From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2152#note_1179932973
@Conan_Kudo - the concern with that approach is, RHEL is not staffed to
support all new technologies, we only support a small subset of what Fedora
enables. We do not want to get our
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2158#note_1179935350
Unfortunate. Let's discuss this at our weekly meeting.
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2165#note_1188537102
@jmflinuxtx - I believe this should catch the failed config MRs during the
cronjob.
___
kernel mailing list -- kernel@lists.fedoraproject.org
From: Don Zickus
Check for kernel config git-push failures
An automated nightly job is run to merge the upstream tree and check for
new configs to be set. The new configs are automatically added to the
redhat/configs/pending-ark directory and pushed. A second script then
comes along and walks
From: Don Zickus
Check for kernel config git-push failures
An automated nightly job is run to merge the upstream tree and check for
new configs to be set. The new configs are automatically added to the
redhat/configs/pending-ark directory and pushed. A second script then
comes along and walks
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2160#note_1181763580
@prarit I think we should be ok with folks providing arch specific stuff.
Towards GA we can evaluate_configs to move the configs higher if most arches
support the option. Then run
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2236#note_1252596373
/unblock
hehe.
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email to kernel-le...@lists.fedoraproject.org
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2236#note_1252599534
@ptalbert @jwilsonrh - um help me unblock this thread?
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2236#note_1252887556
/me wears the cone of shame.
Thanks @ptalbert @jwilsonrh !
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2135#note_1164129535
Silly nitpick, seeing that BTF is a default technology for RH, it wasn't
obvious to me which arch/variant would _not_ have it enabled as I thought that
was _the_ bug. Once I figured out
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2144#note_1171514727
Oddly commenting that out, allows the dist-configs-check target to run and
successfully generate the kernel-*.configs that are failing to copy over.
Still investigating
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2144#note_1171535058
@jtoppins_redhat - found it. https://gitlab.com/cki-project/kernel-
ark/-/blob/os-build/redhat/Makefile#L22-26
You are being bitten by @prarit RHTEST checker. It looks for a '-n
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2144#note_1171547161
@prarit @jtoppins_redhat - this change should fix it
-ifneq (,$(findstring n,$(MAKEFLAGS)))
+ifneq (,$(findstring n,$(filter-out --%,$(MAKEFLAGS
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2149#note_1172697891
@prarit - I don't think it will because we are running a 'cat' command not a
'git' command.
___
kernel mailing list -- kernel
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2370#note_1318936416
@jmflinuxtx - this should fix cki's problem. are you ok with tagging this in
'include in release'?
@scweaver @mh21 - this works around the cki problem and is probably more
correct
From: Don Zickus
Use RHJOBS for create-tarball
rpm-4.18.1 recently modified how _smp_mflags is defined and it breaks
tarball creation: rpm --eval %{_smp_mflags} because the output is
'-j${RPM_BUILD_NCPUS}' instead of '-j8'.
While that is being worked on, let's simplify things and use what we
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2370#note_1318986202
@mh21 - this will be single thread for cki unless cki uses 'make -j dist-srpm'
(no number to default to nproc).
___
kernel mailing list
From: Don Zickus
Use _smp_build_nthreads for rpm eval
rpm-4.18.1 recently modified how _smp_mflags is defined and it breaks
tarball creation: rpm --eval %{_smp_mflags} because the output is
'-j${RPM_BUILD_NCPUS}' instead of '-j8'.
While that is being worked on, let's simplify things and use
From: Don Zickus
Use _smp_build_nthreads for rpm eval
rpm-4.18.1 recently modified how _smp_mflags is defined and it breaks
tarball creation: rpm --eval %{_smp_mflags} because the output is
'-j${RPM_BUILD_NCPUS}' instead of '-j8'.
While that is being worked on, let's simplify things and use
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2370#note_1319161685
ugh nproc --all is wrong i believe
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email to kernel-le
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2370#note_1319106161
@mh21 - the man pages say nproc returns the number of cpus for the _process_
which works well for containers. Contrast this with /proc/cpuinfo which gives
all available cpus
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2375#note_1321278887
Hmm, as an end user it seems odd to have to know what cgroup version and other
quirks are on the system to make correctly calculations. I would think that
would be part of nproc's
From: Don Zickus
Make RHJOBS container friendly
RHJOBS uses 'nproc --all' which is not friendly to process like
containers that use taskset. Instead just use 'nproc'. Folks can
override this with rhpkg.mk file if they choose.
Signed-off-by: Don Zickus
diff --git a/redhat/Makefile b/redhat
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2344#note_1297508006
@clrkwllms - the only tricky part with this change is around the rebase.
originally we put the change in a place that minimized the GA rebase merge
conflicts. This may cause a conflict
From: Don Zickus
Remove cc lines from automatic configs
The kernel-webhook is smarter and knows who the right reviewers are.
The webhooks can also dynamically update them. Just remove the static
cc assisgnment in the nightly automated config generation.
Signed-off-by: Don Zickus
diff --git
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2458#note_1375015427
DevelDir I believe is a CKI env var? @veruu @mh21 ?
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2386#note_1341364121
@prarit - ah ok. I misunderstood how the version worked. maximum rhel
version wasn't obvious to me what that meant. It makes sense
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2386#note_1341177765
Also by using RHEL_MAJOR you have packages kernel-redhat-0.1.x.el9 and kenrel-
redhat-1.1.x.el10 which are basically identical other than configs. What are
the customer expectations
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2386#note_1341172348
@jstancek - so for rhel-9.6 we can expect a kernel package name to be 'kernel-
redhat-0.1.x.el9'? Does anyone intuitively trust a version of 0.X? Should we
start at 1? Just a nitpick
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2386#note_1341362416
This is also something LEAP tests for and would catch if it is a problem. I
am curious too. Though I thought fedora upgrades would have the same problem
though? You update to latest
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2386#note_1342932614
I am ok with the thread being closed.
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email to kernel-le
From: Don Zickus
Remove scmversion from kernel.spec.template
The spec.template file was touching scmversion. The use case of
scmversion was removed from upstream in 6.3. Let's remove it from our
source tree.
Signed-off-by: Don Zickus
diff --git a/redhat/kernel.spec.template b/redhat
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/1617#note_1267368540
Closing, no activity.
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email to kernel-le
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2278#note_1276995815
git grep 'pending-ark' redhat/
shows redhat/scripts/genspec/gen_config_patches.sh
Can you update that too?
___
kernel mailing list -- kernel
From: Don Zickus
CI: ark-create-release: Add a robust check if nothing changed
There are rare cases where upstream hasn't updated and no new 'include
in release' patches were created. The nightly cron job may stumble upon
no new changes. This is blocked by the failure of 'make
dist-release
From: Don Zickus
CI: sync ark-create-release push commands with output
A user can run the ci scripts and either auto push the changes or not.
If the changes are not pushed, a message is printed explaining the steps
needed to properly push the changes.
Over time the real push and the output got
From: Don Zickus
CI: Call script directly
Originally I thought it would be safer to use a Makefile to control the
variables to the ci scripts. Now I realize I lose the ability to pass
arguments unless I use environment variables.
Let's call the script directly in preparation for use arguments
From: Don Zickus
CI: Introduce simple environment script
Simple code re-use across the ci scripts. The same variables and setup
functions are used. Update the ci scripts to source the new script.
This changes the API to ark-create-release.sh, so modify the call to the
script to accept
From: Don Zickus
CI: Remove legacy tag check cruft
Originally code was added to handle the script being repeatedly called
by detecting an existing tag and aborting to prevent a mess.
Over the years we removed the ark-patches, did more git merges and
strengthened dist-release. This simplified
From: Don Zickus
CI: ark-update-configs: misc changes
Simple changes in preparation to help with readability with later
changes.
No behavior changes.
Signed-off-by: Don Zickus
diff --git a/redhat/scripts/ci/ark-update-configs.sh
b/redhat/scripts/ci/ark-update-configs.sh
index blahblah
From: Don Zickus
CI: ark-update-configs: sync push command and output
A rather large change that just moves everything into variables. The
end result is the same except the push command is also echo'd before
being eval'd. This allows the script to keep the command and output in
sync
From: Don Zickus
CI: Move os-build tracking to common area
Future changes will use also use os-build, move the tracking to a common
area to reduce code duplication.
Signed-off-by: Don Zickus
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index blahblah..blahblah 100644
--- a/.gitlab-ci.yml
From: Don Zickus
CI: Remove stale TAG and Makefile cruft
The TAG variable turned into _TAG but the ci scripts were never updated
to reflect that. No one noticed for over a year, therefore it must not
be important.
Also dist-merge-upstream had a branch check as a safety valve to prevent
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2518
These changes remove a lot of legacy cruft and simplify reading of the code.
The goal is to make it easier to understand what is going on and run things
locally. Because of the code
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2517
A collection of simple ci scripts cleanups in preparation for future
automation and easy maintenance.
Signed-off-by: Don Zickus
---
redhat/Makefile | 12 ++--
.gitlab-ci.yml
From: Don Zickus
CI: Move os-build tracking to common area
Future changes will use also use os-build, move the tracking to a common
area to reduce code duplication.
Signed-off-by: Don Zickus
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index blahblah..blahblah 100644
--- a/.gitlab-ci.yml
From: Don Zickus
CI: Remove stale TAG and Makefile cruft
The TAG variable turned into _TAG but the ci scripts were never updated
to reflect that. No one noticed for over a year, therefore it must not
be important.
Also dist-merge-upstream had a branch check as a safety valve to prevent
From: Don Zickus
CI: Call script directly
Originally I thought it would be safer to use a Makefile to control the
variables to the ci scripts. Now I realize I lose the ability to pass
arguments unless I use environment variables.
Let's call the script directly in preparation for use arguments
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2517#note_1455506742
probably not. just legacy cruft. will remove it.
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email
From: Don Zickus
CI: Remove stale TAG and Makefile cruft
The TAG variable turned into _TAG but the ci scripts were never updated
to reflect that. No one noticed for over a year, therefore it must not
be important.
Also dist-merge-upstream had a branch check as a safety valve to prevent
From: Don Zickus
CI: Call script directly
Originally I thought it would be safer to use a Makefile to control the
variables to the ci scripts. Now I realize I lose the ability to pass
arguments unless I use environment variables.
Let's call the script directly in preparation for use arguments
From: Don Zickus
CI: Move os-build tracking to common area
Future changes will use also use os-build, move the tracking to a common
area to reduce code duplication.
Signed-off-by: Don Zickus
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index blahblah..blahblah 100644
--- a/.gitlab-ci.yml
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2517
A collection of simple ci scripts cleanups in preparation for future
automation and easy maintenance.
Signed-off-by: Don Zickus
---
redhat/Makefile | 12 ++--
.gitlab-ci.yml
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2153#note_1455325138
@msalter - they are questions with this MR that haven't been answered. And
folks are not sure if this makes sense. Closing for now. Feel free to re-
open with answers if you still
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2518#note_1467409482
this was addressed in ci-cleanups part1 and committed with the suggestion.
Resolving thread.
___
kernel mailing list -- kernel
From: Don Zickus
CI: sync ark-create-release push commands with output
A user can run the ci scripts and either auto push the changes or not.
If the changes are not pushed, a message is printed explaining the steps
needed to properly push the changes.
Over time the real push and the output got
From: Don Zickus
CI: ark-update-configs: sync push command and output
A rather large change that just moves everything into variables. The
end result is the same except the push command is also echo'd before
being eval'd. This allows the script to keep the command and output in
sync
From: Don Zickus
CI: Remove legacy tag check cruft
Originally code was added to handle the script being repeatedly called
by detecting an existing tag and aborting to prevent a mess.
Over the years we removed the ark-patches, did more git merges and
strengthened dist-release. This simplified
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2519
This change uses the update git push changes for ark-create-release.sh. The
behavior moves the git push for gitlab-ci.yml to the script itself. No
expected behavior fallout
From: Don Zickus
CI: Introduce simple environment script
Simple code re-use across the ci scripts. The same variables and setup
functions are used. Update the ci scripts to source the new script.
This changes the API to ark-create-release.sh, so modify the call to the
script to accept
From: Don Zickus
CI: ark-create-release: Add a robust check if nothing changed
There are rare cases where upstream hasn't updated and no new 'include
in release' patches were created. The nightly cron job may stumble upon
no new changes. This is blocked by the failure of 'make
dist-release
From: Don Zickus
CI: ark-update-configs: misc changes
Simple changes in preparation to help with readability with later
changes.
No behavior changes.
Signed-off-by: Don Zickus
diff --git a/redhat/scripts/ci/ark-update-configs.sh
b/redhat/scripts/ci/ark-update-configs.sh
index blahblah
From: Don Zickus
CI: rawhide_release: switch to using script to push
Currently the gitlab-ci.yml script pushes the gitlab changes. Switch to
using the script to make testing and debugging easier.
No functional changes, but script pushing is minimally tested.
Signed-off-by: Don Zickus
diff
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2518
These changes remove a lot of legacy cruft and simplify reading of the code.
The goal is to make it easier to understand what is going on and run things
locally. Because of the code
From: Don Zickus
CI: Introduce simple environment script
Simple code re-use across the ci scripts. The same variables and setup
functions are used. Update the ci scripts to source the new script.
This changes the API to ark-create-release.sh, so modify the call to the
script to accept
From: Don Zickus
CI: Remove legacy tag check cruft
Originally code was added to handle the script being repeatedly called
by detecting an existing tag and aborting to prevent a mess.
Over the years we removed the ark-patches, did more git merges and
strengthened dist-release. This simplified
From: Don Zickus
CI: ark-update-configs: sync push command and output
A rather large change that just moves everything into variables. The
end result is the same except the push command is also echo'd before
being eval'd. This allows the script to keep the command and output in
sync
From: Don Zickus
CI: sync ark-create-release push commands with output
A user can run the ci scripts and either auto push the changes or not.
If the changes are not pushed, a message is printed explaining the steps
needed to properly push the changes.
Over time the real push and the output got
From: Don Zickus
CI: ark-update-configs: misc changes
Simple changes in preparation to help with readability with later
changes.
No behavior changes.
Signed-off-by: Don Zickus
diff --git a/redhat/scripts/ci/ark-update-configs.sh
b/redhat/scripts/ci/ark-update-configs.sh
index blahblah
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2605#note_1475031711
I believe @prarit was referring to weak-updates which is a RHEL thing for 3rd
party drivers. Thanks @knurd42 for the simple test and confirming the basics
still work
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2575#note_1463813068
@mh21 - can i use this for scheduled jobs too? Right now scheduled job
failures are only sent to owner of scheduled job (ie Justin
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2458#note_1423778786
Yes, thanks for the explaination @dvlasenk !
___
kernel mailing list -- kernel@lists.fedoraproject.org
To unsubscribe send an email to kernel
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2502#note_1422351094
@ecurtin - you would need the RHEL team to sign up to maintain it. And if it
requires other config options too, you would also need those RHEL teams. I
would poke @msalter, I believe
From: Don Zickus
[CI] add exit 0 to the end of CI scripts
Both of the CI scripts (ark-create-release and ark-update-configs) end
on a conditional that is usually non-zero in the normal case. That
leads that non-zero status to be the exit code of the overall script and
thus tells gitlab
From: Don Zickus
[CI] Make sure we are on correct branch before running script
In commit db7320bd04bf53f61a1fde45c9ef5cd2b1666d42, I tried to optimize
things by move the 'git checkout origin/os-build' to the common setup
section. The side affect of this move meant that the last git checkout
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2606#note_1509536950
/block
temporarily blocking as I have heard hesitation about if there is anything
else to do here. Let's discuss this at the weekly meeting.
I am wondering about the data piece
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2607#note_1509449222
@ptalbert - we have a cs9 webhook that detects an MR rebase and keeps the acks
if it detects no changes, right? I thought it would apply here too
From: Don Zickus
CI: rawhide_release: switch to using script to push
Currently the gitlab-ci.yml script pushes the gitlab changes. Switch to
using the script to make testing and debugging easier.
No functional changes, but script pushing is minimally tested.
Signed-off-by: Don Zickus
diff
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2504#note_1438601446
hi @e-milne, for starters, i would recommend wrapping this with "#ifdef
CONFIG_RHEL_DIFFERENCES", so @jmflinuxtx can easily remove it for fedora
systems.
is there a pa
From: Don Zickus on gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2518#note_1440561343
@ezulian - I understand the concern. My main reason for this flexibility (and
initially when we wrote this years ago) was the ability to use a different
area to test. Currently my
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2519
This change uses the update git push changes for ark-create-release.sh. The
behavior moves the git push for gitlab-ci.yml to the script itself. No
expected behavior fallout
From: Don Zickus
CI: Remove stale TAG and Makefile cruft
The TAG variable turned into _TAG but the ci scripts were never updated
to reflect that. No one noticed for over a year, therefore it must not
be important.
Also dist-merge-upstream had a branch check as a safety valve to prevent
From: Don Zickus
CI: Introduce simple environment script
Simple code re-use across the ci scripts. The same variables and setup
functions are used. Update the ci scripts to source the new script.
This changes the API to ark-create-release.sh, so modify the call to the
script to accept
From: Don Zickus
CI: Call script directly
Originally I thought it would be safer to use a Makefile to control the
variables to the ci scripts. Now I realize I lose the ability to pass
arguments unless I use environment variables.
Let's call the script directly in preparation for use arguments
From: Don Zickus
CI: Move os-build tracking to common area
Future changes will use also use os-build, move the tracking to a common
area to reduce code duplication.
Signed-off-by: Don Zickus
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index blahblah..blahblah 100644
--- a/.gitlab-ci.yml
From: Don Zickus
CI: Call script directly
Originally I thought it would be safer to use a Makefile to control the
variables to the ci scripts. Now I realize I lose the ability to pass
arguments unless I use environment variables.
Let's call the script directly in preparation for use arguments
From: Don Zickus
CI: ark-update-configs: misc changes
Simple changes in preparation to help with readability with later
changes.
No behavior changes.
Signed-off-by: Don Zickus
diff --git a/redhat/scripts/ci/ark-update-configs.sh
b/redhat/scripts/ci/ark-update-configs.sh
index blahblah
From: Don Zickus
CI: ark-update-configs: sync push command and output
A rather large change that just moves everything into variables. The
end result is the same except the push command is also echo'd before
being eval'd. This allows the script to keep the command and output in
sync
From: Don Zickus
CI: Remove legacy tag check cruft
Originally code was added to handle the script being repeatedly called
by detecting an existing tag and aborting to prevent a mess.
Over the years we removed the ark-patches, did more git merges and
strengthened dist-release. This simplified
From: Don Zickus
CI: Introduce simple environment script
Simple code re-use across the ci scripts. The same variables and setup
functions are used. Update the ci scripts to source the new script.
This changes the API to ark-create-release.sh, so modify the call to the
script to accept
From: Don Zickus
CI: Move os-build tracking to common area
Future changes will use also use os-build, move the tracking to a common
area to reduce code duplication.
Signed-off-by: Don Zickus
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index blahblah..blahblah 100644
--- a/.gitlab-ci.yml
From: Don Zickus
CI: Remove stale TAG and Makefile cruft
The TAG variable turned into _TAG but the ci scripts were never updated
to reflect that. No one noticed for over a year, therefore it must not
be important.
Also dist-merge-upstream had a branch check as a safety valve to prevent
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2518
These changes remove a lot of legacy cruft and simplify reading of the code.
The goal is to make it easier to understand what is going on and run things
locally. Because of the code
From: Don Zickus
CI: Call script directly
Originally I thought it would be safer to use a Makefile to control the
variables to the ci scripts. Now I realize I lose the ability to pass
arguments unless I use environment variables.
Let's call the script directly in preparation for use arguments
From: Don Zickus
CI: sync ark-create-release push commands with output
A user can run the ci scripts and either auto push the changes or not.
If the changes are not pushed, a message is printed explaining the steps
needed to properly push the changes.
Over time the real push and the output got
From: Don Zickus
CI: ark-create-release: Add a robust check if nothing changed
There are rare cases where upstream hasn't updated and no new 'include
in release' patches were created. The nightly cron job may stumble upon
no new changes. This is blocked by the failure of 'make
dist-release
From: Don Zickus
CI: Remove stale TAG and Makefile cruft
The TAG variable turned into _TAG but the ci scripts were never updated
to reflect that. No one noticed for over a year, therefore it must not
be important.
Also dist-merge-upstream had a branch check as a safety valve to prevent
From: Don Zickus
CI: Move os-build tracking to common area
Future changes will use also use os-build, move the tracking to a common
area to reduce code duplication.
Signed-off-by: Don Zickus
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index blahblah..blahblah 100644
--- a/.gitlab-ci.yml
From: Don Zickus on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2517
A collection of simple ci scripts cleanups in preparation for future
automation and easy maintenance.
Signed-off-by: Don Zickus
---
redhat/Makefile | 12 ++--
.gitlab-ci.yml
From: Don Zickus
CI: rawhide_release: switch to using script to push
Currently the gitlab-ci.yml script pushes the gitlab changes. Switch to
using the script to make testing and debugging easier.
No functional changes, but script pushing is minimally tested.
Signed-off-by: Don Zickus
diff
601 - 700 of 770 matches
Mail list logo