Please review this list of potentially missing .so symlinks.
by Jef Spaleta
Back story:
talking to che in #fedora-devel, and he came across the fact that libgnomebt.so
symlink was not included and thus he was unable to build and link to libgnomebt.
Che did a little digging and discovered a small error in the spec that
left the symlink unpackaged. Appearently, to my great horror, rpm
doesn't notice when symlinks go unpackaged and doesn't warn about it
when building a package. Made me wonder if there were other missing
.so symlinks.
The idea:
So that got me to thinking, is there a not so clever way for me to get
a summary of potentially missing .so symlinks for a fully installed
fc3 system so we can get these reported in one big push. So i did
myself a full install, ran ldconfig and captured the output of
ldconfig -p. I then parsed the output of ldconfig -p using a drop
dead stupid shell script to get a list of candidate missing .so
symlinks file locations based on the existance of .so.* in the same
directory and compared that list to rpmdb-fedora to doublecheck those
specific file locations were not in an available core package.
The result:
I have a list of 49 POTENTIALLY missing .so files from the fc3 package
set. I stress potential because i know this was a very un-clever way
to approach this, and I'm sure there are several special case
situations that I'm misflagging with my script. For example, anything
like libick-X.Y.so.Z will flags libick-X.Y.so as missing even if
libick.so is there even though its not really a problem in most
circumstances.
But my goal was just to get the list down to something small enough to
be human reviewable in a reasonable amount of time. Che has already
gone over the list once, it was originally 78 or so potentially
missing files. I'm sure there are false alarms in the list, but there
are also some files that apear to both che and myself to be missing
and filable as packaging bugs. There might be situations where the .so
was delibrately left out that we are not aware of so I don't want to
jump the gun and file bug reports without more review. For example
libgnomebt.so and libgnome-window-settings.so jump out to me as real
package errors if they were not delibrately left out.
So please take a look at the list of potentially missing .so files
below and comment on specific files that you think are definitely
missing because of a packaging problem that impacts the ability to
build and link the dynamic library. This was from a complete fc3
install, I haven't had a chance to do this with a full rawhide install
yet.
-jef
ld-linux.so
libamu.so
libblkid.so
libboost_date_time.so
libboost_filesystem.so
libboost_prg_exec_monitor.so
libboost_python.so
libboost_regex.so
libboost_signals.so
libboost_test_exec_monitor.so
libboost_thread.so
libboost_unit_test_framework.so
libc.so
libdw.so
libFS.so
libg2c.so
libgcc_s.so
libgcj.so
libgfortranpreview.so
libGLw.so
libgmodule-1.2.so
libgnarl-3.4.so
libgnat-3.4.so
libgnomebt.so
libgnome-window-settings.so
lib-gnu-java-awt-peer-gtk.so
libgthread-1.2.so
libhpojip.so
liblber-2.2.so
liblber.so
libldap-2.2.so
libldap_r-2.2.so
libldap_r.so
libmDNSResponder-0.9.6.so
libNoVersion.so
libnss_db.so
libnuma.so
libobjc.so
lib-org-w3c-dom.so
lib-org-xml-sax.so
libparted-1.6.so
libptal.so
libpthread.so
libpwdb.so
libstdc++-libc6.1-1.so
libstdc++-libc6.2-2.so
libstdc++.so
libsvn_swig_perl-1.so
libsvn_swig_py-1.so
18 years, 9 months
Re: Suggestion for some network parameter defaults
by "Nils O. Selåsdal"
Ralf Ertzinger wrote:
> Hi.
>
> I have been using the following changes to some network parameters on all
> of my machines for a long time, and I was wondering whether they ought to
> be set by default.
>
> net.ipv4.conf.all.accept_redirects (current: 1, proposed: 0)
Any good reason for this ? I find it nice if a router tells me to
send stuff elsewhere rather.
> net.ipv4.icmp_echo_ignore_broadcasts (current: 0, proposed: 1)
And this ?
18 years, 10 months
Firefox fonts
by Avi Alkalay
Somebody knows how to make Gnome apps (firefox) to use the Gnome font
settings (and others like themes, etc) while on a KDE session?
Thanks,
Avi
18 years, 10 months
Where can I find kernel.2.6.9-1.681_FC3.src.rpm ?
by casimiro barreto
Where can I find kernel-2.6.9-1.681_FC3.src.rpm ?
[ ]s
Casimiro
--
The information contained in this message is confidential and
intended to the recipients specified in the headers. If you received
this message by error, notify the sender immediately. The
unauthorized use, disclosure, copy or alteration of this message
are strictly forbidden and subjected to civil and criminal sanctions.
18 years, 10 months
rawhide report: 20041129 changes
by Build System
Updated Packages:
lvm2-2.00.29-1.2
----------------
* Sun Nov 28 2004 Alasdair Kergon <agk(a)redhat.com> - 2.00.29-1.2
- Try excluding more archs.
* Sat Nov 27 2004 Alasdair Kergon <agk(a)redhat.com> - 2.00.29-1.1
- Exclude s390x which fails.
* Sat Nov 27 2004 Alasdair Kergon <agk(a)redhat.com> - 2.00.29-1
- Fix last fix.
memtest86+-1.40-1
-----------------
* Sun Nov 28 2004 Warren Togami <wtogami(a)redhat.com> - 1.40-1
- 1.40
- remove arch patch, now upstream
openoffice.org-1.1.2-14.7
-------------------------
* Wed Nov 24 2004 Dan Williams <dcbw(a)redhat.com> - 1.1.2-14
- #rh129719# Upstream indic translations need to be incorporated into OOo packages
(Fix transex3 to not screw up the merges)
- #rh136990# Backspace fails after IIIMF input
- Sort and prune duplicate entries from dictionary.lst file
* Tue Nov 23 2004 Dan Williams <dcbw(a)redhat.com> - 1.1.2-13
- #rh109628# Font 'Symbol' not displayed correctly (Caolan)
- #rh129719# Upstream indic translations need to be incorporated into OOo packages
(Actually merge the translations this time)
- Accomodate gcc 3.2.3 on gcc 3.4 systems (ie, RHEL4)
rpmdb-fedora-1:4-0.20041129
---------------------------
spamassassin-3.0.1-0.FC3
------------------------
* Sun Oct 31 2004 Warren Togami <wtogami(a)redhat.com> - 3.0.1-0.FC3
- 3.0.1
18 years, 10 months
Missing xorg header files
by Christian Fredrik Kalager Schaller
Hi,
I am trying to make a RPM for NVU (the mozilla based webeditor from
linspire), but the compilation fails with due to missing:
X11/extensions/Print.h
Googling I find other people mentioning the that the headers are missing
but no reply on why or when they can be expected to be back. Or if they
are gone for good and app developers need to stop using them.
Hope someone here can provide some insight.
Christian
18 years, 10 months
APT repository problems with up2date
by Jos Vos
Hi,
Before I start digging in the up2date code myself, maybe someone
can comment on this:
I have a problem when using up2date up2date with my self-created
APT repositories. When updating, it complains about conflicting
files: it just thinks that a long list of directories that are
shared among packages conflict with each other.
Apt-get itself (and synaptic) seems to work fine with the same
repository, only with up2date there is a problem. Also, up2date
with a yum repository of then same package set works fine, as
does yum itself.
FWIW: I generate my APT repositories with --flat --bloat.
Cheers,
--
-- Jos Vos <jos(a)xos.nl>
-- X/OS Experts in Open Systems BV | Phone: +31 20 6938364
-- Amsterdam, The Netherlands | Fax: +31 20 6948204
18 years, 10 months
Laptop mode and SCSI disks
by W. Michael Petullo
I am interested in using the kernel's laptop mode to reduce the power
usage of a server. The server has an external IEEE 1394 drive that could
easily be spun down for large periods of time. The IEEE 1394 drive is,
of course, treated as a SCSI disk by the kernel.
I have found that spinning down SCSI disks is not as easy as spinning
down IDE disks. I have found utilities to do this, but they require
kernel patches and these patches do not seem to exist for the 2.6 kernel's
SCSI sub-system. Scsi-spin, scsi-idle and noflushd are some of the
options I've looked at.
Is anyone working on automatically spinning down SCSI (including USB
and IEEE 1394) disks? Would it be worthwhile for me to try and port
the 2.4 "spindown ioctl" patch to 2.6? Or does similar functionality
already exist in 2.6?
I also have some questions about laptop mode specifically: Is there
something "magical" that causes disks to spin down after a certain
inactivity period when laptop mode is enabled? Or does "hdparm -S X"
still need to be used to set this?
"hdparm -S 1 /dev/sda" currently says:
/dev/sda:
setting standby to 1 (5 seconds)
HDIO_DRIVE_CMD(setidle1) failed: Invalid argument
would hdparm also need to be patched to support SCSI disks?
--
Mike
:wq
18 years, 10 months
Testing
by M Harelick
Hi:
Can someone direct me to the right contact to discuss volunteering for
QA on Linux components?
Thank You
Matthew Harelick
18 years, 10 months
