Bug#605605: apt: uninteresting NEWS.Debian
David Kalnischkies <email@example.com> writes:
> On Wed, Dec 1, 2010 at 19:26, Jakub Wilk <firstname.lastname@example.org> wrote:
>> | * Already included in the last version but now with better documentation
>> | is the possibility to add/prefer different compression types while
>> | downloading archive information, which can decrease the time needed for
>> | update on slow machines. See apt.conf (5) manpage for details.
> I was told on d-devel that this is a great feature for e.g. s390 machines
> and of general usefulness, thats why it was enhanced and got this entry
> in the first place?
It is a great feature and something users should be made aware of. They
do need to modify their config to benefit from it. This should be
mentioned in NEW once. Only remove this if the last version already had
an entry for it please.
>> | * APT manages his manpage translations now with po4a, thanks to Nicolas
>> | François and Kurasawa Nozomu, who also provide the ja translation.
>> | Thanks to Christian Perrier we have already a fr translation and
>> | a few more are hopefully added in the near future.
> Having a translation of a manpage looks for me like a great feature.
> People get pinning for example always wrong, now they can read in their
> native tongue what is wrong. ;) Not to mention the various config options?
> And last but not least it encourages maybe more translations? :)
> We have btw currently complete documentation in de, es, fr and pt - ja only
> manpages and pl only some manpages?
> After all, that are currently 1048 paragraphs(!) to translate?
> In my humble opinion, thats a pretty big achievement.
Obviously manpages are a good thing. But no reason for a NEWS entry. It
really does not matter to the user how the manpages for apt where
generated. Just that they are there.
>> | * This version also introduces some _experimental_ configuration options
>> | to make more aggressive use of dpkg's triggers. If you want to help
>> | testing these _experimental_ options see apt.conf (5) manpage.
> Again something people asked for (e.g. for possible use in d-i).
> The major obstacle against defaulting to them left is the progress reporting?
> Something which isn't shown in APT at all but its front-ends?
> It avoids some pretty hard bugs caused by the need to split too long
> dpkg calls, which happens sometimes at the wrong position - and is
> even a small speed improvement - maybe even bigger if #526774 would
> show progress?
> But yes, this one could possible be removed.
> Properly we could/should add others?
> It very much depends on which angle you look at APT and as it used in many
> ways you can have many angles? (best example is the first point)
> Given that this news-entry is more than a year old,
> it can't be useless for everyone after all.
I think this should stay for unstable/experimental. But it has no place
in stable/testing. The feature is experimental, it says so itself, and
stable/testing users should not be invited to play with such dangerous
Just my 2c,