Hi Holger,
> │ │ │ │ │ --rw-r--r-- 0 1111 1111 255 Jan 1 1970 epub/META
> │ │ │ │ │ +-rw-r--r-- 0 2222 2222 255 Jan 1 1970 epub/META
Hm, I can confirm this in Git (currently 3f4ba2fb8). Indeed, it
appears to be already handled with:
my $uid_len = ord(substr($field, $pos + 5, 1));
my $gid_len = ord(substr($field, $pos + 6 + $uid_len, 1));
$result
.= pack("CCx${uid_len}Cx${gid_len}", 1, $uid_len, $gid_len);
… and this code is run, padding with enough nulls (the "x" format
identifier for Perl's pack[0] method). However, it does not seem to
result in a normalised file when shown with bsdtar -tvf.
(I would immediately commit an "xfailing" test but the .epub in
question is 800K so I will hold off on that until I find a smaller
testcase. Do let me know if you have one.)
[0] http://perldoc.perl.org/functions/pack.html
> [..] this 'find epub -exec touch -d @0 {} \;' also looks like
> something strip-nondeterminism should do.
We normalise the timestamps of .zip/.epub members so this should
not be a problem.
Regards,
--
,''`.
: :' : Chris Lamb
`. `'` la...@debian.org 🍥 chris-lamb.co.uk
`-
_______________________________________________
Reproducible-builds mailing list
Reproducible-builds@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/reproducible-builds
