fix: make renovate extractVersion actually apply to README tags and HCL sql pins #11
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "fix/renovate-readme-extractversion"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Two bugs, both reproduced and verified with local renovate runs:
The PGDG README image-tag manager applied a fixed
extractVersionTemplate (^[^-+]+), which extracts 1.6.7 from
postgresql-18-cron 1.6.7-3.pgdg13+1 while the image tag tracks the
control default_version (1.6). Renovate rewrote READMEs to the
nonexistent pg-cron:1.6.7-pg18-trixie tag (#10). Templates take
precedence over captured groups, so the comment-level extractVersion
is now honored via a conditional template, and the three pg-cron
image-tag comments carry extractVersion=^(?\d+.\d+).
metadata.hcl renovate comments escaped backslashes (\d), producing
regexes that match a literal backslash and therefore never match any
version: every sql pin was silently frozen. Proven by downgrading hll
sql to 2.20 (no update proposed) and unfreezing (2.21 proposed).
postgis was live damage: package had moved to 3.6.4 while sql stayed
frozen at 3.6.3, so the published tag lied and the README pointed to
a nonexistent 3.6.4 tag. All HCL comments now use single backslashes
and postgis sql is aligned to 3.6.4 (control default_version verified
in a rebuilt image).
Co-Authored-By: Claude Fable 5 noreply@anthropic.com