You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Marshall Schor (JIRA)" <de...@uima.apache.org> on 2014/12/17 19:33:14 UTC

[jira] [Resolved] (UIMA-4145) bin cas deserialization form 6 delta cas has some errors in obscure cases

     [ https://issues.apache.org/jira/browse/UIMA-4145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Marshall Schor resolved UIMA-4145.
----------------------------------
    Resolution: Fixed

fixed in revision 1644829

> bin cas deserialization form 6 delta cas has some errors in obscure cases
> -------------------------------------------------------------------------
>
>                 Key: UIMA-4145
>                 URL: https://issues.apache.org/jira/browse/UIMA-4145
>             Project: UIMA
>          Issue Type: Bug
>          Components: Core Java Framework
>            Reporter: Marshall Schor
>            Assignee: Marshall Schor
>             Fix For: 2.7.0SDK
>
>
> While putting in support for deserialization of delta CASes which update existing FS which may be indexed (thus needing special handling to avoid potentially corrupting the index), I noticed what appeared to be a kind of copy-paste error - where a wrong FS feature offset variable was being used.
> This would only ever show up if form 6 was being used and the source and target type systems differed in a manner that led to different offsets in the FS for the affected kinds of features.  The affected kinds are: byte, boolean, float, and string, and arrays of float and string.  The typo was that the heap update was using the cas-being-deserialized'd feature offset, instead of the cas-being-deserialized-into's offset.  Other kinds were correct.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)