[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713176#comment-16713176 ] ASF subversion and git services commented on LUCENE-8592: - Commit 719cde97f84640faa1e3525690d262946571245f in lucene-solr's branch refs/heads/branch_7_6 from [~jim.ferenczi] [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=719cde9 ] LUCENE-8592: switch the corrupted sorted index to a 7.6 version (instead of 7x). > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713174#comment-16713174 ] ASF subversion and git services commented on LUCENE-8592: - Commit d9cd9f78b1182125a7fb02d724608d5f355df785 in lucene-solr's branch refs/heads/branch_7x from [~jim.ferenczi] [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=d9cd9f7 ] LUCENE-8592: switch the corrupted sorted index to a 7.6 version (instead of 7x). > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713175#comment-16713175 ] ASF subversion and git services commented on LUCENE-8592: - Commit 3098af2b76e447e72f11d5de509ae154e0cec644 in lucene-solr's branch refs/heads/branch_7_6 from [~jim.ferenczi] [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=3098af2 ] LUCENE-8592: Fix index sorting corruption due to numeric overflow The merge sort of sorted segments can produce an invalid sort if the sort field is an Integer/Long that uses reverse order and contains values equal to Integer/Long#MIN_VALUE. These values are always sorted first during a merge (instead of last because of the reverse order) due to this bug. Indices affected by the bug can be detected by running the CheckIndex command on a distribution that contains the fix (7.6+). > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713172#comment-16713172 ] ASF subversion and git services commented on LUCENE-8592: - Commit 16a76883eba7a6c70a70b634fbe6cf89712f2d97 in lucene-solr's branch refs/heads/branch_7x from [~jim.ferenczi] [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=16a7688 ] LUCENE-8592: Fix index sorting corruption due to numeric overflow The merge sort of sorted segments can produce an invalid sort if the sort field is an Integer/Long that uses reverse order and contains values equal to Integer/Long#MIN_VALUE. These values are always sorted first during a merge (instead of last because of the reverse order) due to this bug. Indices affected by the bug can be detected by running the CheckIndex command on a distribution that contains the fix (7.6+). > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713169#comment-16713169 ] ASF subversion and git services commented on LUCENE-8592: - Commit df84a3c9815a85aa140e5013b44488f45de9a203 in lucene-solr's branch refs/heads/master from [~jim.ferenczi] [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=df84a3c ] LUCENE-8592: switch the corrupted sorted index to a 7.6 version (instead of 7x). > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713160#comment-16713160 ] ASF subversion and git services commented on LUCENE-8592: - Commit 9f29ed0757eae12d8311ffd6891f7032370ea39a in lucene-solr's branch refs/heads/master from [~jim.ferenczi] [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=9f29ed0 ] LUCENE-8592: Fix index sorting corruption due to numeric overflow The merge sort of sorted segments can produce an invalid sort if the sort field is an Integer/Long that uses reverse order and contains values equal to Integer/Long#MIN_VALUE. These values are always sorted first during a merge (instead of last because of the reverse order) due to this bug. Indices affected by the bug can be detected by running the CheckIndex command on a distribution that contains the fix (7.6+). > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713130#comment-16713130 ] Michael McCandless commented on LUCENE-8592: +1, patch looks great! Thanks [~jim.ferenczi]. > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16713127#comment-16713127 ] Michael McCandless commented on LUCENE-8592: Thanks [~jim.ferenczi], I'll look now! > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Affects Versions: 7.5, master (8.0) >Reporter: Jim Ferenczi >Priority: Blocker > Fix For: 7.6, master (8.0) > > Attachments: LUCENE-8592.patch, LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[
https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16712958#comment-16712958
]
Michael McCandless commented on LUCENE-8592:
Phew this is evil, nice catch [~jim.ferenczi]! Thank you for adding reverse
true/false testing through all the test cases. The patch looks good – moving
the {{reverseMul}} logic to after the comparison.
This may have high impact? Any time the user sorts by an int field, reversed,
and has missing values, their index will now fail {{CheckIndex}} on upgrade and
require reindexing from their original docs.
Whether this fix is buggy depends on the return values of e.g.
{{Integer.compareTo}}, {{String.compareTo}}, etc.? I.e. do those methods ever
return {{Integer.MIN_VALUE}}. Looking at [Integer.java in at least
JDK8|http://hg.openjdk.java.net/jdk8/jdk8/jdk/file/687fd7c7986d/src/share/classes/java/lang/Integer.java#l1233]
it seems we are good – it returns 0, 1, -1. And [String.java returns the
difference of two
{{chars}}|http://hg.openjdk.java.net/jdk8/jdk8/jdk/file/687fd7c7986d/src/share/classes/java/lang/String.java#l1140].
Hopefully all other native {{compareTo}} impls are similar.
Should we add a test case confirming {{CheckIndex}} detects this bug? Create a
broken index and zip it up and commit that, with a test that unzips and
confirms {{CheckIndex}} fails on it?
Instead of the (int) casts e.g. in {{return (int)
globalOrds.get(readerValues.ordValue())}}, maybe we should switch to
{{Math.toIntExact}}? Because these are single values, it should never happen
that the long ord exceeds positive int space, so the conversion should always
work safely.
> MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural
> sort is reversed
> --
>
> Key: LUCENE-8592
> URL: https://issues.apache.org/jira/browse/LUCENE-8592
> Project: Lucene - Core
> Issue Type: Bug
>Reporter: Jim Ferenczi
>Priority: Major
> Attachments: LUCENE-8592.patch
>
>
> MultiSorter#getComparableProviders on an integer or long field doesn't handle
> MIN_VALUE correctly when the natural order is reversed. To handle reverse
> sort we use the negation of the value but there is no check for overflows so
> MIN_VALUE for ints and longs are always sorted first (even if the natural
> order is reversed).
> This method is used by index sorting when merging already sorted segments
> together. This means that a sorted index can be incorrectly sorted if it uses
> a reverse sort and a missing value set to MIN_VALUE (long or int or values
> inside the segment that are equals to MIN_VALUE).
> This a bad bug because it affects the documents order inside segments and
> only a reindex can restore the correct sort order.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16712938#comment-16712938 ] Michael McCandless commented on LUCENE-8592: I'll have a look; this is sneaky. > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Reporter: Jim Ferenczi >Priority: Major > Attachments: LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16712870#comment-16712870 ] Simon Willnauer commented on LUCENE-8592: - the patch looks good to me. Yet, I am not 100% on top of this code if there are other places that need to be fixed. Still +1 to commit. > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Reporter: Jim Ferenczi >Priority: Major > Attachments: LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[
https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16712613#comment-16712613
]
Jim Ferenczi commented on LUCENE-8592:
--
{quote}
Does CheckIndex fail on broken indices? That could be helpful to know whether
users can trust it to check whether they are affected and need to reindex.
{quote}
Yes, with this patch CheckIndex will fail on indices affected by this bug.
> MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural
> sort is reversed
> --
>
> Key: LUCENE-8592
> URL: https://issues.apache.org/jira/browse/LUCENE-8592
> Project: Lucene - Core
> Issue Type: Bug
>Reporter: Jim Ferenczi
>Priority: Major
> Attachments: LUCENE-8592.patch
>
>
> MultiSorter#getComparableProviders on an integer or long field doesn't handle
> MIN_VALUE correctly when the natural order is reversed. To handle reverse
> sort we use the negation of the value but there is no check for overflows so
> MIN_VALUE for ints and longs are always sorted first (even if the natural
> order is reversed).
> This method is used by index sorting when merging already sorted segments
> together. This means that a sorted index can be incorrectly sorted if it uses
> a reverse sort and a missing value set to MIN_VALUE (long or int or values
> inside the segment that are equals to MIN_VALUE).
> This a bad bug because it affects the documents order inside segments and
> only a reindex can restore the correct sort order.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-8592) MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural sort is reversed
[ https://issues.apache.org/jira/browse/LUCENE-8592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16712609#comment-16712609 ] Adrien Grand commented on LUCENE-8592: -- I see, the patch is applying reverseMul on top of the comparison result rather than on values themselves, like search-time sort. That is still subject to bugs but much less likely I guess. +1 to merge. Let's get someone else to have a look at it first, but it would be nice to have it in 7.6 too. Does CheckIndex fail on broken indices? That could be helpful to know whether users can trust it to check whether they are affected and need to reindex. > MultiSorter#sort incorrectly sort Integer/Long#MIN_VALUE when the natural > sort is reversed > -- > > Key: LUCENE-8592 > URL: https://issues.apache.org/jira/browse/LUCENE-8592 > Project: Lucene - Core > Issue Type: Bug >Reporter: Jim Ferenczi >Priority: Major > Attachments: LUCENE-8592.patch > > > MultiSorter#getComparableProviders on an integer or long field doesn't handle > MIN_VALUE correctly when the natural order is reversed. To handle reverse > sort we use the negation of the value but there is no check for overflows so > MIN_VALUE for ints and longs are always sorted first (even if the natural > order is reversed). > This method is used by index sorting when merging already sorted segments > together. This means that a sorted index can be incorrectly sorted if it uses > a reverse sort and a missing value set to MIN_VALUE (long or int or values > inside the segment that are equals to MIN_VALUE). > This a bad bug because it affects the documents order inside segments and > only a reindex can restore the correct sort order. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
