ExceptionInInitializerError in CommonMarkFlavourDescriptor.createInlinesLexer - MarkdownParser.doParseInline

[eboudrant]

We're seeing a crash with the markdown library used in Android/Compose UI, we can't really reproduce it but I'll try to add more informations when I can. The the crash happened with strings that are about 40-50 char length and contains no markdown chars like ** * [ ] ( ) _. We can't get the actual content of the string.

We're using markdown 0.6.1.

java.lang.ExceptionInInitializerError
        at org.intellij.markdown.flavours.commonmark.CommonMarkFlavourDescriptor.createInlinesLexer(CommonMarkFlavourDescriptor:35)
        at org.intellij.markdown.parser.MarkdownParser.doParseInline(MarkdownParser:88)
        at org.intellij.markdown.parser.MarkdownParser.parseInline(MarkdownParser:45)
        at org.intellij.markdown.parser.MarkdownParser$InlineExpandingASTNodeBuilder.createLeafNodes(MarkdownParser:123)
        at org.intellij.markdown.parser.TopLevelBuilder.createASTNodeOnClosingEvent(TopLevelBuilder:21)
        at org.intellij.markdown.parser.TreeBuilder.buildTree(TreeBuilder:48)
        at org.intellij.markdown.parser.MarkdownParser.doParse(MarkdownParser:83)
        at org.intellij.markdown.parser.MarkdownParser.parse(MarkdownParser:33)
        at org.intellij.markdown.parser.MarkdownParser.buildMarkdownTreeFromString(MarkdownParser:28)
        ...
Caused by: java.lang.NullPointerException: Attempt to invoke virtual method 'char java.lang.String.charAt(int)' on a null object reference
        at org.intellij.markdown.lexer._MarkdownLexer$Companion.zzUnpackTrans(_MarkdownLexer:560)
        at org.intellij.markdown.lexer._MarkdownLexer$Companion.zzUnpackTrans(_MarkdownLexer:551)
        at org.intellij.markdown.lexer._MarkdownLexer$Companion.access$zzUnpackTrans(_MarkdownLexer:447)

[fnfunfunc]

We've got this issue on 0.7.3 too

[AlexanderGH]

I suspect this is a Dexguard/Proguard bug. Adding explicit keep rules for those classes seems to fix the issue.

[fnfunfunc]

I suspect this is a Dexguard/Proguard bug. Adding explicit keep rules for those classes seems to fix the issue.

Which fields/functions should I keep?

[AlexanderGH]

The one(s) throwing the exception:

org.intellij.markdown.lexer._MarkdownLexer
org.intellij.markdown.lexer._MarkdownLexer$Companion

Can you confirm if you're using Dexguard, or Proguard, or just R8? And if it's Dexguard, would you mind filing a ticket with them?

[fnfunfunc]

The one(s) throwing the exception:

org.intellij.markdown.lexer._MarkdownLexer
org.intellij.markdown.lexer._MarkdownLexer$Companion

Can you confirm if you're using Dexguard, or Proguard, or just R8? And if it's Dexguard, would you mind filing a ticket with them?

I'm using Proguard and I will try to keep these classes, thanks