v2.9.0

Fixed

• gno update no longer loads the PDF, Word and PowerPoint parsers before indexing a collection that has none of those files. Every update used to load them at its first file, to check for multi-record formats. On a cold Windows machine that took 9 to 11 seconds before the update printed anything. The parsers now load when the first such file is converted. A build that cannot load them still fails only those file types, as before.

Added

• GNO_PHASE_TRACE=1 prints one line per startup phase of a CLI command to stderr (write lease, index open, migrations, walk, each file, file processor), so a slow or stuck run shows where it is. See Troubleshooting, "Slow Indexing".