plougher
commented
3 years ago You raise some interesting issues. First whether it should be possible to (re-)create a Squashfs filesystem from pseudo file definitions. Second, the lack of timestamp information in pseudo definitions makes it impossible to recreate a filesystem using pseudo definitions as timestamping cannot be preserved.
As I felt this is a useful thing to do, I spent some time adding the necessary functionality. This additional functionality is split into three areas:
-
Adding timestamps. The pseudo definitions have been extended to add timestamps. New upper-case pseudo definitions have been added which take timestamps. For example, the new set of pseudo definitions is
filename d mode uid gid filename m mode uid gid filename b mode uid gid major minor filename c mode uid gid major minor filename f mode uid gid command filename s mode uid gid symlink filename i mode uid gid [s|f] filename l filename filename L pseudo_filename filename D time mode uid gid filename M time mode uid gid filename B time mode uid gid major minor filename C time mode uid gid major minor filename F time mode uid gid command filename S time mode uid gid symlink filename I time mode uid gid [s|f] filename R time mode uid gid length offset
where the upper-case versions take a timestamp. This can be specified as an absolute numeric value, which is taken to be the number of seconds since the epoch of 00:00:00 UTC on 1 January 1970. A string can also be used, which is passed to the "date" command for interpretation, and can be anything that "date" considers a valid date.
- Adding pseudo definitions. The previous pseudo definitions lacked the ability to create hard-links, and also lacked the ability to recreate regular files, as there was no capability to store data in a pseudo definition.
Without that ability it is impossible to recreate a Squashfs filesystem using pseudo definitions. These pseudo definitions have been added.
- An option has been added to Unsquashfs which will produce a "pseudo file" definition of the input Squashfs filesystem. This pseudo-file can be passed to Mksquashfs to re-create the Squashfs filesystem.
Usage:
% unsquashfs -pseudo-file name squashfs-filesystem
This will create a pseudo file called "name", which is a representation of the supplied "squashfs-filesystem".
% mksquashfs - squashfs-filesystem -pf name
This will take the pseudo-file "name", and it will produce the squashfs filesystem called "squashfs-filesystem". This will be identical in content to the original filesystem supplied to Unsquashfs.
keeely
The pseudo file options allow one to create block devices, files, directories and (with v4.4) symlinks. However they don't allow one to specify the file timestamp as far as I can see, so you have to have (presumably) whatever time the input file or directory has. If the file is dynamically generated, I'm going to assume it takes the system time at the point it was generated. This may not be what you want, so the ability to specify timestamp in the pseudo-file specification line would be a welcome feature.
I could find no easy way of generating a file full of pseudo-file entries from an existing squashfs, so I started writing something to do so. Then it struck me someone must have already written this before. Has someone done this?