Android Lint: Unknown Issue IDs should NOT break the build [158295165]

Fixed

Bug

Status Update

No update yet.

Description

br...@capitalone.com

created issue #1

Jun 5, 2020 04:48PM

Build: AI-192.7142.36.36.6392135, 202004132352,

AI-192.7142.36.36.6392135, JRE 1.8.0_212-release-1586-b4-5784211x64 JetBrains s.r.o, OS Mac OS X(x86_64) v10.15.4, screens 1920x1200; Retina

AS: 3.6.3; Kotlin plugin: 1.3.71-release-Studio3.6-1; Android Gradle Plugin: 3.6.3; Gradle: 6.4.1; NDK: from local.properties: (not specified), latest from SDK: (not found); LLDB: pinned revision 3.1 not found, latest from SDK: (package not found); CMake: from local.properties: (not specified), latest from SDK: (not found), from PATH: (not found)

IMPORTANT: Please read https://developer.android.com/studio/report-bugs.html carefully and supply all required information.

See conversation here on the lint-dev Google Group.

Current behavior: When a module inherits lint suppressions that suppress a lint rule that is not actually applied to the given module, the build fails with an error like:

Error: Unknown issue id "SomeIssueId", found in /home/tools/jenkins/workspace/.../lint.xml [LintError]

Expected behavior: Because the lint rule is not applied to the module, it doesn't matter that the suppression isn't targeting any rule for that module. Just ignore (and possibly place something in the lint report) and allow the build to succeed.

Comments

uc...@google.com <uc...@google.com> Jun 5, 2020 05:26PM

Assigned to an...@google.com.

tn...@google.com <tn...@google.com> Jun 22, 2020 10:20PM

Reassigned to tn...@google.com.

tn...@google.com <tn...@google.com> #2Jun 24, 2020 01:29AM

Marked as fixed.

This is fixed by Change-Id: Icdccf15d6fb09b0ed8239e8237736c2c485042b4 for Studio 4.2. It's now a separate issue (instead of using the general LintError issue), such that it can be enabled/disabled individually, and it's an issue that has default severity of warning, which means by default it won't break the build.

But, not that this is basically the opposite of the issue which requested this behavior to begin with: Issue 37070812: Lint does not fail when invalid issue ID is referenced in XML.

Since that issue was filed by Jake, who now seems to agree with the opposite behavior (see https://groups.google.com/g/lint-dev/c/Ok00I-tOdNg) changing this seems like the right behavior, and since it's still a warning it will help catch typos etc.

go...@jakewharton.com <go...@jakewharton.com> #3Jun 24, 2020 02:28AM

Oh wow. That's legitimately funny.

I guess in my defense the original issue was for a typo of a built-in check and this one (at least for me) is for a third-party check that may not be available in the dependency graph.

Somewhat related: maybe it would be useful to try and suggest valid lint IDs when an invalid one is found using something like edit distance and a cutoff?

br...@capitalone.com <br...@capitalone.com> #4Jun 24, 2020 12:24PM

Somewhat related: maybe it would be useful to try and suggest valid lint IDs when an invalid one is found using something like edit distance and a cutoff?

Agreed, this would be a good enhancement to whatever is put in the report for the new Issue. Something like "Invalid Issue Id UnuseResource, did you instead mean UnusedResource?"

Thanks for making this change, will make our lives a lot easier when inheriting rules from 3P artifact upticks.

tn...@google.com <tn...@google.com> #5Jun 24, 2020 04:53PM

Good idea! Lint already has an edit distance implementation and uses it in some other lint checks (such as WrongIdDetector) so this was easy to hook up; should also be in 4.2 via Change-Id: Ic4d6c86a4394c8823b8e5e45e569a452b11e9de7.

Issue 158295165