Kai Krakow

BTW: There may be a problem with 0.6.2, it has a bug somewhere with cmdline parameters. Could you try 0.6.1, or preferably master which should become 0.7 at some time...

Your data is safe at any point (except for kernel bugs): bees always uses the kernel to deduplicate extents, this is transaction safe, concurrent writes won't damage any data, bees...

@Zygo but for the "--help" case this looks a lot like what I fixed with https://github.com/Zygo/bees/pull/136... ``` # bees --help # [...] 2020-04-19 00:11:17 1166966.1166968 status_report: Exiting thread status_report, 0.001...

You may want to use long options with syntax `--long-option=NUMBER`... Usually, both formats are accepted but it may confuse some parsers if you omit `=`. But please check your options...

LTO with gcc 11+ is really problematic... If you compile the toolchain itself with lto, it will make libstdc++ completely unusable because most stdlib symbols are missing for linking.

> Which distro is having problems? Gentoo here... The problem is compiling gcc itself with `-flto` - it breaks the symbol visibility. Some other libraries may be affected, too. I've...

I'd prefer an option to make bees use read-only extents as the base extent for dedup. Am I guessing right this wouldn't mutate read-only snapshots then? OTOH, I myself am...

@PeeJay Why not meson/ninja? It seems even simpler from what I've seen, plus it seems to have better dependency tracking.

@HaleTom Bees dedupe offsets can only be as fine-grained as the block size so it's unlikely to match anything inside and outside a tar file unless your tar headers are...

BTW: Thinking about it and using variable chunking (instead of static/block-based chunking) should enable bees to work around boundary shifting problems as introduced by tar files you mentioned. It should...