Vladimir Kotal

Also, some level of UI testing would be useful for this change, I believe. In particular of the slider for the search results.

Of course, running this on Travis might not buy us much as I only reproduced the issues when running them on Solaris machines with at least 24 CPU cores.

The `IndexDatabase#indexParallel()` implementation uses [custom ForkJoin pool which some places do not recommend](https://www.baeldung.com/java-when-to-use-parallel-stream#3-custom-thread-pool). There are some indications that the [intended parallelism level does not have to be achieved in ForkJoin...

Is the CVS Root remote or local ?

It's definitely weird that logs do not give any hints whatsoever. I'd try simulating this on command line - start with one repo and keep adding. One hypothesis would be...

Also, for the cases where CVS HistoryException with exitcode 1 is thrown could you share the log context ?

I did a small experiment. My laptop has 4 virtual CPUs; according to `Indexer.java#main()` the number of threads is determined as: `126 int noThreads = 2 + (2 * Runtime.getRuntime().availableProcessors());`...

It depends on what is correct for CVS or what do the users expect. Certainly we can modify the history executor for CVS to treat `HEAD` in special way.

I mean if the branch of the repo is `HEAD` (i.e. was checked out with `-rHEAD` so there exists `CVS/Tag` file which contains `THEAD`) then instead of running `hg log...

This was likely reintroduced with #4516. It may need the changes #4524 for proper fix.