Compilation allowed for illegal color values [280475500]

Fixed

Bug

Status Update

No update yet.

Description

ev...@gmail.com

created issue #1

May 3, 2023 01:20PM

Build: AI-221.6008.13.2211.9619390, 202302170051, 

AI-221.6008.13.2211.9619390, JRE 11.0.15+0-b2043.56-8887301x64 JetBrains s.r.o., OS Mac OS X(aarch64) v13.3.1, screens 2560.0x1440.0, 1440.0x2560.0

AS: Electric Eel | 2022.1.1 Patch 2
Kotlin plugin: 221-1.8.0-release-for-android-studio-AS5591.52
Android Gradle Plugin: 7.4.2
Gradle: 7.5
Gradle JDK: version 11.0.16
NDK: from local.properties: (not specified), latest from SDK: (not found)
CMake: from local.properties: (not specified), latest from SDK: (not found), from PATH: 3.26.3


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

The short version: there's no error raised for something like this:

<color name="purple_200">#0xFFBB86FC</color>

#0xFFBB86FC is not a valid hex code, and should not be allowed to be compiled. Instead, the app compiles successfully and fails at runtime. I've made tiny little demo app here: https://github.com/evancharlton/IllegalColor if that's helpful.

Comments

jg...@google.com <jg...@google.com> May 4, 2023 02:32PM

Assigned to an...@google.com.

je...@google.com <je...@google.com> May 8, 2023 04:43PM

Reassigned to lu...@google.com.

lu...@google.com <lu...@google.com> #2May 9, 2023 04:42PM

Accepted by lu...@google.com.

I tried compiling a single color resource with the example hex value in comment 1 with the kotlin resource compiler and AAPT2 (version 34). Both failed as expected.

I will begin reproducing using the github project.

ev...@gmail.com <ev...@gmail.com> #3May 10, 2023 05:53AM

Very interesting! It's possible that we're using an outdated setup which allows this to be possible.

It's good to hear that this might not be an issue going forward, at least :)

lu...@google.com <lu...@google.com> #4May 15, 2023 01:11PM

#0xFFBB86FC -> build (AGP 7.4.2) fails [expected], aapt2 (34.0.0) fails [expected]

#blah -> build (AGP 7.4.2) success [not expected], fails with AAPT2 [expected]

##0xE1E1E1 -> build (AGP 7.4.2) success [not expected], fails with AAPT2 (34.0.0) [expected]

I'll take a look at our resource compilation error catching in AGP, as it might be ignoring AAPT fails. Another possibility is there is a mismatch (or bug rather) between the JVM resource compiler.

Although #0xFFBB86FC is compiling successfully in on my machine in the repro project, differing bug; but it does look like there are problems in this area.

Message last modified on May 15, 2023 01:11PM

ev...@gmail.com <ev...@gmail.com> #5May 15, 2023 01:58PM

That's great! Thanks for the confirmation :)

lu...@google.com <lu...@google.com> #6May 16, 2023 01:57PM

So the issue seems to be that the JVM (or Kotlin based) resource compiler sets the resource value of a color value resource to NULL rather than throwing an exception.

All that needs to be done here is to throw an exception when this is the case, so we can keep parity with AAPT behaviour.

lu...@google.com <lu...@google.com> May 23, 2023 05:30PM

Marked as fixed.