[
https://issues.apache.org/jira/browse/LUCENE-1654?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714068#action_12714068
]
Uwe Schindler edited comment on LUCENE-1654 at 5/28/09 10:50 AM:
-----------------------------------------------------------------
I found a bug for indexes with the string-based userdata:
{code}
if (format <= FORMAT_USER_DATA) {
if (format <= FORMAT_DIAGNOSTICS) {
userData = input.readStringStringMap();
} else {
userData = Collections.EMPTY_MAP;
if (0 != input.readByte()) {
userData.put("userData", input.readString());
}
}
} else {
userData = Collections.EMPTY_MAP;
}
{code}
But the empty maps of Collections are read-only, so put() throws UOE.
I fix this and commit: userData should be a singleton map, if 0!=readByte(),
otherwise EMPTY_MAP.
was (Author: thetaphi):
I found a bug for indexes with the string-based userdata:
{code}
if (format <= FORMAT_USER_DATA) {
if (format <= FORMAT_DIAGNOSTICS) {
userData = input.readStringStringMap();
} else {
userData = Collections.EMPTY_MAP;
if (0 != input.readByte()) {
userData.put("userData", input.readString());
}
}
} else {
userData = Collections.EMPTY_MAP;
}
{code}
I fix this and commit. userData should be a singleton map, if 0!=readByte() or
EMPTY_MAP. But the empty maps of Collections are read-only, so put() throws UOE.
> Include diagnostics per-segment when writing a new segment
> ----------------------------------------------------------
>
> Key: LUCENE-1654
> URL: https://issues.apache.org/jira/browse/LUCENE-1654
> Project: Lucene - Java
> Issue Type: Improvement
> Reporter: Michael McCandless
> Assignee: Michael McCandless
> Fix For: 2.9
>
> Attachments: LUCENE-1654.patch, LUCENE-1654.patch
>
>
> It would be very helpful if each segment in an index included
> diagnostic information, such as the current version of Lucene.
> EG, in LUCENE-1474 this would be very helpful to see if certain
> segments were written under 2.4.0.
> We can start with just the current version.
> We could also consider making this extensible, so you could provide
> your own arbitrary diagnostics, but SegmentInfo/s is not public so I
> think such an API would be "one-way" in that you'd have to use
> CheckIndex to check on it later. Or we could wait on such extensibility
> until we provide some consistent way to access per-segment details
> in the index.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
