<feed xmlns='http://www.w3.org/2005/Atom'>
<title>buildbot, branch v25</title>
<subtitle>LEDE buildbot configuration</subtitle>
<id>https://git.openwrt.org/buildbot/atom?h=v25</id>
<link rel='self' href='https://git.openwrt.org/buildbot/atom?h=v25'/>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/'/>
<updated>2025-12-14T05:09:49Z</updated>
<entry>
<title>phase1: add support for overriding feeds host</title>
<updated>2025-12-14T05:09:49Z</updated>
<author>
<name>Petr Štetiar</name>
</author>
<published>2025-12-12T21:41:01Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=057e8adc9d49eda73499b5b771cba5df5e8f80d1'/>
<id>urn:sha1:057e8adc9d49eda73499b5b771cba5df5e8f80d1</id>
<content type='text'>
Currently git.openwrt.org is sometimes overloaded, so lets implement a
mechanism to override the source host for feeds.

This introduces a `feeds_host_override` configuration option. When set,
the buildmaster will temporarily modify `feeds.conf.default` to point to
the alternate host (e.g., GitHub) before updating feeds, and restore the
original configuration afterwards.

This is particularly useful when the primary git server is returning 503
errors:

    Updating feed 'packages' from 'https://git.openwrt.org/feed/packages.git' ...
    Cloning into './feeds/packages'...
    fatal: unable to access 'https://git.openwrt.org/feed/packages.git/': The requested URL returned error: 503
    failed.

Signed-off-by: Petr Štetiar &lt;ynezz@true.cz&gt;
</content>
</entry>
<entry>
<title>phase2: fix signing steps when only apk_key is defined</title>
<updated>2025-12-13T08:19:17Z</updated>
<author>
<name>Petr Štetiar</name>
</author>
<published>2025-12-13T08:02:33Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=b0a3bf3f9b2acf5be391ed1d684e135f1161af19'/>
<id>urn:sha1:b0a3bf3f9b2acf5be391ed1d684e135f1161af19</id>
<content type='text'>
Signing steps are currently skipped if only APK signing is configured,
because phase2 effectively enables signing only when `usign` is present.

Fix this by making `IsSignEnabled` explicitly cover APK signing too.

While at it, refactor the signing checks into dedicated helper functions
`IsUsignEnabled`, `IsApkSigningEnabled`, and `IsGpgSigningEnabled`, and
use them consistently to align phase2 with the phase1 implementation.

Signed-off-by: Petr Štetiar &lt;ynezz@true.cz&gt;
</content>
</entry>
<entry>
<title>phase1: fix CONFIG_SIGNED_PACKAGES for apk signing</title>
<updated>2025-12-12T20:49:48Z</updated>
<author>
<name>Petr Štetiar</name>
</author>
<published>2025-12-12T20:46:05Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=a51561bb76cdd9759c693f99527f88a6287532cf'/>
<id>urn:sha1:a51561bb76cdd9759c693f99527f88a6287532cf</id>
<content type='text'>
Currently `CONFIG_SIGNED_PACKAGES` is only enabled when `usign_key` is
configured. However, we also want to enable it when `apk_key` is
configured, as we support signing APK packages as well.

So lets fix it by adding `IsSignedPackagesEnabled` helper which checks
for both keys and updates the .config generation to use it.

Signed-off-by: Petr Štetiar &lt;ynezz@true.cz&gt;
</content>
</entry>
<entry>
<title>phase1: fix signing steps when only apk_key is defined</title>
<updated>2025-12-11T19:05:20Z</updated>
<author>
<name>Petr Štetiar</name>
</author>
<published>2025-12-11T18:13:05Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=b92996c987fa8d09e927be15e5bde3fc405b55b9'/>
<id>urn:sha1:b92996c987fa8d09e927be15e5bde3fc405b55b9</id>
<content type='text'>
Signing steps are currently skipped if only `apk_key` is configured for a
branch, because `IsSignEnabled` only checks for `usign_key` or `gpg_key`.

Fix this by explicitly checking for `apk_key`.

While at it, refactor the signing checks into dedicated helper functions
`IsApkSigningEnabled` and `IsGpgSigningEnabled` for better readability.

Signed-off-by: Petr Štetiar &lt;ynezz@true.cz&gt;
</content>
</entry>
<entry>
<title>CI(deps): Bump actions/checkout from 4 to 5</title>
<updated>2025-08-12T12:16:13Z</updated>
<author>
<name>dependabot[bot]</name>
</author>
<published>2025-08-12T04:55:28Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=54e80480d2cb1cd4f723d89addf65058b0e56709'/>
<id>urn:sha1:54e80480d2cb1cd4f723d89addf65058b0e56709</id>
<content type='text'>
Bumps [actions/checkout](https://github.com/actions/checkout) from 4 to 5.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](https://github.com/actions/checkout/compare/v4...v5)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-version: '5'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] &lt;support@github.com&gt;</content>
</entry>
<entry>
<title>docker,worker: install python3-cryptography</title>
<updated>2024-11-25T10:12:20Z</updated>
<author>
<name>Robert Marko</name>
</author>
<published>2024-11-25T09:40:03Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=7f966275500d5cd94cc4b70e0e6bb203d6f569ae'/>
<id>urn:sha1:7f966275500d5cd94cc4b70e0e6bb203d6f569ae</id>
<content type='text'>
Addition of OPTEE support along with STM32 added a requirment on
python3-cryptography and without it buildbots for STM32 will fail with:
Checking 'python3-cryptography'... failed.
Checking 'python3-pyelftools'... ok.
optee-os: Please install the Python3 cryptography module
make[3]: *** [/builder/shared-workdir/build/include/prereq.mk:9: prereq] Error 1

So install python3-cryptography to satisfy it.

Signed-off-by: Robert Marko &lt;robimarko@gmail.com&gt;
Link: https://github.com/openwrt/buildbot/pull/61
Signed-off-by: Petr Štetiar &lt;ynezz@true.cz&gt;
</content>
</entry>
<entry>
<title>scripts: signall: skip updating hash of previous kmods/</title>
<updated>2024-11-18T08:08:57Z</updated>
<author>
<name>Christian Marangi</name>
</author>
<published>2024-11-17T15:13:02Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=31eb26b97e7cf8f202c8bd2c65d550edd6e28374'/>
<id>urn:sha1:31eb26b97e7cf8f202c8bd2c65d550edd6e28374</id>
<content type='text'>
Handle case where a sha256sums might contain previous kmods/. In such
case packages.adb for previous kmods/ needs to be skipped as not
included in the sign tar and already signed by previous runs.

Skip is limited to kmods/ entry as those are the only entry expected to
be present in the sha256sums but not included in sign tar.

Signed-off-by: Christian Marangi &lt;ansuelsmth@gmail.com&gt;
</content>
</entry>
<entry>
<title>phase1: correctly exclude kmods files from rsync target upload and prune</title>
<updated>2024-11-18T08:08:56Z</updated>
<author>
<name>Christian Marangi</name>
</author>
<published>2024-11-15T15:41:52Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=a75ce026ca9bd5df828e9e51db08f209a716fe82'/>
<id>urn:sha1:a75ce026ca9bd5df828e9e51db08f209a716fe82</id>
<content type='text'>
There is currently a BUG in the --exclude pattern for the targetupload step.
The /kmods/ directory is actually never excluded. This was never notice
as the rsynclist never actually had kmods entry as make checksum step
was done before moving the kmods to the dedicated directory.

This is caused by the fact that with --files-from, not only the /kmods/
directory needs to be excluded but also every content in it, hence the
additional --exclude "/kmods/**" is required to correcly skip the entry
present in rsynclist.

Signed-off-by: Christian Marangi &lt;ansuelsmth@gmail.com&gt;
</content>
</entry>
<entry>
<title>pahse1: merge sha256sums with remote kmods entry if present</title>
<updated>2024-11-18T08:08:55Z</updated>
<author>
<name>Christian Marangi</name>
</author>
<published>2024-11-15T14:30:09Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=9f22657ef41bb599558aaeffda451353eaf35199'/>
<id>urn:sha1:9f22657ef41bb599558aaeffda451353eaf35199</id>
<content type='text'>
Add additional logic to merge the sha256sums with the remote kmods entry
(in remote sha256sums) already present.

This is to produce a more consistent sha256sums that actually reflect
what is present in the remote server by including also the sha of the
kmods for each kernel version supported.

Signed-off-by: Christian Marangi &lt;ansuelsmth@gmail.com&gt;
</content>
</entry>
<entry>
<title>phase1: remove kmods in target packages if archive is enabled</title>
<updated>2024-11-18T08:08:54Z</updated>
<author>
<name>Daniel Golle</name>
</author>
<published>2024-11-15T14:27:55Z</published>
<link rel='alternate' type='text/html' href='https://git.openwrt.org/buildbot/commit/?id=ec2bd8a07e88f4e251fccb99142bd46c816d62ec'/>
<id>urn:sha1:ec2bd8a07e88f4e251fccb99142bd46c816d62ec</id>
<content type='text'>
OPKG gets confused if kmod packages are present in both, target packages
as well as kernel version specific folder.
Remove them from target packages to make opkg pick the kmods from kmod
archive folder only.

This also affects APK that picks packages from the first entry in the
repository file and doesn't support checking the same package from
multiple entry.

Signed-off-by: Daniel Golle &lt;daniel@makrotopia.org&gt;
[ add --remove-source-files and improve implementation ]
Signed-off-by: Christian Marangi &lt;ansuelsmth@gmail.com&gt;
</content>
</entry>
</feed>
