more notes for oracle/opengrok#4317
Using the code fixed to perform correct Path comparisons in both places (on the set of files from FileCollector as well as in IndexDatabase#processFileIncremental()
, the script number 3 (oracle/opengrok#4317 (comment)) this still leads to cases with duplicate documents. This does not happen every time, it might be timing related.
The script goes through solaris-userland Git repository using particular sequence of changesets that were previously known to reproduce the duplicate documents (before the Path comparator fixes): 3b039b16f 170ebb43b 80cc6ae18 52c217090 a609ca278 69a8daead 07e01a4e6 154009177 794af3182 c073248f7 823f2c28e 4a5e3cb85 341f9beb2 653378bce 4f8fe9ee8
Changes to the code:
diff --git a/opengrok-indexer/src/main/java/org/opengrok/indexer/index/IndexDatabase.java b/opengrok-indexer/src/main/java/org/opengrok/indexer/index/IndexDat