Re: "Calculated and written ECMA-119 tree end differ" under very specific circumstances

[Thomas Schmitt]

Hi,

Valtteri Vuorikoski wrote:
> -for_backup with zisofs disabled (dev set to /dev/zero) also fails:
> [...]
> With only "-xattr any" and zisofs disabled: [failure]
> [...]
> Invocation with "-xattr on" finishes successfully.

Does "-for_backup -xattr on" plus zisofs work ?
("-xattr on" will override the "-xattr any" of "-for_backup".)


I can reproduce the problem by using my $HOME as carrier tree for long
xattr:

  dummy="123456789012345678901234567890123456789012345678901234567890"
  dummy="$dummy""123456789012345678901234567890123456789012345678901234567890"

  xorriso -for_backup -outdev stdio:/dev/null \
          -map /home/thomas / \
          -setfattr_r user.dummy "$dummy" / --

Again, xorriso-1.3.8 of 2014 does this without error and 1.4.0 of 2015
shows the failure.
So that development cycle not only introduced the bug about zisofs.
(This all was done because Linux isofs has a very peculiar interpretation
of the SUSP specs. So using 1.3.8 would not be an option unless you
intent to never mount the result on Linux.)

I will next try to reduce the size of the carrier tree so that i have
a better chance with the debugger.


Have a nice day :)

Thomas