On Thu, Aug 15, 2002 at 06:45:05PM +0100, Colin Watson wrote:
> It should be safe to make libc depend on this now, in order to pull in
> libdb.so.2.
>
> Accepted:
> libdb1-compat_2.1.3-4_i386.deb
> to pool/main/d/db1-compat/libdb1-compat_2.1.3-4_i386.deb
Upgrading from woody to the current glibc (-13, from -14's changelog I presume
it's the same) breaks:
apache
apache-common
apache-dev
apache-perl
apache-ssl
caudium-php4
gnucash
kpackage
libapache-csacek
libapache-mod-auth-mysql
libapache-mod-auth-pam
libapache-mod-auth-shadow
libapache-mod-auth-useragent
libapache-mod-backhand
libapache-mod-cgi-debug
libapache-mod-dav
libapache-mod-filter
libapache-mod-gzip
libapache-mod-index-rss
libapache-mod-interchange
libapache-mod-ldap
libapache-mod-lisp
libapache-mod-mp3
libapache-mod-python
libapache-mod-random
libapache-mod-relocate
libapache-mod-repository
libapache-mod-speedycgi
libapache-mod-text2html
libapache-mod-trigger
libapache-mod-webapp
libapache-mod-witch
libapache-mod-xslt
libdb2-util
libdb3-util
medusa
mmorph
nmh
ocaml-base
perspic
php3
php4
php4-cgi
python1.5
python2.1
python2.2
radiusd-livingston
smail
webalizer
xkbsel
xkbsel-gnome
I can't see any obvious way to conflict with each of those separately, and
I guess you don't want to add well over ten lines of Conflicts: to take care
of them. So please add this dependency, already...
According the update_excuses, there are also the following RC bugs:
* #153263: libc6-prof: unusable on arm
* #151804: libc6: libbind security hole
* #152099: an empty directory string in LD_LIBRARY_PATH is
interpreted as '.'
* #155606: provide symbols .hidden in gcc 3.1/3.2 when building
glibc for ppc
* #156841: glibc_2.2.5-13(alpha/unstable): FTBFS: non-PIC code in a
dynamic lib
Presumably 156841 was fixed in -14; I've heard rumours that 153263 isn't
an issue anymore either. 152099 seems like it's not being treated by
you guys as "grave", so it should probably be downgraded or closed so
the testing scripts know what's going on. 151804 has a remark indicating
it's been fixed, and should be closed. I don't really follow why 155606
is critical (important, sure, but "breaks unrelated software, the whole
system, causes serious data loss, or introduces a security hole" ?).
Since -13 bumped the shlibs, this has been blocking every package updated
in the last two weeks from entering testing, which isn't healthy.
Cheers,
aj
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
