[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ia32-libs depends on ia32-apt-get ?



Goswin von Brederlow wrote:

> Didier 'OdyX' Raboud <didier@raboud.com> writes:
>> Norbert Preining wrote:
>>> - calling /usr/share/ia32-apt-get/convert-all-sources.list
>>
>> Which horribly breaks with anything a little custom (proxies, custom
>> repositories, ...) and fills your /etc/apt/sources.list.d/ with ia32-apt-
>> get.{i386,amd64} copies of all your pet sources.
> 
> Examples please.

Hi Goswin, 

Here is an example from my laptop :

I don't have an /etc/apt/sources.list , so installation fails (#534979). In 
my /etc/apt/sources.list.d, I have 4 files :

* 00_particulars.list, which contains various repositories, some of them
  commented, some of them not, mostly unused now.
* 10_apt-proxy.list, which contains lines for my home apt-proxy-ng with
  testing, t-p-u, unstable, experimental and testing/updates.
* 20_std.list.dis and 30_local_pbuilder.list.dis, both disabled for aptitude
  (fallback and intent).

You'll find them in the attached apt_sources_list_d.tar.bz2.

# touch /etc/apt/sources.list; aptitude install ia32-apt-get
(Does it really generate a gpg key as root, asking me to move the mouse ?)

Then, in /etc/apt/sources.list.d/, in addition to _MY_ 4 files, I have the 
following list :


ia32-apt-get.amd64.00_particulars.list
	(Completely broken, no valid repository)
ia32-apt-get.amd64.list
	(empty)
ia32-apt-get.i386.ia32-apt-get-transitional.list
	(transitional-i386/main not found => nothing valid)
ia32-apt-get.amd64.10_apt-proxy.list
	(Completely broken, no valid repository)
ia32-apt-get.i386.00_particulars.list
	(Completely broken, no valid repository)
ia32-apt-get.i386.list
	(empty)
ia32-apt-get.amd64.ia32-apt-get-transitional.list
	(transitional-amd64/main not found => nothing valid)
ia32-apt-get.i386.10_apt-proxy.list
	(Completely broken, no valid repository)
ia32-apt-get-transitional.list
	(WORKS ! - Contains ia32-libs{,-gtk} 1:3.0)

So among the 9 packages prepared by ia32-apt-get postinst, one is working. 
All others are broken.

You'll find them in the attached apt_sources_list_d_after.tar.bz2.

Is that enough of an example ?

>>> - calling apt-get update from the commandline
>>
>> It dpkg-diverts apt-get but not aptitude... How can we accept to see
>> apt-get diverted for such a hack ?
> 
> You don't have too but then you won't get 32bit support beyond the
> verry core libs needed for gcc-multilib.

Sorry, but if I install wine, I don't have the choice to have apt-get _not_ 
diverted…

> The reasons for ia32-apt-get are this:
> 
> - multiarch is still not there.
> - ia32-libs source with all the additional request libs grows to about
>   1GB in size of which everything is pure duplication.
> - ia32-libs contains so many libs that it needs a new upload every
>   other day (or is constantly out of sync like it always was).
> - ia32-libs can only cover unstable or testing but not both.
> - ia32-libs has no security support but security bugs.
> - ia32-libs doesn't ensure library versions are new (or old) enough
>   to work with 3rd party debs. They easily miss a library or get a
>   wrong version.
> - ftp-master has vetoed splitting ia32-libs into individual packages
>   and shown a strong dislike to ia32-libs as it was.
> - ia32-libs doesn't allow to install 3rd party 32bit debs or use 3rd
>   party apt repositories with 32bit packages.

I very much understand all those reasons. I still think that the actual 
response ia32-apt-get provides is actually not good enough to let things 
like wine rely on.

I would largely prefer if ia32-* in its actual shape would be released in 
experimental (where, with this level of touching the base of Debian 
repositories handling, it should sit) and version 2.7 uploaded back in 
Sid...

Regards,

-- 
Didier Raboud, proud Debian user.
CH-1802 Corseaux
didier@raboud.com

Attachment: apt_sources_list_d_after.tar.bz2
Description: application/bzip-compressed-tar

Attachment: apt_sources_list_d.tar.bz2
Description: application/bzip-compressed-tar


Reply to: