On 2025-08-09 22:47:30 -0700, Michael Paoli wrote: > Length of a file can change while tar is backing it up, tar generally > handles this, > though it may complain/warn about it. > And to the compression formats, it's just a stream of data, > they know nothing about tar. So, the blocks of compressed data may > have part of one large file, or maybe thousands of files. The compression > formats generally don't know or care when it comes to tar archive or the like, > so there's not really much info they could provide to tar about what's within > ... for the most part all they really know is its data, and perhaps > also bits like > how long the block is compressed, and uncompressed, and perhaps some > hash of that data or the like.
This is why I said that in the hope to offer a better chance of optimization of compression, it is better to make "tar" do the compression (either internally or with an external xz). xz offers a way to choose at what points compressed blocks will be split. -- Vincent Lefèvre <vinc...@vinc17.net> - Web: <https://www.vinc17.net/> 100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/> Work: CR INRIA - computer arithmetic / Pascaline project (LIP, ENS-Lyon)
