DIAMOND v2.2.2

• Fixed a crash in the view workflow (present since v2.2.0).
• Fixed a bug that could cause an error clustering a database with a single sequence.
• Fixed a potential error of too many open files in the clustering workflows.
• Fixed a possible inter-platform inconsistency of the approx_pident output field caused by architectural differences in implementing fused multiply-add.
• Using the --log option will no longer produce a diamond.log file.
• The --verbose/-v option is deprecated and has no effect.
• Removed a potential inter-system non-determinism caused by toggling query-indexed mode by detection of L3 cache size.
• Fixed the --header option not correctly working for clustering workflows.
• Removed the --unaligned-targets option.
• Fixed an error reading FASTQ files with line breaks in the sequence or quality lines.
• Improved the performance of loading FASTA files containing DNA sequences.
• Fixed an issue of not correctly loading compressed streams on Windows.
• Fixed the Windows release version missing zstd support.
• Fixed query-indexed mode not being correctly triggered when using the --dbsize option in conjunction with a FASTA database.
• Fixed a bug that could cause incorrect ordering in the output file when using the --dbsize option in conjunction with a FASTA database.
• Fixed a crash when using --mp-init on Windows.
• Fixed a potential deadlock when using --multiprocessing.
• Added the option --daa-build-version to overwrite the diamond build version written to DAA files.
• Disabled suppression of terminal logging output to stderr when requesting output be written to stdout.