with-multiple-morphs-repos.stdout | 8 ++++----
tests.branching/workflow-petrify.stdout | 6 +++---
tests.merging/rename-stratum.stderr | 2 +-
tests.merging/setup | 2 +-
tests/setup | 2 +-
17 files changed, 33 insertions(+), 34 deletions(-)
--
1.7.10.4
10 years, 7 months
[PATCH 0/3] Add lorry files for building Qt5
by Richard Dale
git@trove.baserock.org:baserock/local-config/lorries
branch: master
commit: 1f48477527c797f8e6994a8e650f90b067c9fcbb
Add lorry files for Qt5 and submodules
Add lorry files for extra X11 repos needed by Qt5
Richard Dale (3):
Add lorry files for qt5 and its submodules
Add some extra x11 lorry files needed for building qt5
Add namespaces to the lorry names for git submodule repos
open-source-lorries/qt5-tools-all.lorry | 78 +++++++++++++++++++++++++++++++
open-source-lorries/util-image.lorry | 6 +++
open-source-lorries/util-keysyms.lorry | 11 +++++
open-source-lorries/util-wm.lorry | 6 +++
4 files changed, 101 insertions(+)
create mode 100644 open-source-lorries/qt5-tools-all.lorry
create mode 100644 open-source-lorries/util-image.lorry
create mode 100644 open-source-lorries/util-keysyms.lorry
create mode 100644 open-source-lorries/util-wm.lorry
--
1.7.10.4
10 years, 7 months
[PATCH 0/3] Changed empty stratum warning to error and added test in tests.build for this case
by Richard Holland
repo: git://git.baserock.org/baserock/baserock/morph
branch: baserock/richardholland/empty-stratum-error
commit: 7c5487a695ab8353e9e98116a7165722cdd9f61e
Richard Holland (3):
Changed empty stratum warning to error
Added new test case for empty strata
Corrections to previous commits
morphlib/builder2.py | 8 +++++++-
tests.build/empty-stratum.exit | 1 +
tests.build/empty-stratum.script | 38 ++++++++++++++++++++++++++++++++++++++
tests.build/empty-stratum.stderr | 1 +
4 files changed, 47 insertions(+), 1 deletions(-)
create mode 100644 tests.build/empty-stratum.exit
create mode 100755 tests.build/empty-stratum.script
create mode 100644 tests.build/empty-stratum.stderr
--
1.7.9-rc0
10 years, 8 months
[PATCH 0/2] Add nfsboot write and configure extensions
by Jonathan Maw
head: f436f85c11f06d3a1a58c64940973ecd786fb107
branch: jonathanmaw/deploy-nfsboot-rebase
repo: ssh://git@trove.baserock.org/baserock/baserock/morph
Actually getting nfsboot to work requires more work than just the write
extension. It also requires a properly configured nfsboot server, the
target device to be configured to boot from that server, and for the
built system to be appropriate to that device.
Jonathan Maw (2):
Add nfsboot configuration extension
Add nfsboot write extension
morphlib/exts/nfsboot.configure | 32 +++++++++
morphlib/exts/nfsboot.write | 146 ++++++++++++++++++++++++++++++++++++++++
2 files changed, 178 insertions(+)
create mode 100755 morphlib/exts/nfsboot.configure
create mode 100755 morphlib/exts/nfsboot.write
--
1.7.11.7
10 years, 8 months