MalformedInputException on valid BSON.

[mikelehen]

Hitting a MalformedInputException on a particular bson file. Unfortunately, I can't share it since it has customer data in it. The BSON seems to be valid (bsondump has no problem with it). I'll try to dig in deeper when I get a chance, but logging an issue in case it's useful.

Exception in thread "main" java.lang.RuntimeException: Input length = 1
at com.fasterxml.jackson.databind.MappingIterator.next(MappingIterator.java:124)
at scala.collection.JavaConversions$JIteratorWrapper.next(JavaConversions.scala:575)
at scala.collection.Iterator$class.foreach(Iterator.scala:772)
at scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573)
at bson2json.App$.main(App.scala:37)
at bson2json.App.main(App.scala)
Caused by: java.nio.charset.MalformedInputException: Input length = 1
at java.nio.charset.CoderResult.throwException(CoderResult.java:277)
at de.undercouch.bson4jackson.io.LittleEndianInputStream.readUTF(LittleEndianInputStream.java:293)
at de.undercouch.bson4jackson.io.LittleEndianInputStream.readUTF(LittleEndianInputStream.java:227)
at de.undercouch.bson4jackson.BsonParser.readString(BsonParser.java:570)
at de.undercouch.bson4jackson.BsonParser.nextToken(BsonParser.java:292)
at com.fasterxml.jackson.databind.deser.std.MapDeserializer._readAndBindStringMap(MapDeserializer.java:422)
at com.fasterxml.jackson.databind.deser.std.MapDeserializer.deserialize(MapDeserializer.java:314)
at com.fasterxml.jackson.databind.deser.std.MapDeserializer.deserialize(MapDeserializer.java:26)
at com.fasterxml.jackson.databind.MappingIterator.nextValue(MappingIterator.java:189)
at com.fasterxml.jackson.databind.MappingIterator.next(MappingIterator.java:120)
... 5 more

[michel-kraemer]

Without test data this is going to be a tough one. MalformedInputException usually means the string contains invalid characters (in this case invalid for UTF-8). Can you confirm this? Would it be possible to provide a test file that contains only this specific string?

Cheers,
Michel

[michel-kraemer]

I'm about to release a new version of bson4jackson that is compatible to Jackson 2.3. I will close this issue due to inactivity. If you get to reproduce the problem with data that can be shared feel free to open the issue again.

Thanks,
Michel

[bguerout]

Hi,

Someone has described same behavior on Jongo ML : https://groups.google.com/forum/?fromgroups=#!topic/jongo-user/Lj8bvsBjEao

I was able to reproduce the bug described in this issue with bson4jackson 2.3.0 using the content.txt (json file) provided into the post.
I'm going to investigate

[bguerout]

Probably related to something like http://stackoverflow.com/questions/14381704/the-type-generates-a-string-that-requires-more-than-65535-bytes-to-encode-in-utf

[bguerout]

Hi @michel-kraemer, can this issue be reopened ?

[michel-kraemer]

@bguerout I'm sorry for the late answer. End of the year is always a busy time at the office! :-/ I'll try to investigate the issue later today or tomorrow. If it can be reproduced, it can be fixed!

[bguerout]

No problem, i know what you are talking about :-)

[michel-kraemer]

The commit I just uploaded should fix the problem. Can you please test it and report if the problem persists? Building bson4jackson should be as simple as running ./gradlew install but I can send you a binary if you like.

[bguerout]

Indeed this is fixed.
Thanks

[michel-kraemer]

Thanks for the feedback. If you need a new version just let me know.

[mikelehen]

Thanks! Whenever you release a new version (to maven), I can easily check to see if my case is fixed as well.

[michel-kraemer]

Version 2.3.1 has just been released. Should be available at Maven central in a couple of minutes.

Cheers,
Michel

[mikelehen]

Sorry for the delay, but 2.3.1 does indeed fix the issue I was hitting. Thanks very much!