[ https://issues.apache.org/jira/browse/LUCENE-8533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16654982#comment-16654982 ]
Vladimir Dolzhenko commented on LUCENE-8533: -------------------------------------------- Patch [^LUCENE-8533_fix_readVInt_javadoc.patch] to fix javadoc is attached. > Incorrect readVInt: negative number could be returned > ----------------------------------------------------- > > Key: LUCENE-8533 > URL: https://issues.apache.org/jira/browse/LUCENE-8533 > Project: Lucene - Core > Issue Type: Improvement > Reporter: Vladimir Dolzhenko > Assignee: Uwe Schindler > Priority: Major > Attachments: LUCENE-8533_fix_readVInt_javadoc.patch, readVInt.patch > > > {{readVInt()}} has to return positive numbers (and zero), throw some > exception in case of negative numbers. > While for the sequence of bytes {{[-1, -1, -1, -1, 15]}} it returns {{-1}}. > simplifying > [readVInt|https://github.com/apache/lucene-solr/blob/1d85cd783863f75cea133fb9c452302214165a4d/lucene/core/src/java/org/apache/lucene/store/DataInput.java#L113] > up to last readByte (exclusive): > {code:java} > int i = ((byte)-1) & 0x7F; > i |= (((byte)-1) & 0x7F) << 7; > i |= (((byte)-1) & 0x7F) << 14; > i |= (((byte)-1) & 0x7F) << 21; > {code} > Here {{i = 268435455}} or in binary format is > {{00001111_11111111_11111111_11111111}} > Keeping in mind that {{int}} is a signed type we have only 3 more bits before > overflow happens or in another words {{(Integer.MAX_VALUE - i) >> 28 == 7}} - > that's max value could be stored in 5th byte to avoid overflow. > Instead of > {code:java} > i |= (b & 0x0F) << 28; > if ((b & 0xF0) == 0) return i; > {code} > has to be > {code:java} > i |= (b & 0x07) << 28; > if ((b & 0xF8) == 0) return i; > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org