You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Shai Erera (JIRA)" <ji...@apache.org> on 2010/05/27 17:44:40 UTC
[jira] Created: (LUCENE-2480) Remove support for pre-3.0 indexes
Remove support for pre-3.0 indexes
----------------------------------
Key: LUCENE-2480
URL: https://issues.apache.org/jira/browse/LUCENE-2480
Project: Lucene - Java
Issue Type: Improvement
Components: Index
Reporter: Shai Erera
Priority: Minor
Fix For: 4.0
We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
# Remove the .zip indexes
# Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
# remove FORMAT_PRE from FieldInfos
# Remove old format from TermVectorsReader
If you know of other places where code can be removed, then please post a comment here.
I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Reopened: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Uwe Schindler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Uwe Schindler reopened LUCENE-2480:
-----------------------------------
I reopen this issue, as the following problem is there:
When somebody tries to open an pre-3.0 index, it should fail with a meaningful exception. At least e.g. FieldsReader (where I found dead code after your removal) does not check that the version is ge the minimum version. We should add a static FORMAT_MINIMUM to all places where formats are used and check it on reading.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Uwe Schindler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Uwe Schindler updated LUCENE-2480:
----------------------------------
Attachment: (was: LUCENE-2480-minimumsupport.patch)
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Uwe Schindler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Uwe Schindler updated LUCENE-2480:
----------------------------------
Attachment: LUCENE-2480-minimumsupport.patch
Here a first patch, that handles this in FieldsReader to show how it should be done elsewhere, too. This patch also removes some "dead" code in FieldReader, too. I left the method in, as we may need canReadRawDocs() for later format changes, too.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480-minimumsupport.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Earwin Burrfoot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873292#action_12873292 ]
Earwin Burrfoot commented on LUCENE-2480:
-----------------------------------------
Doing that now, plus some additions to Shai's patch
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Shai Erera (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873408#action_12873408 ]
Shai Erera commented on LUCENE-2480:
------------------------------------
Solr tests pass too ... dunno what happened, but they started to pass. And the dataimport tests do not seem they should be affected by this issue. So I'd like to commit this soon.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Shai Erera (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Shai Erera updated LUCENE-2480:
-------------------------------
Attachment: LUCENE-2480.patch
Patch removes code which supports old indexes. If there are more places in the code that you know of, please let me know.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Earwin Burrfoot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873465#action_12873465 ]
Earwin Burrfoot commented on LUCENE-2480:
-----------------------------------------
Wow! So fast! :)
bq. You didn't remove the .zip indexes?
Your patch didn't remove them for me, patch file format can't handle binary stuff.
bq. You removed the code from TestBackwardsCompatibilty ...... So I don't think we should remove that piece of code.
Several lines later there sits a piece of absolutely identical commented-out code. The only difference is that it doesn't use "preLockless" in method names.
If someone really needs it later, the code I left over suits him no less than the code I removed.
bq. TestIndexFileDeleter failed 'cause of deletable file - fixed it.
Cool! It's very strange I missed the failure.
bq. Renamed FORMAT_FLEX_POSTING to FORMAT_4_0
FORMAT_4_0_FLEX_POSTINGS? I loved the self-descriptiveness :)
And, yup. Bye-bye to the byte.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Resolved: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Uwe Schindler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Uwe Schindler resolved LUCENE-2480.
-----------------------------------
Resolution: Fixed
Closing again.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Uwe Schindler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12889619#action_12889619 ]
Uwe Schindler commented on LUCENE-2480:
---------------------------------------
Oh, I move over to this issue. I am almost finished!
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Earwin Burrfoot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873469#action_12873469 ]
Earwin Burrfoot commented on LUCENE-2480:
-----------------------------------------
bq. Strange, there were lines in my patch file which indicated they are removed, but were absent from yours.
{code}
Index: lucene/src/test/org/apache/lucene/index/index.19.cfs.zip
===================================================================
Cannot display: file marked as a binary type.
svn:mime-type = application/octet-stream
Index: lucene/src/test/org/apache/lucene/index/index.19.nocfs.zip
===================================================================
Cannot display: file marked as a binary type.
svn:mime-type = application/octet-stream
Index: lucene/src/test/org/apache/lucene/index/index.20.cfs.zip
===================================================================
Cannot display: file marked as a binary type.
svn:mime-type = application/octet-stream
{code}
These lines in your patch mean that diff noticed the differences, but failed to express them - nothing more. You could modify all these files instead of deleting, and get the same result.
bq. I don't see it.
Ctrl+F, "testCreateCFS"
Ok, they're not completely identical. But I don't really care either way.
bq. .... FORMAT stuff .....
Ok.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Resolved: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Shai Erera (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Shai Erera resolved LUCENE-2480.
--------------------------------
Assignee: Shai Erera
Resolution: Fixed
Thanks Mike and Earwin for your help !
Committed revision 949509.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Shai Erera (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Shai Erera updated LUCENE-2480:
-------------------------------
Attachment: LUCENE-2480.patch
Thanks Earwin ! Few comments:
* You didn't remove the .zip indexes?
* You removed the code from TestBackwardsCompatibilty which creates the older indexes. While this isn't needed currently for trunk - the code is commented out, and I don't think it bothers anyone - one day, after 4.0 is released, we will need to use it. So I don't think we should remove that piece of code.
* TestIndexFileDeleter failed 'cause of deletable file - fixed it.
Attached is the merged patch, w/ the above fixes. Solr tests fail (*.dataimport.* tests) for some reason. I don't understand this yet - so if someone more knowledgeable about Solr than me cares to take a look, I'd appreciate it. I see many things like that (not sure how those tests are related to this issue):
{code}
[junit]
[junit] request was: q=id:500&start=0&qt=standard&version=2.2&rows=20
[junit] junit.framework.AssertionFailedError: now it should query failed XPath: //result[@numFound=1]
[junit] xml response was: <?xml version="1.0" encoding="UTF-8"?>
[junit] <response>
[junit] <lst name="responseHeader"><int name="status">0</int><int name="QTime">16</int></lst><result name="response" numFound="0" start="0"/>
[junit] </response>
[junit]
[junit] request was: q=id:500&start=0&qt=standard&version=2.2&rows=20
[junit] at org.apache.solr.util.AbstractSolrTestCase.assertQ(AbstractSolrTestCase.java:236)
[junit] at org.apache.solr.update.AutoCommitTest.testMaxTime(AutoCommitTest.java:218)
{code}
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Earwin Burrfoot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Earwin Burrfoot updated LUCENE-2480:
------------------------------------
Attachment: LUCENE-2480.patch
Here we go. Pre-utf8/compressed fields support removed, microcleanups/additions to previous patch.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Michael McCandless (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12889617#action_12889617 ]
Michael McCandless commented on LUCENE-2480:
--------------------------------------------
I already opened LUCENE-2523 for this, but it's OK if we do the work here and resolve that one as a dup.
I think you should see an IndexTooOldException instead of CorruptIndexException (though the former could subclass the latter?)
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480-minimumsupport.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Michael McCandless (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873285#action_12873285 ]
Michael McCandless commented on LUCENE-2480:
--------------------------------------------
Great work Shai!
Another place we can remove is the preUTF8Strings in oal.store.DataInput, and anywhere that calls setModifiedUTF8StringsMode, and I think also FieldsWriter.FORMAT_VERSION_UTF8_LENGTH_IN_BYTES?
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Earwin Burrfoot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Earwin Burrfoot updated LUCENE-2480:
------------------------------------
Comment: was deleted
(was: Doing that now, plus some additions to Shai's patch)
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Updated: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Shai Erera (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Shai Erera updated LUCENE-2480:
-------------------------------
Attachment: LUCENE-2480.patch
* Removed the byte written by SegmentInfo (hasSingleNormsFile - redundant).
* Added live migration code to SI.ctor to read that byte if pre-4.0 index
* Renamed FORMAT_FLEX_POSTING to FORMAT_4_0. I think we should do this going forward in more places, so that it's clear to which version this format belongs and will make the task of removing support for old indexes in the future easier. The jdoc of the constant needs to describe what is covered by that format.
Besides the Solr tests, everything passes.
BTW, I don't think we should stop writing that byte in 3.1 - not sure if hasSingleNormsFile is always true, and even if so, 3.0 indexes still need to be supported and they write the byte. And it does not add much of complexity.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2480) Remove support for pre-3.0 indexes
Posted by "Shai Erera (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873466#action_12873466 ]
Shai Erera commented on LUCENE-2480:
------------------------------------
bq. Your patch didn't remove them for me, patch file format can't handle binary stuff.
Strange, there were lines in my patch file which indicated they are removed, but were absent from yours. Anyway, they're removed now.
bq. Several lines later there sits a piece of absolutely identical commented-out code.
I don't see it. All I see is a comment and two methods that are used to generate the old indexes.
bq. FORMAT_4_0_FLEX_POSTINGS
While I don't mind that sort of descriptiveness, it does not fit in all cases, such as this case -- this format is related to both flex postings, but also to that byte removed from the SegmentInfo :). So I think we should keep the names simple, and have a useful javadoc. It's the sort of constants no one really cares about in his daily work, only when you handle file format backwards stuff :).
Also, if by chance (like this case, again) you get to change the file format version twice, by two different people, with long time interval between the two, a FORMAT_4_0 should alert you that that's used for the unreleased index, and so you don't need to bump it up again (here, jumping from -9 to -11) accidentally.
> Remove support for pre-3.0 indexes
> ----------------------------------
>
> Key: LUCENE-2480
> URL: https://issues.apache.org/jira/browse/LUCENE-2480
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Shai Erera
> Assignee: Shai Erera
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch, LUCENE-2480.patch
>
>
> We should remove support for 2.x (and 1.9) indexes in 4.0. It seems that nothing can be done in 3x because there is no special code which handles 1.9, so we'll leave it there. This issue should cover:
> # Remove the .zip indexes
> # Remove the unnecessary code from SegmentInfo and SegmentInfos. Mike suggests we compare the version headers at the top of SegmentInfos, in 2.9.x vs 3.0.x, to see which ones can go.
> # remove FORMAT_PRE from FieldInfos
> # Remove old format from TermVectorsReader
> If you know of other places where code can be removed, then please post a comment here.
> I don't know when I'll have time to handle it, definitely not in the next few days. So if someone wants to take a stab at it, be my guest.
--
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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org