[Pkg-ime-devel] Reminder: Freeze of Wheezy is approaching

[Aron Xu]

On Sun, Jan 29, 2012 at 09:43, Osamu Aoki <osamu at debian.org> wrote:
> Hi,
>
> On Sat, Jan 28, 2012 at 10:21:59PM +0800, Aron Xu wrote:
>> On Sat, Jan 28, 2012 at 21:20, Osamu Aoki <osamu at debian.org> wrote:
>> > HI,
>> >
>> > On Sat, Jan 28, 2012 at 08:10:19PM +0800, Aron Xu wrote:
>> >>
>> >> [...]
>> >> Package scim-python needs some love because it is RC buggy. Though the
>> >> package is not that widely used today, I still wonder if there are
>> >> people running it and would like to have it for Wheezy.
>> >
>> > My honest recommendation is to drop whole SCIM packages.
>> >
>> > After spending enough time for making SCIM for squeeze while struggling
>> > with their old autotools codes, I was going to ask removal of them.
>> > SCIM lacks gtk3 support now and scim-bridge which was the last hope for
>> > gtk3 and qt4 support is about to be dropped.
>> >
>> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=642371
>> >
>> > Zombie upstream is something very difficult to work with.  They are
>> > there but not as much help as the original authors.
>> > ...
>
> (Just to be sure: SCIM now has active upstream.  But new upstream
> admitted he is not strong in technical background to keep up source with
> new GTK or QT tools.)
>

And still, there are some complicated problems in scim-python.

>> I don't want SCIM either, since it is actually dead and ancient. But
>> I'm in doubt that whether there are still someone need it for some
>> features other IME (ibus, for example) doesn't have?
>
> I do not think there is anything left which we need SCIM... With current
> situation, security support is weak from upstream.
>

I sent some emails to Chinese communities asking the users opinon
yesterday, and I got some objections that said SCIM supports some
features that other existing input method frameworks do not have:
1.Uniform user defined words
2.Many many usable tables (as in table based input methods)

> Rolf is the one who insisted to keep SCIM. (He pushed many patches
> upstream but seems to have difficulties to package new upstream.
> Anyway, I cced him with this email so he will know this issue.
>
> As I remember for scim (1.4.9-3), in which I used autoreconf.  It caused
> all sorts of headache.  autoreconf causes build failure if autotools
> changes significantly unless upstream is active enough to keep up these
> changes.
>

I can help him for a single version if he insist on keeping SCIM, but
I don't want to `maintain' it.

> I also realized ibus is FTBFS now.  I will work on this FTBFS and maybe
> on packaging ibus-fbterm.  This seems missing in ibus family. (SCIM does
> not have equivalent)
>

IIRC ibus-fbterm is already abandoned upstream for sometime.

> FYI: (From Daiki)
> FTBFS of ibus can be fixed by using fix documented in automake upstream
> http://git.savannah.gnu.org/cgit/automake.git/commit/?id=4e4dae50
>
> Osamu



-- 
Regards,
Aron Xu