uim-common hangs in old prerm
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
uim (Ubuntu) |
Confirmed
|
Undecided
|
Unassigned |
Bug Description
dist-upgrading from dapper to edgy stops while running the prerm script of the old uim-common from 1.0.0-1u1 to 1.2.1-3:
--- snip ---
Préparation du remplacement de uim-common 1:1.0.0-1ubuntu1 (en utilisant .../uim-
ERROR: unbound variable (errobj is-set-ugid?)
*backtrace*
>>(is-set-ugid?)
>>(if (is-set-ugid?) (list (string-append (sys-pkglibdir) "/plugin")) (filter string? (append (list (getenv "LIBUIM_
>>(define uim-plugin-
>>(require "plugin.scm")
>>(require "init.scm")
>>(*catch (quote errobj) (require "init.scm"))
>>(eq? (quote *init.scm-loaded*) (*catch (quote errobj) (require "init.scm")))
ERROR: unbound variable (errobj custom-
ERROR: unbound variable (errobj custom-
ERROR: unbound variable (errobj plugin-alist)
ERROR: wta to car (errobj t)
[hangs; I pressed ctrl-C here]
dpkg : erreur de traitement de /var/cache/
dpkg : avertissement - ancien script pre-removal a été tué par le signal (Interruption)
--- snip ---
[the last dpkg message says that the pre-rm script of the old package has been interrupted]
The same error occurs for uim-anthy, uim-canna, uim-m17nlib and uim-skk.
Changed in uim: | |
status: | Unconfirmed → Confirmed |
Right, the problem is it wants libuim0, which was in Dapper but has been replaced by something else in Edgy (libuim3, I believe). It seems the prerm script needs this, and as it's been removed, it fails. So, go to /var/lib/ dpkg/info/ uim-common. prerm, put exit 0 right after the comments (lines beginning with '#'). I haven't had any other problems like you have, but I assume the same applies for those packages, just edit /var/lib/ dpkg/info/ <package> .prerm instead.
You are now able to remove uim-common (or the other packages) and them apt-get install it again, replacing it with the updated and working version, or replace it by apt-get update-ing it. It will replace the prerm script, so you don't need to edit it back, and I'm guessing, though I can't be sure, not excecuting the prerm script is ok so long as the package is replaced, so do this at you own risk, but it worked fine for me.