From 7bcc508dab2941549d0d359274f21f9389f75e3f Mon Sep 17 00:00:00 2001 From: Marcus Holland-Moritz Date: Sat, 23 Aug 2025 07:10:30 +0200 Subject: [PATCH] chore: update TODO --- TODO | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) diff --git a/TODO b/TODO index d1e03f3c..d79efb59 100644 --- a/TODO +++ b/TODO @@ -1,6 +1,14 @@ - More "real" benchmarks -- Entry v2 +- Entry v2 - more memory efficient implementation + +- Support xattrs + +- Proper memory manager so `-L` will actually limit memory correctly + +- Incremental filesystem layers; this might support layer trees rather + than just a single linear layer sequence, i.e. we can add multiple + different child layers on top of a common parent layer. - Mounting lots of images with shared cache? - Communication could be implemented using xattrs @@ -12,17 +20,6 @@ - However, inode numbers aren't straightforward in case of adding/removing images. -- Metadata rebuilding (see also below). It should be - possible to rebuild the metadata block and change - features like packing without having to rebuild the - whole image. - -- Allow changing the block size during recompression. - This should be possible, along with updating the chunks - in the metadata, and the operation should be perfectly - reversible. At block boundaries, chunks must be split - or merged, but that should be easily doable as well. - - Sparse files: implement by adding a "chunk type" field to each chunk; for now, there are only "normal" and "sequence" chunks, with sequence chunks encoding as @@ -40,8 +37,6 @@ - Add support for logging to file (with different level?) -- Add support for libarchive filters in dwarfsextract. - - When hashing, start by only hashing the first, say, 4KiB, and only if the hashes are identical, hash the whole file