Skip to content

Android: List.of() (API 30+) used in core since 0.24.0 contradicts the README's "minimum API level is 19" #457

Description

@Jimmy-Jung

Summary

Since 0.24.0 the core commonmark module uses java.util.List.of(...) (and a few Set.of/Map.of). On Android these methods exist only from API 30 (Android 11). The README still says:

It works on Android too, but that is on a best-effort basis, please report problems. For Android the minimum API level is 19

An app with minSdk < 30 that does not enable core library desugaring crashes on the first parse on API < 30 devices:

java.lang.NoSuchMethodError: No static method of(Ljava/lang/Object;)Ljava/util/List;
    at org.commonmark.internal.InlineParserImpl.parse(InlineParserImpl.java)

This is the same class of problem as #366 / #373 (Objects.requireNonNullElseGet, fixed by #369). The List.of usages were introduced deliberately in #322 as a Java 11 clean-up, so this is a report that the clean-up and the Android support statement now disagree — not a claim that the change was wrong.

Steps to reproduce

Android app, minSdk 24, no coreLibraryDesugaring, org.commonmark:commonmark:0.30.0, run on an API 24–29 emulator:

Parser.builder().build().parse("*a*")   // NoSuchMethodError: List.of

With isCoreLibraryDesugaringEnabled = true + com.android.tools:desugar_jdk_libs:2.1.5 the same code works on API 24 (verified on an API 24 emulator).

Where the API 30+ calls live (0.30.0, commonmark module)

List.of / Set.of / Map.of occurrences by file:

File Count
internal/InlineParserImpl.java 8
renderer/markdown/CoreMarkdownNodeRenderer.java 3
renderer/html/CoreHtmlNodeRenderer.java 3
parser/Parser.java 3 (incl. Javadoc example)
internal/ParagraphParser.java 2
internal/IndentedCodeBlockParser.java 2
renderer/text/CoreTextContentNodeRenderer.java, renderer/markdown/MarkdownRenderer.java, renderer/html/HtmlWriter.java, renderer/html/DefaultUrlSanitizer.java, parser/block/AbstractBlockParser.java, node/SourceSpans.java 1 each

0.22.0 had none; 0.24.0 introduced 8 in InlineParserImpl.

Why CI did not catch it

commonmark-android-test is only run through ./gradlew :app:lint in .github/workflows/ci.yml (minSdk 19, no emulator run), so a runtime NoSuchMethodError never surfaces there.

Possible resolutions

Either would be fine from a consumer's point of view — the important part is that the README and the code agree:

  1. Keep API 19 support — replace the factory calls with Java 8 equivalents (Collections.emptyList() / singletonList(x) / unmodifiableList(Arrays.asList(a, b)), same for Set/Map). Mechanical, ~26 sites, keeps immutability. I can send a PR if you want this.
  2. Document the new requirement — state in the README that Android needs API 30+ or core library desugaring, and ideally have commonmark-android-test run on an emulator (or fail lint on NewApi) so future regressions are visible.

Context: I hit this while building an Android Markdown renderer on top of commonmark-java (https://github.com/Jimmy-Jung/RichMarkdown-Android). We chose minSdk 30 for now, so this is not blocking us — reporting because the README statement is likely to mislead other Android users.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions