[jira] [Comment Edited] (CSV-102) Commons CSV does not properly handle record separators

["Roger Whitcomb (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CSV-102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13851057#comment-13851057 ] 

Roger Whitcomb edited comment on CSV-102 at 12/17/13 11:08 PM:
---------------------------------------------------------------

One reason for doing this, however, is that the CSVFormat class claims (already) to handle a different record separator, but in fact it doesn't really handle anything but CR/LF or some combination.  The record separator setting in CSVFormat is completely ignored in Lexer.java and it only uses CR/LF in every place where it is identifying the end/start of a record.


was (Author: rwhitcomb):
One reason for doing this, however, is that the CSVFormat class claims (already) to handle a different record separator, but in fact it doesn't really handle anything but CR/LF or some combination.  

> Commons CSV does not properly handle record separators
> ------------------------------------------------------
>
>                 Key: CSV-102
>                 URL: https://issues.apache.org/jira/browse/CSV-102
>             Project: Commons CSV
>          Issue Type: Bug
>    Affects Versions: 1.x
>         Environment: ALL
>            Reporter: Oliver Oyston
>            Priority: Minor
>         Attachments: Changes.patch, nonstandard.patch
>
>
> Any specified record separator appear to be ignored and the code just assumes that standard line endings are the record delimiters.
> I have a patch to partly address the problems that 'falls back' to the current behavior. The patch only allow a maximum of a one character record separator.
> The approach needs verifying and also needs improving to be more generic.



--
This message was sent by Atlassian JIRA
(v6.1.4#6159)