[
https://issues.apache.org/jira/browse/ORC-589?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Zoltán Borók-Nagy resolved ORC-589.
-----------------------------------
Fix Version/s: 1.7.0
1.6.3
Resolution: Fixed
> [C++] ORC doesn't check for negative dictionary entry lengths anymore
> ---------------------------------------------------------------------
>
> Key: ORC-589
> URL: https://issues.apache.org/jira/browse/ORC-589
> Project: ORC
> Issue Type: Bug
> Reporter: Zoltán Borók-Nagy
> Assignee: Zoltán Borók-Nagy
> Priority: Major
> Fix For: 1.6.3, 1.7.0
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> The constructor of StringDictionaryColumnReader used to have a check for
> negative dictionary entry lengths:
> [https://github.com/apache/orc/blob/branch-1.5/c%2B%2B/src/ColumnReader.cc#L523-L524]
> {noformat}
> for (uint64_t i = 1; i < dictionaryCount + 1; ++i) {
> if (lengthArray[i] < 0)
> throw ParseError("Negative dictionary entry length");
> lengthArray[i] += lengthArray[i - 1];
> }{noformat}
> However we don't have it on current master. It causes a problem because it
> will use a negative value (casted to unsigned) to allocate a DataBuffer:
>
> {noformat}
> dictionary->dictionaryBlob.resize(static_cast<uint64_t>(lengthArray[dictSize]))
> {noformat}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
