TestDoc failure [LUCENE-8096]

ant test -Dtestcase=TestDoc -Dtests.method=testIndexAndMerge -Dtests.seed=8E42CE481D9FEA92 -Dtests.slow=true -Dtests.locale=es-CO -Dtests.timezone=Pacific/Fiji -Dtests.asserts=true -Dtests.file.encoding=US-ASCII

[junit4] ERROR 0.18s J0 | TestDoc.testIndexAndMerge <<<
[junit4] > Throwable #1: java.io.IOException: file "_merge3.fdx" was already written to
[junit4] > at __randomizedtesting.SeedInfo.seed([8E42CE481D9FEA92:E7185A13B36323A7]:0)
[junit4] > at org.apache.lucene.store.MockDirectoryWrapper.createOutput(MockDirectoryWrapper.java:654)
[junit4] > at org.apache.lucene.store.TrackingDirectoryWrapper.createOutput(TrackingDirectoryWrapper.java:43)
[junit4] > at org.apache.lucene.codecs.compressing.CompressingStoredFieldsWriter.<init>(CompressingStoredFieldsWriter.java:113)
[junit4] > at org.apache.lucene.codecs.compressing.CompressingStoredFieldsFormat.fieldsWriter(CompressingStoredFieldsFormat.java:128)
[junit4] > at org.apache.lucene.codecs.lucene50.Lucene50StoredFieldsFormat.fieldsWriter(Lucene50StoredFieldsFormat.java:183)
[junit4] > at org.apache.lucene.codecs.asserting.AssertingStoredFieldsFormat.fieldsWriter(AssertingStoredFieldsFormat.java:48)
[junit4] > at org.apache.lucene.index.SegmentMerger.mergeFields(SegmentMerger.java:213)
[junit4] > at org.apache.lucene.index.SegmentMerger.merge(SegmentMerger.java:103)
[junit4] > at org.apache.lucene.index.TestDoc.merge(TestDoc.java:227)
[junit4] > at org.apache.lucene.index.TestDoc.testIndexAndMerge(TestDoc.java:188)
[junit4] > at java.lang.Thread.run(Thread.java:748)

---
Migrated from [LUCENE-8096](https://issues.apache.org/jira/browse/LUCENE-8096) by Adrien Grand (@jpountz)

Start by running the reported ant test command with seed 8E42CE481D9FEA92. Inspect TestDoc.testIndexAndMerge and TestDoc.merge at the reported stack-trace locations, then trace the duplicate _merge3.fdx creation through MockDirectoryWrapper.createOutput. Done means the failure is understood and the test no longer reports that the file was already written.

Written by the indexing model from the issue text.