<feed xmlns='http://www.w3.org/2005/Atom'>
<title>mirrors/arti.git/.gitlab-ci.yml, branch arti-v1.1.9</title>
<subtitle>mirror of https://gitlab.torproject.org/tpo/core/arti
</subtitle>
<id>http://git.dilluti0n.com/mirrors/arti.git/atom?h=arti-v1.1.9</id>
<link rel='self' href='http://git.dilluti0n.com/mirrors/arti.git/atom?h=arti-v1.1.9'/>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/'/>
<updated>2023-09-18T13:42:58Z</updated>
<entry>
<title>CI: Remove amd64 tags</title>
<updated>2023-09-18T13:42:58Z</updated>
<author>
<name>Ian Jackson</name>
<email>ijackson@chiark.greenend.org.uk</email>
</author>
<published>2023-09-18T13:42:58Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=f300fe459bb9674d26fd34f56c016794adfcb008'/>
<id>urn:sha1:f300fe459bb9674d26fd34f56c016794adfcb008</id>
<content type='text'>
We have a new runner, and apparently to use it better we want to
remove these tags.

TBH this seems rather unprincipled.  In theory, shouldn't we tag with
`amd64` all the jobs that we definitely want to run on that arch, and
the others can run on whatever is the default?  Then in the glorious
distant RISCV future everything will run on riscv64 by default?  But
maybe we wouldn't want to switch that as part of the infrastructure
anyway?

Anyway, for discussion see
  https://gitlab.torproject.org/tpo/tpa/team/-/issues/41295#note_2943129
</content>
</entry>
<entry>
<title>CI: Ensure libssl3 is installed in shadow and chutney tests</title>
<updated>2023-09-05T13:20:47Z</updated>
<author>
<name>Jim Newsome</name>
<email>jnewsome@torproject.org</email>
</author>
<published>2023-09-05T13:20:47Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=acf87aa13a969fdab957fd193824f078971757bf'/>
<id>urn:sha1:acf87aa13a969fdab957fd193824f078971757bf</id>
<content type='text'>
</content>
</entry>
<entry>
<title>CI: Ensure libsqlite3 is installed in shadow and chutney tests</title>
<updated>2023-09-05T13:07:00Z</updated>
<author>
<name>Jim Newsome</name>
<email>jnewsome@torproject.org</email>
</author>
<published>2023-09-05T13:07:00Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=63351fc1d6c0e4451b1c37f5f5fbab0ac7c430be'/>
<id>urn:sha1:63351fc1d6c0e4451b1c37f5f5fbab0ac7c430be</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Update CI to use Debian bookworm instead of bullseye</title>
<updated>2023-08-29T14:08:41Z</updated>
<author>
<name>Jim Newsome</name>
<email>jnewsome@torproject.org</email>
</author>
<published>2023-08-25T19:16:33Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=95067549c6548023929144b0505ade71907274bf'/>
<id>urn:sha1:95067549c6548023929144b0505ade71907274bf</id>
<content type='text'>
</content>
</entry>
<entry>
<title>shadow test: use libigraph from bullseye</title>
<updated>2023-08-29T14:08:37Z</updated>
<author>
<name>Jim Newsome</name>
<email>jnewsome@torproject.org</email>
</author>
<published>2023-08-25T21:42:42Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=26ac5eea78b5c513349db8edadb0b192138d9359'/>
<id>urn:sha1:26ac5eea78b5c513349db8edadb0b192138d9359</id>
<content type='text'>
Workaround for https://github.com/shadow/tgen/issues/44
</content>
</entry>
<entry>
<title>CI: Replace all uses of `debian:stable-slim` with `bookworm`</title>
<updated>2023-08-25T16:29:31Z</updated>
<author>
<name>Ian Jackson</name>
<email>ijackson@chiark.greenend.org.uk</email>
</author>
<published>2023-08-25T15:48:51Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=4a9724733b3fe470eaecdb2ab9e6c83e0cf99a8a'/>
<id>urn:sha1:4a9724733b3fe470eaecdb2ab9e6c83e0cf99a8a</id>
<content type='text'>
Thesee jobs are ones which don't interact with our Rust executables.

Since `debian:stable-slim` has already been updated I believe, I think
this actually makes no difference right now.  But it will prevent us
unexpectedly picking up Debian updates.
</content>
</entry>
<entry>
<title>CI: Replace all uses of `rust:latest` image with `rust:bullseye`</title>
<updated>2023-08-25T16:29:23Z</updated>
<author>
<name>Ian Jackson</name>
<email>ijackson@chiark.greenend.org.uk</email>
</author>
<published>2023-08-25T15:42:04Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=7f70ab83c3ebeea91f0d33e1f24167a9ded8093a'/>
<id>urn:sha1:7f70ab83c3ebeea91f0d33e1f24167a9ded8093a</id>
<content type='text'>
`rust:latest` updates to a new Debian willy-nilly, which broke our CI
due to it being an uncontrolled dependency.
</content>
</entry>
<entry>
<title>CI: cargo install caching: Use cache/ directory, per-image</title>
<updated>2023-08-25T16:28:16Z</updated>
<author>
<name>Ian Jackson</name>
<email>ijackson@chiark.greenend.org.uk</email>
</author>
<published>2023-08-25T15:59:04Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=f45e53b15f0fd142a5403ed58af6a4a471f0fbf4'/>
<id>urn:sha1:f45e53b15f0fd142a5403ed58af6a4a471f0fbf4</id>
<content type='text'>
This avoids us trying to run binaries built on one image, on a
different one.  I believe that this is supposed to work according to
Rust doctrine, but empirically it doesn't.
</content>
</entry>
<entry>
<title>CI: Use apt, not pip, to install python packages.</title>
<updated>2023-08-25T12:17:54Z</updated>
<author>
<name>Nick Mathewson</name>
<email>nickm@torproject.org</email>
</author>
<published>2023-08-25T12:17:54Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=381bb3118a708baba8c70b0ffe749f280e6eb8a4'/>
<id>urn:sha1:381bb3118a708baba8c70b0ffe749f280e6eb8a4</id>
<content type='text'>
Debian's pip3 now apparently complains when root is running it,
which breaks our old coverage CI.
</content>
</entry>
<entry>
<title>CI: Remove unneeded install of git in maint-checks</title>
<updated>2023-08-09T16:16:58Z</updated>
<author>
<name>Ian Jackson</name>
<email>ijackson@chiark.greenend.org.uk</email>
</author>
<published>2023-08-09T13:31:13Z</published>
<link rel='alternate' type='text/html' href='http://git.dilluti0n.com/mirrors/arti.git/commit/?id=879a3e8acf78194d9dfabac1ce2ee32f238af3a9'/>
<id>urn:sha1:879a3e8acf78194d9dfabac1ce2ee32f238af3a9</id>
<content type='text'>
This was added in 9357a8fd6b22 "ci: add shebang to the GitLab CI" as
part of !990 to the `maint-checks` job; but the actual additional
check was added to the `doc-features` job (by mistake, fixed in
!1490); and, that shebang check script doesn't need git anyway.
</content>
</entry>
</feed>
