F23 install from hard disk
by Amadeus W.M.
I haven't upgraded my machines in a while and I couldn't find any
reference on how to install directly from the hard disk, without media.
Is this still possible? None of my pcs boot off of a usb stick and the
cdrom I think is failing on one of them, so I really need to install from
the disk. Any suggestions? Thanks!
8 years
any simpler option to overlaying header file dirs than funionfs?
by Robert P. J. Day
a project i was recently introduced to uses, quite heavily, the
notion of overlaying header file directories with the now-obsolete gcc
"-I-" option, and i'm curious if there are equivalent solutions that
don't use that feature.
summary: sizable C (and possibly, not sure, some C++) code base
contains numerous, quote-surrounded header file includes of the form:
h1.h;
#include "h2.h"
... etc etc ...
where the header file "h2.h" is frequently in the same directory as
"h1.h" *but* it may not be the one that should be included.
instead, the build system is currently replete with gcc build
commands of the form:
$ gcc -I- -Id1/ -Id2/ -Id3/ ... and so on and so on ...
and "-I-" is used to force the preprocessor to resolve references to
"headerfile.h" not by looking in the current directory first, but by
scanning a list of "overlay" directories to see if someone has
registered a "higher-priority" header file by that name. it's not the
most intuitive build structure i've ever seen but it's been around for
quite some time and it works and its users really don't want to have
to mess with it.
[SIDE NOTE: the current usage of "-I-" in this project is not so much
to "split" the header file search path as that option is used right at
the *beginning* of all the "-I" options, so its use is to simply
refuse to look locally for header files and immediately start scanning
the "-I" directories. but that doesn't really change the problem to be
solved.]
predictably, now that gcc 5 has obsoleted the "-I-" option, one
would like to find an alternative that doesn't require a wholesale
reworking of the build system. since the build system runs on centos,
my current thought is to use "funionfs" to unionize those overlay
directories in the same priority sequence.
i just tried a trivial funionfs example and it seems to work, but is
there a simpler solution? the gcc folks, in deprecating, then
obsoleting, the "-I-" option, insist that one can use "-iquote"
instead, but that is clearly incorrect -- the requirement here is that
resolving a preprocessor include of the form:
#include "rday.h"
*must* *not* look in the current directory first. if there is some way
to get that effect without "-I-" and without some sort of unionfs, i'd
like to know what it is.
thank you kindly.
rday
--
========================================================================
Robert P. J. Day Ottawa, Ontario, CANADA
http://crashcourse.ca
Twitter: http://twitter.com/rpjday
LinkedIn: http://ca.linkedin.com/in/rpjday
========================================================================
8 years
cpio Digest mismatch error?
by Tom Horsley
Anyone have any idea why only fedora systems would give this
error installing an rpm?
Installing : ccur-ntrace-ai-7.5-0.beta.75.x86_64 19/32
Error unpacking rpm package ccur-ntrace-ai-7.5-0.beta.75.x86_64
error: unpacking of archive failed on file
/usr/lib/NightTrace/illuminators/ccur_rt/ccur_rt.h;56fd10d8: cpio: Digest
mismatch
error: ccur-ntrace-ai-7.5-0.beta.75.x86_64: install failed
The exact same rpm installs fine on a centos 7.2 system.
Even on fedora, if I use rpm2cpio and extract the files, there is
absolutely nothing wrong with the ccur_rt.h header file.
8 years
Getting rtl8192eu driver.
by Michael D. Setzer II
I just got 4 usb 802.11n nics, and they were not recognized by fedora 23??
Windows 10 did seem to have no problems, and they are rtl8192eu.
Did find a page that talked about getting them to work, but the install script
failed with errors. First Error was about __DATE__ and __TIME__ macros,
so found an option to add to the Makefile to ignore that.
The second error involved strnicmp issues, and eventually found a reference
to replace that with strncasecmp. With that change the script does seem to
complete. First time it didn't seem to work, but then the wlassistant was able
to see the device and was able to connect to a wireless hub.
Wondering if someone knows of a better solution?
Thanks.
+----------------------------------------------------------+
Michael D. Setzer II - Computer Science Instructor
Guam Community College Computer Center
mailto:mikes@kuentos.guam.net
mailto:msetzerii@gmail.com
Guam - Where America's Day Begins
G4L Disk Imaging Project maintainer
http://sourceforge.net/projects/g4l/
+----------------------------------------------------------+
http://setiathome.berkeley.edu (Original)
Number of Seti Units Returned: 19,471
Processing time: 32 years, 290 days, 12 hours, 58 minutes
(Total Hours: 287,489)
BOINC@HOME CREDITS
ABC 16613838.513356 | EINSTEIN 93508184.955695
ROSETTA 42378641.476131 | SETI 78948334.109332
8 years
NFS from a VM
by John J. McDonough
I have a small server that offers a number of directories over NFS. I
am running a virtual machine on another physical machine.
When I try to mount an NFS directory on the VM, the NFS server refuses,
claiming a bad port. Successive tries result in different reported
port numbers.
What is going on? Is there some sort of virtio setting I am missing?
NFS works fine to the machine hosting the VM, but not to another
machine.
--McD
8 years
Re: [talk-au] Distinguishing between low-friction and high-friction shared paths
by Simon Slater
On Mon, 4 Apr 2016 04:28:45 PM Sam Russell wrote:
> NSW's specifications for bicycle infrastructure are… interesting.
Slightly off-topic ... a Dutch planner comparing cycleways in Holland and
Australia said that in Holland the cyclists are protected by barriers like the
parked cars or even a gutter in the least. Yet in Australia we protect the
parked cars by the cyclists.
--
Regards
Simon Slater
Registered Linux User #463789
http://linuxcounter.net
8 years
Is the flash plugin no longer getting security updates?
by Ranbir
Hi Everyone,
Everywhere I go that uses useless flash, I get an error saying my flash
version is out of date. When I look for an update from Adobe (i.e. back
ported security fixes), there isn't one. I'm already running the latest
release for Linux.
Did Adobe drop the security updates to flash sooner than they promised?
--
Ranbir
8 years
Re: devtmpfs on /dev type devtmpfs
by Paul Schroeder
On Mon, Apr 4, 2016 at 10:02 AM, Paul Schroeder <paul.schroeder(a)nimbix.net>
wrote:
>
>
> On Fri, Apr 1, 2016 at 8:21 PM, Joe Zeff <joe(a)zeff.us> wrote:
>
>> On 04/01/2016 08:05 AM, Paul Schroeder wrote:
>>
>>> I tried both. Neither solution seemed to keep the mount from happening.
>>>
>>
>> Have you considered using umount to remove the unwanted mount in
>> /etc/rc.d/rc.local? It's not exactly elegant, but it should do the trick.
>>
>> Yea. That won't really work:
>
> # umount /dev
> umount: /dev: target is busy.
> (In some cases useful info about processes that use
> the device is found by lsof(8) or fuser(1))
>
>
>
It would appear that the LXC config option for the container,
lxc.automount, determines whether or not LXC decides to mount devtmpfs on
/dev. When lxc.automount = 1, it is obviously on and always does this.
After digging through the LXC code, if the option is not specified, LXC
autodetects whether or not systemd is used inside the container. If it is
detected, LXC automatically turns automount on. (I wish this had been
documented somewhere.) So to stop LXC from mounting /dev, one has to
explicitly set lxc.automount = 0..
However, this causes problems inside the container, causing it to freeze on
boot. It would appear that systemd within my CentOS container is still
trying to mount /dev when it can't. I believe that I have all of the udev
services masked properly, so this must be coming from somewhere else within
systemd. Any further ideas out there?
$ sudo lxc-start -n c7
Failed to mount devtmpfs at /dev: Permission denied
Failed to mount cgroup at /sys/fs/cgroup/systemd: Permission denied
[!!!!!!] Failed to mount API filesystems, freezing.
Thanks! Paul...
8 years
OT: Suspend/restore and USB devices
by Patrick O'Callaghan
This is possibly a hardware issue rather than a Fedora problem, but I
thought I'd bring it up anyway:
I have a Bluetooth mouse connected via a BT dongle. The dongle normally
sits on the back of my cpu case on the floor beneath my desk, but I've
noticed this often produces lag in mouse movement, presumably caused by
interference from the desk's metal frame (BTW it took me months to
figure this out). I bought a USB extension cable (cheap Amazon Basics
model) and now have the dongle sitting above the desk and the lag
problems have gone.
So far so good. However now when I suspend the system and try to
restart it, nothing will happen until I physically remove the dongle
from the extension cable. The screen will then come back to life and I
can plug the dongle in again to get the mouse working. This never
happened when the dongle was directly connected to the cpu.
Do I need to get a better cable? If not, what can I do?
poc
8 years