build: ship the dictionary from a committed corpus, refreshed monthly by pull request

This commit is contained in:
tiennm99 committed 2026-10-02 13:46:25 +07:00
1 parent a947f5ce2e
commit f37e3e69f4
13 files changed
+36670 -95

No files matched your search

+8 -14
View File
@@ -1,9 +1,9 @@
# Tests and packaging.
#
# Nothing here downloads the upstream wordlist except the release image
# build. Everything else plays against the small database derived from
# the checked-in word sample, which goes through the same builder the real one
# does.
# Nothing here downloads the upstream wordlist. The image is built from the
# committed corpus, data/dictionary.txt, exactly as it ships; the test suites
# play against the small database derived from the checked-in word sample,
# which goes through the same builder the real one does.
#
# The wire contract has its own workflow: see proto.yml.
name: ci
@@ -159,17 +159,11 @@ jobs:
with:
persist-credentials: false
# On a release the image is built from the real upstream release, which
# is the only job in this file that downloads it. Every other run builds
# the same Dockerfile against the fixture word list, so a broken image is
# caught on the pull request rather than at release time.
# The real dictionary, from the committed corpus: the image every run
# builds is the one that ships, so a corpus the builder rejects fails
# here rather than at deploy time.
- name: Build
run: |
if [ "${{ github.event_name }}" = "release" ]; then
docker build -t noitu:ci .
else
docker build --build-arg FIXTURE_DICT=1 -t noitu:ci .
fi
run: docker build -t noitu:ci .
# CC BY-SA 4.0 applies to the derived wordlist wherever it is
# distributed, and an image is distribution. This is the assertion that