at bottom :-
On 3/11/15, A Mennucc1 <mennu...@debian.org> wrote:
> Il 07/03/2015 10:20, shirish शिरीष ha scritto:
>> I am still not sure why it says 'Delta is not present' - I wish I
>> could get some more precise output/statement -
> one possible cause: debdelta will not created deltas for packages that
> are too small; but I am not sure that this covers all cases...
>
> a.
Dear Andrea,
In the above case at least, the .deb package was about 4.1 MB
(compressed size).
See :-
[/var/cache/apt/archives] - [188]
└─[$] ll -h | grep python2.7-doc
[21:50:06]
-rw-r--r-- 1 root root 4.1M Mar 1 17:59 python2.7-doc_2.7.9-2_all.deb
I do understand that for small packages it doesn't do the diff. (the
ones which are 7KB or 8KB or something like that.)
It is very difficult to know why sometimes delta takes forever to
happen or doesn't happen and why sometimes it happens instantly (apart
from the size, for e.g. I waited for over 24 hrs. for the gcc deltas
to come but none of them came).
One case where there will be a failure is if the package used from
'experimental' suite or 'unstable' suite a new version comes in
'unstable' while your suite is 'testing'
[/etc/apt] - [199]
└─[$] cat preferences
[21:56:00]
Package: *
Pin: release testing
Pin-Priority: 600
Package: *
Pin: release experimental
Pin-Priority: 1
Package: *
Pin: release unstable
Pin-Priority: 1
In such a case the user will never come to know that the package can't
be diffed because it's from a different suite altogether. AFAIK the
diffing works only when both the packages are from the same suite
'testing.old.version.1 < testing.new.version.2 = diff > 80 or
'unstable/sid.old.version.1 < testing.new.version.2 = diff > 80. (but
that too if your /etc/apt/sources.list uses unstable/sid as your repo.
and it's total sid repo.)
In the case where the package is from 'experimental.old.version.1 <
unstable.new.version.2 don't think we would have debdelta as this has
not been done and most of the time it might be a failure as well
(dunno) or 'experimental.old.version.1 < testing.new.version.2' where
again such diffs. wouldn't happen or lastly 'unstable.old.version.1 <
unstable/testing.new.version.2' .
While the experimental ones are clearly out of the picture, last
couple of corner scenarios may be something that can be worked with
assuming you have good enough hardware to run those permutations and
combinations and getting the deltas as well as enough bandwidth to
distribute them.
It would at the very least make an interesting read of what you think
of the above test-cases/scenarios and if there are any thoughts of
tackling them in the near future.
Looking forward to hear from you.
--
Regards,
Shirish Agarwal शिरीष अग्रवाल
My quotes in this email licensed under CC 3.0
http://creativecommons.org/licenses/by-nc/3.0/
http://flossexperiences.wordpress.com
EB80 462B 08E1 A0DE A73A 2C2F 9F3D C7A4 E1C4 D2D8
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
