Vladimir Kotal
Vladimir Kotal
There are 2 areas where test coverage of `DirectoryListingTest` can be improved: - test subdirs - currently all the `FileEntry` constructors are invoked with `subdirs` being `null` - validate Lines#/LOC...
As identified when [indexing AOPS](https://github.com/oracle/opengrok/discussions/4038#discussioncomment-5781989), the first phase of indexing (when the history cache is generated) can take a long time due to one of the repositories having much bigger...
While working on the annotation cache, I noticed a fair bit of confusion/deficiency in the code when handling history related tunables. Specifically, there are 2 problems: 1. confusion between general...
As suggested in https://github.com/oracle/opengrok/discussions/4262#discussioncomment-5507017 , it should be possible to support history based reindex even without history cache. This would require some other way to store the latest indexed changeset...
As revealed in #1491, the history generate flags (as specified with `OPENGROK_GENERATE_HISTORY` in the `OpenGrok` shell script or with `-r` option for indexer directly) fail to be honored for dirbased...
The `ConfigMerge` tool should make sure that (canonicalized) paths for the base and new files are different to avoid nasty surprises. Inspired by https://github.com/oracle/opengrok/discussions/4133#discussioncomment-4402540_
As mentioned in PR #4232, the dependency check should be run periodically, outside of the usual builds.
Report changeset ID next to the version when starting the Docker container.
The properties in `Configuration` pointing to various roots can be defined as `Path` objects which they really are.
Specifically, Docker build fails when running `./dev/dockerhub_readme.py` with: ``` INFO:__main__:DOCKER_USERNAME is empty, exiting Error: Process completed with exit code 1. ``` Seen e.g. in #4139.