[
https://issues.apache.org/jira/browse/LUCENE-3014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13035370#comment-13035370
]
Shai Erera commented on LUCENE-3014:
------------------------------------
Hey guys, does this affect LUCENE-2921 (or vice versa)?
Basically, I thought that we should stop writing version header in files and
just use the release version as a header.
Robert, I don't think we are allowed to change index format versions on bug-fix
releases and even if we do, that same bug fix would go into the 3.x release so
it would still know how to read 3.1.1? Perhaps that was your point and I missed
it ...
> comparator API for segment versions
> -----------------------------------
>
> Key: LUCENE-3014
> URL: https://issues.apache.org/jira/browse/LUCENE-3014
> Project: Lucene - Java
> Issue Type: Task
> Reporter: Robert Muir
> Assignee: Uwe Schindler
> Priority: Critical
> Fix For: 3.2
>
> Attachments: LUCENE-3014.patch
>
>
> See LUCENE-3012 for an example.
> Things get ugly if you want to use SegmentInfo.getVersion()
> For example, what if we committed my patch, release 3.2, but later released
> 3.1.1 (will "3.1.1" this be whats written and returned by this function?)
> Then suddenly we broke the index format because we are using Strings here
> without a reasonable comparator API.
> In this case one should be able to compute if the version is < 3.2 safely.
> If we don't do this, and we rely upon this version information internally in
> lucene, I think we are going to break something.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
