Skip to content

Do not rewrite Docker FROM images built from variables - #1212

Merged
timtebeek merged 2 commits into
mainfrom
tim/worcester-v2
Aug 20, 2026
Merged

Do not rewrite Docker FROM images built from variables#1212
timtebeek merged 2 commits into
mainfrom
tim/worcester-v2

Conversation

@timtebeek

@timtebeek timtebeek commented Aug 20, 2026

Copy link
Copy Markdown
Member

Fixes #1211

What's changed

UpgradeDockerImageVersion built a ChangeFrom sub-recipe for every (image, oldVersion) pair — 170 of them when targeting Java 25.

DockerFrom's matcher substitutes a * for each environment variable and then matches bidirectionally, so for

ARG IMAGE_NAME
ARG IMAGE_TAG
FROM ${IMAGE_NAME}:${IMAGE_TAG}

the image name matched as matchesGlob("openjdk", "*") and the tag as matchesGlob("8*", "*"). The first ChangeFrom in the list therefore fired, the $1 capture resolved to nothing, and the reference was replaced with a hardcoded eclipse-temurin:25.

This replaces the sub-recipe list with a single visitor that:

  • skips any FROM whose image name or tag contains a $, since the value can not be determined statically;
  • parses the leading version from the tag and only bumps it when it sits between 8 and the target version, preserving the suffix (-jre-alpine, -jdk-jammy, ...);
  • maps openjdk/adoptopenjdk to eclipse-temurin;
  • drops any digest pin when it bumps the tag. The digest, not the tag, decides which image is pulled, so eclipse-temurin:25-jre@sha256:<openjdk-11-digest> would keep resolving to the pre-upgrade image.

As a side effect the recipe list of the aggregate migrations shrinks considerably: UpgradeToJava25 drops from 1398 to 488 recipes.

Tests

Added cases covering ${VAR} and $VAR in the image name, the tag, and both; a Dockerfile that mixes a variable FROM with a literal one; unrelated images; and digest pin removal. The existing parameterized cases are unchanged and still pass.

`UpgradeDockerImageVersion` composed a `ChangeFrom` per (image, oldVersion)
pair. `DockerFrom`'s matcher treats an environment variable as a `*` wildcard
and matches bidirectionally, so `FROM ${IMAGE_NAME}:${IMAGE_TAG}` matched every
one of those globs and was rewritten to a hardcoded `eclipse-temurin:25`.

Replace the ~170 sub-recipes with a single visitor that reads the image name and
tag, skips any reference containing a variable, and bumps only tags whose
leading version is between 8 and the target version.
The digest, not the tag, decides which image is pulled, so carrying the old
digest over would keep resolving to the pre-upgrade image.
@timtebeek
timtebeek marked this pull request as ready for review August 20, 2026 13:43
@timtebeek
timtebeek merged commit 8b1f865 into main Aug 20, 2026
1 check passed
@timtebeek
timtebeek deleted the tim/worcester-v2 branch August 20, 2026 13:44
@github-project-automation github-project-automation Bot moved this from In Progress to Done in OpenRewrite Aug 20, 2026
timtebeek added a commit that referenced this pull request Aug 21, 2026
* Upgrade `ARG` default values used in Docker `FROM` instructions

Follow up to #1212, which left any `FROM` built from a variable untouched.
When the variable is a global `ARG` with a literal default, that default can
be upgraded instead, so `ARG java_version=17` used as
`FROM eclipse-temurin:${java_version}` becomes `ARG java_version=25`.

* Do not bump `ARG` defaults shared with images we leave alone

Three follow ups from review:

- A `FROM` whose image we do not upgrade now vetoes the arguments feeding it,
  so `ARG VERSION=11` used by both `eclipse-temurin:${VERSION}` and
  `node:${VERSION}` is left alone rather than turning the latter into `node:25`.
- Drop the digest pin when an argument holding a whole `name:tag` reference is
  upgraded, as the stale digest would keep resolving to the old image.
- Upgrade quoted default values, keeping their quotes. The parser hands an
  `ARG` value to us as a single literal with the quotes still in its text and
  no quote style, so `ARG JAVA_VERSION="11"` never matched a version before.

* Trim comments and javadoc that restate the code

* Plan `ARG` upgrades before rewriting any `FROM`

Rewriting a `FROM` as it was visited, and only withholding the matching `ARG`
bump after the traversal, left half applied edits behind: a dropped digest pin
or a rename to `eclipse-temurin` next to an argument still holding the old
version.

The whole file is now planned up front, and replayed until the set of withheld
arguments stops growing, as withholding one argument can rule out the images
that depend on it. Only the surviving plan is applied. Every give up path now
withholds the arguments that `FROM` reads.

* Read `ARG` default values through the new `Argument` accessors

openrewrite/rewrite#8576 routes `ARG` values through the same path as every
other argument, so quotes are modelled in `Literal.quoteStyle` and `$VAR`
references become `EnvironmentVariable`. The local `QuotedText` that took the
quotes off the literal text by hand is no longer needed, and reading a value
is now `Argument.getText()`, which gives up on a reference the same way.

* Leave a `FROM` with a dangling tag separator alone

openrewrite/rewrite#8590 splits an image reference in the grammar, so
`FROM eclipse-temurin:` now parses rather than failing. Cover the shape that
newly reaches the recipe.

* Do not bump `ARG` defaults shared with images we leave alone

An argument holding only an image name carries no version, so renaming a
deprecated image holds for every `FROM` that reads it, whatever tag each one
carries. Withhold a version alone:

  ARG BASE=openjdk          ARG BASE=eclipse-temurin
  FROM ${BASE}:11-jre  ->   FROM ${BASE}:25-jre
  FROM ${BASE}:latest       FROM ${BASE}:latest

A `FROM` whose image we can not resolve, or is not one we upgrade, still
withholds both, as we know too little to rename either.

Hoist the recipe into `defaults`, leaving the one test that varies the target
version to override it.

* Bump an `ARG` default shared with an image we leave alone

An image we do not recognise may well be an internal one built on the same
Java version, and if it shared a version argument before it should move along
with it:

  ARG VERSION=11                       ARG VERSION=25
  FROM eclipse-temurin:${VERSION}  ->  FROM eclipse-temurin:${VERSION}
  FROM acme/base:${VERSION}            FROM acme/base:${VERSION}

Each `FROM` now stands on its own, so the repeated planning that let one
withheld argument withhold the next is gone, and a visitor reads the file in
one pass.

Group the tests that assert no change into a nested class.

* Upgrade a Java image pulled through a registry

Mirroring stock images through an internal registry is ordinary, so read the
repository past a registry rather than treating the whole reference as a name
we do not know:

  ARG REGISTRY                                 ARG REGISTRY
  FROM ${REGISTRY}/eclipse-temurin:11-jre  ->  FROM ${REGISTRY}/eclipse-temurin:25-jre
  FROM docker.io/openjdk:11-jre                FROM docker.io/eclipse-temurin:25-jre

Which leading segment is a registry follows Docker's own rule: one holding a
`.` or a `:`, or `localhost`. Anything else belongs to the repository, so
`azul/zulu-openjdk` still reads as a name and `mycompany/eclipse-temurin` is
left alone.

Also rewrite only the `ARG` declaration a default was read from, as a name may
be declared more than once and the others carry no value to upgrade.

* Read a registry through rewrite-docker's `ImageName`

The rule for which leading segment of an image name is a registry now lives
upstream, in `org.openrewrite.docker.trait.ImageName`, so read the registry
and repository from it rather than keeping a second copy of the rule here.

That also covers a segment upstream reads as a registry and the copy here did
not: a repository name may not be uppercase, so `MyRegistry/openjdk` names a
registry, not a repository.

Reading the name through `getTextWithVariables()` lets one parse serve a
spelt-out registry and a `${REGISTRY}` alike, which retires the second helper
that recognised the latter by the shape of its contents.

* Upgrade a `FROM` as it is visited

Planning every `FROM` before rewriting any was there to withhold an argument
another `FROM` ruled out. Nothing is withheld now, so each one can be upgraded
where it is met, which retires the second traversal, the plan holder and the
map of replacements keyed by id.

Read the declaration an argument's default came from by its value rather than
by its id, so the two maps become one.
timtebeek added a commit that referenced this pull request Aug 22, 2026
…ns (#1213)" (#1215)

This reverts commit d33eca3.

The recipe reached for rewrite-docker API that no released version of rewrite
carries yet. `Docker.Argument.getText()`, `getTextWithVariables()` and
`hasEnvironmentVariables()` arrived in openrewrite/rewrite#8576, and
`org.openrewrite.docker.trait.ImageName` in openrewrite/rewrite#8599, both
landed 2026-08-21, one day after v8.90.3.

The Moderne CLI loads the LST classes in its own classloader, so a recipe runs
against the rewrite-docker the CLI bundles rather than the one this artifact
resolves. CLI 4.6.3 bundles rewrite-docker 8.90.3, where `Docker.Argument`
exposes only `getContents()` and `ImageName` does not exist. `visitFile` reads
a global `ARG` through `getText()` and `visitFrom` opens with
`hasEnvironmentVariables()`, so the first `FROM` of every Dockerfile would
raise `NoSuchMethodError`, surfacing as error markup on every Dockerfile of
every `UpgradeToJava*` run.

Nothing is lost by waiting: v3.42.1 predates this commit, so the breakage has
not shipped. The state restored here is the #1212 fix, which reads an image
reference through `DockerFrom` alone and so links against 8.90.3.

Reapplied in a follow-up PR, to merge once a rewrite release carries #8576,
#8590 and #8599 and the CLI picks it up.
timtebeek added a commit that referenced this pull request Aug 22, 2026
… for rewrite release) (#1216)

* Revert "Upgrade `ARG` default values used in Docker `FROM` instructions (#1213)"

This reverts commit d33eca3.

The recipe reached for rewrite-docker API that no released version of rewrite
carries yet. `Docker.Argument.getText()`, `getTextWithVariables()` and
`hasEnvironmentVariables()` arrived in openrewrite/rewrite#8576, and
`org.openrewrite.docker.trait.ImageName` in openrewrite/rewrite#8599, both
landed 2026-08-21, one day after v8.90.3.

The Moderne CLI loads the LST classes in its own classloader, so a recipe runs
against the rewrite-docker the CLI bundles rather than the one this artifact
resolves. CLI 4.6.3 bundles rewrite-docker 8.90.3, where `Docker.Argument`
exposes only `getContents()` and `ImageName` does not exist. `visitFile` reads
a global `ARG` through `getText()` and `visitFrom` opens with
`hasEnvironmentVariables()`, so the first `FROM` of every Dockerfile would
raise `NoSuchMethodError`, surfacing as error markup on every Dockerfile of
every `UpgradeToJava*` run.

Nothing is lost by waiting: v3.42.1 predates this commit, so the breakage has
not shipped. The state restored here is the #1212 fix, which reads an image
reference through `DockerFrom` alone and so links against 8.90.3.

Reapplied in a follow-up PR, to merge once a rewrite release carries #8576,
#8590 and #8599 and the CLI picks it up.

* Upgrade `ARG` default values used in Docker `FROM` instructions

Reapplies #1213, reverted in #1215 because the rewrite-docker API it reads
`ARG` defaults through had not been released yet.

Hold until a rewrite release carries openrewrite/rewrite#8576 (the
`Docker.Argument` accessors), #8590 (the image reference grammar) and #8599
(`ImageName`), and confirm the Moderne CLI bundles that release, as the CLI
loads the LST classes in its own classloader and so decides which
rewrite-docker a recipe actually links against.

* Revert "Merge branch 'main' into tim/docker-arg-defaults-redo"

This reverts commit 901210a, keeping the branch side.

Merging main in pulled #1215 across, and #1215 is the revert of the very
commit this branch exists to reapply. The merge base still carried the `ARG`
work and main had removed it, so the merge resolved to main's removal and
emptied the branch: `git diff main...HEAD` came back with nothing, leaving the
pull request proposing no change at all.

Reverting the merge rather than dropping it also settles the branch. The merge
stays in history, so main's revert counts as already merged here and undone on
purpose. A later `main` merge brings its new commits without resurrecting the
removal, which resetting the branch would leave it open to on the next
`Update branch`.

* Read an argument's text through `ArgumentContents`

openrewrite/rewrite#8608 moves `getText()`, `getTextWithVariables()`,
`getQuoteStyle()` and `hasEnvironmentVariables()` off `Docker.Argument` and
into `org.openrewrite.docker.internal.ArgumentContents`, then drops them from
the LST type.

That matters because of how the Moderne CLI splits one rewrite-docker jar
across two classloaders: `org.openrewrite.docker.tree` resolves to the
rewrite-docker the CLI bundles, while recipes, traits and `internal` load
child-first from the recipe artifact. Reading an argument through the LST type
therefore linked against the CLI's copy, which is where #1215 came from. The
helpers now sit on the recipe's side of that split and read only members that
predate the CLIs in the field, so they travel with this artifact.

`ImageName` already sits on that side, so it needed no change.

Verified against the pull request rather than assumed: rewrite-docker built at
6f5fd253 and published locally, and every one of the fifteen
`org.openrewrite.docker.tree` members this recipe links against confirmed
present in the 8.90.3 jar CLI 4.6.3 bundles, matching on descriptor.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

UpgradeDockerImageVersion hardcoded Dockerfile image issue

1 participant