Itermittent NPE in remove

[cpotter1]

In DataService, on line 227, seeing this problem periodically.

java.lang.NullPointerException
at org.dizitart.no2.internals.DataService.remove(DataService.java:228)
at org.dizitart.no2.internals.NitriteService.remove(NitriteService.java:372)
at org.dizitart.no2.internals.DefaultNitriteCollection.remove(DefaultNitriteCollection.java:360)
at org.dizitart.no2.internals.DefaultNitriteCollection.remove(DefaultNitriteCollection.java:349)
at

Looks like you are not checking for null on "removed" before calling getRevision:
Document removed = underlyingMap.remove(nitriteId);
int rev = removed.getRevision();

[cpotter1]

This is using 3.0.0

[anidotnet]

Can you please give me a reproducible code? The region from where it is throwing is as follows:

for (Document document : cursor) {
    NitriteId nitriteId = document.getId();
    indexingService.removeIndexEntry(document, nitriteId);

    Document removed = underlyingMap.remove(nitriteId);
    int rev = removed.getRevision();  <<< NPE

It means that nitriteId is not there in the underlyingMap, but the nitriteId is obtained from a document iterating the cursor which is generated from the same underlyingMap.

So I am curious why is it happening. My initial assumption is, you are removing the document from a different thread and it is hitting a race condition. So If you could give me a reproducible code it would be better for me to dig down and take proper measure.

[cpotter1]

Hi Anindya, I am thinking the same thing as you... However, the problem is very hard to reproduce. Was hoping you could just write some defensive code for this case.

…

On Wed, May 30, 2018 at 3:18 AM, Anindya Chatterjee < ***@***.***> wrote: Can you please give me a reproducible code? The region from where it is throwing is as follows: for (Document document : cursor) { NitriteId nitriteId = document.getId(); indexingService.removeIndexEntry(document, nitriteId); Document removed = underlyingMap.remove(nitriteId); int rev = removed.getRevision(); <<< NPE It means that nitriteId is not there in the underlyingMap, but the nitriteId is obtained from a document iterating the cursor which is generated from the same underlyingMap. So I am curious why is it happening. My initial assumption is, you are removing the document from a different thread and it is hitting a race condition. So If you could give me a reproducible code it would be better for me to dig down and take proper measure. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#76 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAoz9_0y9m_icqKvDrxK5ZSzlUk9KEb2ks5t3ke0gaJpZM4USExO> .

[anidotnet]

@cpotter1 can you test the latest snapshot now? - 3.0.1-SNAPSHOT