[SLUG] Filesystem encryption
Hi all, I'm considering using device mapper's crypto support to encrypt the entirety of my laptop's drive. This is a fairly permanent thing to do, so I'm seeking some experiences with it to help me decide if it's a good idea or not. I used it a few years back and found that it didn't play nice with XFS, causing frequent lockups, which wasn't really what I was looking for. It also burned a lot of cycles, making stuff like my frequent grepping through source trees and image processing impractical. Now, the target machine is much faster (a Thinkpad x61, C2D), but I don't really know how C2D crypto performance compares to Pentium M, so it'd be good to hear about that too. Finally, if I do go ahead with it, what's the easiest way to do it? I recall Ubuntu having an alternative installer that could do it for me. Is that the best way to go? Any and all insights appreciated. Please CC me, as I'm not subscribed to the list. Thanks, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] xterm fonts
Sorry I didn't get to this sooner; just getting back to email after the Christmas break. I've found that a good solution to this problem is to use colours to indicate bold and underlined, as actual "bolding" just makes the text less clear. I'm unfortunately not exactly sure which of the settings in my .Xdefaults is responsible for doing this, so the bits that *look* relevant follow: XTerm*highlightSelection: true XTerm*VT100*colorBDMode: on XTerm*VT100*colorBD: DodgerBlue1 XTerm.VT100*colorULMode: on XTerm.VT100*underLine: off XTerm*VT100*colorUL: red1 XTerm.VT100*dynamicColors: on The following colour settings work quite well with those options, but it's purely a question of taste: *VT100*color0: black *VT100*color1: red1 *VT100*color2: green3 *VT100*color3: yellow3 *VT100*color4: DodgerBlue1 *VT100*color5: magenta3 *VT100*color6: cyan3 *VT100*color7: gray90 *VT100*color8: gray30 *VT100*color9: red *VT100*color10: green *VT100*color11: yellow *VT100*color12: blue *VT100*color13: magenta *VT100*color14: cyan *VT100*color15: white *VT100*colorUL: red1 *VT100*colorBD: DodgerBlue1 So anyway you dump all that stuff in your .Xdefaults file in your home directory and run xrdb .Xdefaults to import the settings, if your login scripts don't do so automatically (I'm not sure what current distros do on that front). HTH, James. On Sat, 2007-12-08 at 14:40 +1100, Richard Heycock wrote: > Does anyone know how to set up a proper bold font in xterm? I've got a > low res screen so I've got everything as small as possible but bold > fonts in xterm are pretty much unreadable. > > I can get a decent bold font if I use eterm but I don't know how to find > out which font eterm uses. > > rgh > > -- > "Wickedness is a myth invented by good people to account for the > curious attractiveness of others" -- Oscar Wilde > > Richard Heycock > Blog: http://blog.twosecondmemory.org/ -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Ubuntu Gutsy MD Houdini act
On Thu, 2007-10-25 at 16:26 +0930, Glen Turner wrote: > On Thu, 2007-10-25 at 14:22 +1000, James Gregory wrote: > > > > Has the md raid system been deprecated in Gutsy? Is it likely that I > > need to do something to migrate this to device-mapper or something? > > Hi James, > > It looks like they've reworked the assembly of RAID during boot for > Gutsy. See I finally got time to work this one out. Sorta. Thought I'd post for the sake of the archives. There were two problems: - There was an extraneous kopt_alt_26 (I think) line in /boot/grub/menu.lst, which means that instead of having 'root=UUID=xyz', it had root=/dev/md5. I removed this line, leaving just the kopt= line, with the correct root= argument and reran update-grub. - My /etc/mdadm/mdadm.conf file was subtly wrong: the last two colon-separated fields on the ARRAY lines for md4 and md5 were wrong. The first two were correct. I have no idea how it got into this state. Anyway, I ran 'mdadm -QE --scan', and replaced the relevant lines in /etc/mdadm/mdadm.conf with that data. Then I ran mkinitramfs to push that change back into the initrd. Et voila! It boots again. I can only assume that the reason the old kernel still worked was because it somehow had a correct mdadm.conf in its initrd, though I'm utterly puzzled at how it ended up in this state Thanks to everyone for their help on this. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] null-modem cable
On Tue, 2007-11-27 at 00:17 +1100, Nick Croft wrote: > Or any pointers to a sydney retailer would be equally appreciated. I know of two: http://www.elx.com.au/item/NM9SS2 and https://www.lindy.com.au/ (you'll need to dig through their products list). We've used the latter one to order a bunch of them at work, and they've been quite good. HTH, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] Sponsor me for Movember! http://www.movember.com/au/donate/ Registration number: 117052 -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] batch rename
On Sat, 2007-11-03 at 16:30 +0900, Kevin Shackleton wrote: > suggestions to rename multiple files in a single directory to individual > shorter names? Hey Kevin, The tool you want is 'rename'. It lets you rename a set of files using a PCRE to determine the new file name. For example, to do what you've suggested you could use something like this: rename 's/^proj1//' * The manual page has some other useful examples. HTH, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] Sponsor me for Movember! http://www.movember.com/au/donate/ Registration number: 117052 -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Ubuntu Gutsy MD Houdini act
Hi Glen, On Thu, 2007-10-25 at 01:13 +0930, Glen Turner wrote: > On Wed, 2007-10-24 at 01:18 +1000, James Gregory wrote: > > I upgraded one of my computers to Ubuntu Gutsy with aptitude recently > > (the update-manager -d method refused to work), and am now unable to > > boot my computer into the new kernel (thankfully one of the old ones > > still works). > > I found "do-release-upgrade" to be the next best approach if > update-manager > fails. Thanks. I didn't know about that; I'll give it a go. > If you are having trouble booting a new kernel with a slightly odd I/O > system > then it is likely to be the initramfs lacking a needed module -- I had > this > problem on my PowerBook Mac which tossed me into a busybox shell (rescue > mode?). I then insmod-ed the ide_core module, continued to boot, and > generated a correct initrd. Yep, I agree that the problem is the initramfs, but it *can* find all the /dev/hd* entries, it has even created /dev/md* special files, what it hasn't done is gone ahead and assembled the actual md devices that those special files are meant to refer to (ie, running cat /proc/mdstat shows no arrays). I'll check, but that sounds to me like the raid personality drivers are loaded; there's just a step missing where it assembles the devices... Unless mdadm loads the raid personalities when it does its thing; I'll check that tonight. Incidentally, this is software raid, not hardware raid. Thankyou very much for your reply, but I don't think this is it; the md devices can be created manually, no insmodding required. Has the md raid system been deprecated in Gutsy? Is it likely that I need to do something to migrate this to device-mapper or something? I did think of another way to solve this problem, which I'll go to as a last resort if I have to: do a new installation of Gutsy onto one of the drives, copy the data across and then add the other drive as its mirror. Something I'd like to avoid if possible though. Thanks for your help, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Ubuntu Gutsy MD Houdini act
On Wed, 2007-10-24 at 01:18 +1000, James Gregory wrote: > Hi guys, > > I upgraded one of my computers to Ubuntu Gutsy with aptitude recently > (the update-manager -d method refused to work), and am now unable to > boot my computer into the new kernel (thankfully one of the old ones > still works). Hmm. Alternatively, is there a way I can apply the changes that update-manager -d would have done if I could have got it to work? Ta, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Ubuntu Gutsy MD Houdini act
Hi guys, I upgraded one of my computers to Ubuntu Gutsy with aptitude recently (the update-manager -d method refused to work), and am now unable to boot my computer into the new kernel (thankfully one of the old ones still works). The problem is that my raid devices aren't coming up (the whole system, root included, is on RAID1). I can confirm this by waiting for the "waiting for root" timeout during boot, and inspecting the system from the shell that comes up. The device files exist (/dev/md5 is the important one), but the array has not been assembled. I can manually assemble it at this point, and then mount it successfully. Haven't tried booting from that stage yet. So, any ideas on how I can get RAID working on this machine? It was setup by the ubuntu alternative installer from 7.04 (on Jeff's sage advice). I've looked around and the two common remedies (removing evms and adding arbitrary sleep times to initramfs) don't help (evms wasn't ever there, and the /dev/md5 *file* exists; it just hasn't been assembled. Any help appreciated; as this is the second machine I've broken by installing Gutsy, I'm beginning to believe that I'm cursed! And the next machine to get this upgrade is the important one. That will be gutsy indeed. Thanks. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Suspenseful laptops
On Fri, 2007-09-14 at 20:04 +1000, Lindsay Holmwood wrote: > On 9/13/07, James Gregory <[EMAIL PROTECTED]> wrote: > > Can anyone recommend a laptop that is small (11.1 - 13.3 inches), > > reasonably fast, and can do suspend (both to RAM and disk) easily? Oh, > > and that I can buy now; old models aren't so interesting to me. Ubuntu > > is my OS of preference, but if it makes a difference, I'd consider > > switching to something else. > > > > I've been eyeing off similar configurations the last few months. My > recommendations: > > IBM x61(s) > HP Presario B1973TU > HP NC4400, TC4400 > Toshiba Portege R500 > Sony Vaio TZ90 (Japan import only) I saw the TZ90 when I was in Japan, and very nearly impulse bought it. I'm fairly sure that the TZ90 is this computer: http://www.sony.com.au/vaio/category.jsp?id=34641 with a Japanese keyboard. I've certainly seen the same 1.2GHz, solid state drive configuration in .au. The American page for it has a bit more detail: http://tinyurl.com/2fjerp I currently have a HP nc4010, which is ok. The battery life is a bit disappointing, but I was dealing ok with that, until I bought a new camera. Now I'm finding that this machine is struggling just to generate thumbnails of my photos. It's this processor speed concern that makes me wary of the TZ; jpeg processing (AFAIK) is not parallelisable, so I'd be stuck with the speed of a single core, and whilst the instruction latency is substantially lower on the Core architecture, I think I'd probably only just break even. Larger L2 cache might help, but the memory access patterns for JPEG are well understood, so presumably any modern implementation would use prefetch instruction to mitigate that. Does anyone know if this is the case in the standard OSS jpeg implementations? > I have no idea about how well Linux works but I assume the IBM x61 and > HP TC and NCs are well supported. Well, the Sony seems to be full of Intel bits, so it would probably be ok. I'm just a little wary of it because we have a Vaio SZ in the office, which has very odd bits of proprietary stuff around the Intel goodness (like a switch to toggle between two video cards; toggling this switch does not do good things to the Linux). I hadn't thought of the Thinkpad. That might be a good option. And it's got a "real" Core chip at 2GHz. Does anyone know if it does suspend/resume ok? Thanks for your help here Lindsay. Those are some good suggestions. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Suspenseful laptops
On Thu, 2007-09-13 at 12:30 +1000, Jeff Waugh wrote: > > > > Can anyone recommend a laptop that is small (11.1 - 13.3 inches), > > reasonably fast, and can do suspend (both to RAM and disk) easily? Oh, and > > that I can buy now; old models aren't so interesting to me. Ubuntu is my > > OS of preference, but if it makes a difference, I'd consider switching to > > something else. > > I'm very happy with the Dell D420, which has been updated to Core 2 Duo with > the newer Dell D430. Robert Collins has one of these, so he can tell you how > sweet it is. :-) Mine works really nicely, with Intel top-to-bottom. Neat. Thanks dude. Looks a little slow though -- 1.2 GHz. Do you find it so? > The Dell M1330 is also pretty intriguing -- backlight and SSD storage! It is > 13", slightly bigger than the D430. Yeah, that does look interesting, but it seems a bit weird having a wedge-shaped laptop. Still a contender though. This little one from Sony: http://tinyurl.com/26fu4d would be a good option if I decide I can deal with the slow-clocked chip. The battery life claims are pretty amazing. Has anyone used one of those with Ubuntu? James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Suspenseful laptops
Can anyone recommend a laptop that is small (11.1 - 13.3 inches), reasonably fast, and can do suspend (both to RAM and disk) easily? Oh, and that I can buy now; old models aren't so interesting to me. Ubuntu is my OS of preference, but if it makes a difference, I'd consider switching to something else. Ta, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Audio skipping on Ubuntu kernels
Just a quick follow-up on this.
I discovered last night that the reason that switching to 'performance'
made the problem go away is *not* that more CPU power was available, it
was that the CPU frequency no longer changed. My correlation of 'more
CPU usage == music skipping' was in fact basing my diagnosis on a
symptom. In fact, it's more like "more CPU usage == clock speed changes
== music skipping". It occasionally skips still, but it's nowhere near
as bad as it used to be.
So, for now, I've set it to use 'powersave', which is a hack, but a
manageable one.
Other discoveries last night:
- If you disable the fans, the BIOS will turn them on if it really needs
to (but don't blame me if *your* laptop explodes). For some reason,
Ubuntu sets up my computer to always spin the fans slowly rather than
turn them off, so this is preferable for me (it's quieter, and it has
longer battery life).
- Despite DMA being turned on on the hard-disk, there were heaps of
other things that were not. Notably, whilst DMA transfers were enabled,
the drive was programmed to use PIO mode (I actually didn't know that
went as deep as the driver). My laptop is a lot happier with these lines
in my /etc/hdparm.conf file:
/dev/hda {
mult_sect_io = 16
write_cache = off
dma = on
read_ahead_sect = 3072
lookahead = on
io32_support = 1
spindown_time = 6
interrupt_unmask = on
write_cache = on
transfer_mode = 34
}
(again, be careful setting those. The wrong value for transfer_mode will
freeze your computer; it was however, the one that made DMA transfers
worthwhile). The read_ahead value is something I'm experimenting with.
It is my hope that using OS read-ahead will mean that the disk spins
down a bit more.
- Adding 'noatime' to the options in fstab is a 100% awesome thing to do
on a laptop.
- Even though with the new drivers my GPU is quite capable of doing
'Desktop Effects', it makes the GPU chip seriously hot on this machine.
If you're on an older laptop (mine's about 2 years old), this might be a
good thing to do.
HTH someone.
James.
On Thu, 2007-08-02 at 11:04 +1000, James Gregory wrote:
> On Thu, 2007-08-02 at 09:44 +1000, Scott Ragen wrote:
> > James Gregory <[EMAIL PROTECTED]> wrote on 01/08/2007 04:56:26 PM:
> >
> > > So that does substantially help matters -- I have to try pretty hard to
> > > make it skip in that situation. It unfortunately also chews through
> > > battery life and makes the fan scream like some kind of gently blowing
> > > banshee. Needing 1.7GHz of processing power to download email and play
> > > music seems a bit overkill.
> > >
> > > But ok, it may be *switching* performance levels that is the problem
> > > (since that will occur when my mail client wakes up and does stuff). If
> > > that is the case, what kind of things could I do? I've previously tried
> > > re-nicing rhythmbox and esd to -19 and it seemed to have no measurable
> > > effect.
> > >
> > Which driver are you using for the cpufreq?
> > cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver
>
> It's set to 'centrino' atm.
>
> > If your using the generic acpi, try the driver specific for your
> > cpu/chipset. If not, try using acpi-cpufreq.
>
> Is this more complicated than echoing the appropriate string into that
> file? I get the following:
>
> # echo "acpi-cpufreq" > /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver
> bash: /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver: Permission denied
>
> And nothing in 'dmesg' telling me what happened.
>
> I did make a little bit of progress on this yesterday. I found that if I
> use the 'conservative' frequency scaler, and renice my various courier
> processes (I use courier-imap for mail) to 19, it's substantially
> better. Still far from flawless, but only a stone's throw from
> tolerable. I might try configuring Evolution to talk directly to the
> Maildir and see what happens.
>
> Thanks for the pointers.
>
> James.
>
> --
> James Gregory -- http://codelore.com -- [EMAIL PROTECTED]
>
>
--
James Gregory -- http://codelore.com -- [EMAIL PROTECTED]
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Which Distro?
On Wed, 2007-08-29 at 17:08 +1000, Jeff Waugh wrote: > > > > Quick qn: I need a Linux distribution I can install and get RAID-1 > > (software, PATA drives) out of the box. What are my options? IIRC, Ubuntu > > was not able to do this the last time I tried. Is that still the case? > > No, Ubuntu has been able to do this for a while. Not entirely sure if the > GUI installer does it, but the text installer (which is about as easy, and > kicks off way faster) definitely does. If you're installing a desktop, get > the 'alternative' CD to use the text installer. If you're installing a > server, just grab the 'server' CD. Awesome! That's exactly what I want. I did look cursorily at the Ubuntu website, but didn't see it mentioned in the feature list; probably just missed it. Thanks for your help! James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Which Distro?
Hey guys, Quick qn: I need a Linux distribution I can install and get RAID-1 (software, PATA drives) out of the box. What are my options? IIRC, Ubuntu was not able to do this the last time I tried. Is that still the case? Pretty flexible on most things -- just want it to support my not-too-exotic hardware and make RAID easy. Thanks! James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Audio skipping on Ubuntu kernels
On Thu, 2007-08-02 at 09:44 +1000, Scott Ragen wrote: > James Gregory <[EMAIL PROTECTED]> wrote on 01/08/2007 04:56:26 PM: > > > So that does substantially help matters -- I have to try pretty hard to > > make it skip in that situation. It unfortunately also chews through > > battery life and makes the fan scream like some kind of gently blowing > > banshee. Needing 1.7GHz of processing power to download email and play > > music seems a bit overkill. > > > > But ok, it may be *switching* performance levels that is the problem > > (since that will occur when my mail client wakes up and does stuff). If > > that is the case, what kind of things could I do? I've previously tried > > re-nicing rhythmbox and esd to -19 and it seemed to have no measurable > > effect. > > > Which driver are you using for the cpufreq? > cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver It's set to 'centrino' atm. > If your using the generic acpi, try the driver specific for your > cpu/chipset. If not, try using acpi-cpufreq. Is this more complicated than echoing the appropriate string into that file? I get the following: # echo "acpi-cpufreq" > /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver bash: /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver: Permission denied And nothing in 'dmesg' telling me what happened. I did make a little bit of progress on this yesterday. I found that if I use the 'conservative' frequency scaler, and renice my various courier processes (I use courier-imap for mail) to 19, it's substantially better. Still far from flawless, but only a stone's throw from tolerable. I might try configuring Evolution to talk directly to the Maildir and see what happens. Thanks for the pointers. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Audio skipping on Ubuntu kernels
On Wed, 2007-08-01 at 15:56 +1000, Scott Ragen wrote: > [EMAIL PROTECTED] wrote on 01/08/2007 03:05:49 PM: > > > On Wed, 2007-08-01 at 13:31 +1000, Robert Collins wrote: > > > On Wed, 2007-08-01 at 13:27 +1000, James Gregory wrote: > > > > Hey guys, > > > > > > > > I've found that playback of music on my laptop has this horrible, > > > > horrible tendency to "skip" (by which I mean that the music stops > for > > > > some fraction of a second) extremely frequently. This has only been > > > > occurring since Ubuntu's -generic kernels came in (I forget which > > > > version that was). > > > > > > I'd check your disk is using dma; PIO disk IO is a great way to turn > > > your laptop into a snail. > > > This might be way off, but I had intermittant freezing problems with the > kernel using cpufreq ondemand governor. > To see what govenor is used, try > # cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor > > Try setting it to performance if its anything but. So that does substantially help matters -- I have to try pretty hard to make it skip in that situation. It unfortunately also chews through battery life and makes the fan scream like some kind of gently blowing banshee. Needing 1.7GHz of processing power to download email and play music seems a bit overkill. But ok, it may be *switching* performance levels that is the problem (since that will occur when my mail client wakes up and does stuff). If that is the case, what kind of things could I do? I've previously tried re-nicing rhythmbox and esd to -19 and it seemed to have no measurable effect. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Audio skipping on Ubuntu kernels
On Wed, 2007-08-01 at 16:31 +1000, Heracles wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > I have only 1GB of RAM and a 2GB swap. My scaling governor is set to > ondemand and I don't have a problem with audio (playing oggs on vlc) > skipping with disk IO and ftp happening simultaneously. > My kernel is 2.6.20-16-generic on Ubuntu 7.04. > Do the files work properly when they are the only thing running? Yep, if Rhythmbox is the only thing running, it's ok. That's unfortunately a slightly prohibitive solution. Also keep in mind that this used to work, along with all the other stuff I normally get my laptop to do. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[Fwd: Re: [SLUG] Audio skipping on Ubuntu kernels]
Forgot to CC the list on this one. Forwarded Message From: James Gregory <[EMAIL PROTECTED]> To: Denis Crowdy <[EMAIL PROTECTED]> Subject: Re: [SLUG] Audio skipping on Ubuntu kernels Date: Wed, 01 Aug 2007 16:42:23 +1000 On Wed, 2007-08-01 at 16:15 +1000, Denis Crowdy wrote: > James Gregory wrote: > > Hey guys, > > > > I've found that playback of music on my laptop has this horrible, > > horrible tendency to "skip" (by which I mean that the music stops for > > some fraction of a second) extremely frequently. This has only been > > occurring since Ubuntu's -generic kernels came in (I forget which > > version that was). > Might be helpful: > http://ubuntuforums.org/showthread.php?t=136179 - particularly the last > post with the suggested options to the kernel as: > noapic nolapic pci=noacpi acpi=off quiet splash > > Although some of those would mess with power settings on a laptop > wouldn't they? Yeah, I'd prefer not to disable ACPI if I can get away with it. It's also worth noting that this did work, on previous kernels, with ACPI. The other thing about this is that it's solving the problem of the sound-card not working at all, as opposed to that of it working, but stalling. > Can you get jack working? Not very useful for general everyday Desktop > sound granted but might indicate whether a driver or ALSA was at fault? Hm. Similarly, disabling esd would tell me that. I might try disabling esd some time later. > Also, there is a low latency kernel that I use on this machine > (although I can't say I've noticed better audio results). Might be > worth a try. Yeah, I'm currently using it, on the hope that it might help, but it doesn't. Thanks for your suggestions. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Audio skipping on Ubuntu kernels
On Wed, 2007-08-01 at 15:29 +1000, Amos Shapira wrote: > On 01/08/07, James Gregory <[EMAIL PROTECTED]> wrote: > > > > Good thinking, but DMA was enabled. I checked the other settings, and > > twiddled them, but it's still skipping. Current config is as follows: > > > You keep assuming that it's an IO issue - what about creating a short file > which can fit in memory and try to play it repeatedly, just to see that the > system can keep up with the playing without the disk in the equation? Well, actually I kinda assumed it was a scheduler issue, but fair point nonetheless -- measure then optimise. However, all the 4-5MB files I play can well and truly fit into the 1GB of RAM in the system and I have no swap partition. I did the experiment however, and playing the same file repeatedly still suffers the same problem. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Audio skipping on Ubuntu kernels
On Wed, 2007-08-01 at 13:31 +1000, Robert Collins wrote: > On Wed, 2007-08-01 at 13:27 +1000, James Gregory wrote: > > Hey guys, > > > > I've found that playback of music on my laptop has this horrible, > > horrible tendency to "skip" (by which I mean that the music stops for > > some fraction of a second) extremely frequently. This has only been > > occurring since Ubuntu's -generic kernels came in (I forget which > > version that was). > > I'd check your disk is using dma; PIO disk IO is a great way to turn > your laptop into a snail. Hey Rob. Good thinking, but DMA was enabled. I checked the other settings, and twiddled them, but it's still skipping. Current config is as follows: # sudo hdparm /dev/hda /dev/hda: multcount= 8 (on) IO_support = 1 (32-bit) unmaskirq= 1 (on) using_dma= 1 (on) keepsettings = 0 (off) readonly = 0 (off) readahead= 256 (on) geometry = 19457/255/63, sectors = 312581808, start = 0 Thanks, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Audio skipping on Ubuntu kernels
Hey guys,
I've found that playback of music on my laptop has this horrible,
horrible tendency to "skip" (by which I mean that the music stops for
some fraction of a second) extremely frequently. This has only been
occurring since Ubuntu's -generic kernels came in (I forget which
version that was).
It seems that it occurs whenever almost *anything* wants the CPU or disk
-- for example, my fetchmail and procmail setup interrupts music
playback whenever I receive mail.
Is there a quick fix I can apply to improve matters? My guess is that
userspace is being switched out, so having humungous kernel-side buffers
for my audio driver is something I'd consider here.
It's a HP nc4010, and lspci claims that the audio card is an "ALi
Corporation M5451 PCI AC-Link Controller Audio Device (rev 02)". Ubuntu
Feisty. lsmod reports the following 'snd' modules loaded:
# lsmod | grep sn | awk '{print $1}' | sort | uniq
ac97_bus
snd
snd_ac97_codec
snd_ali5451
snd_mixer_oss
snd_page_alloc
snd_pcm
snd_pcm_oss
snd_rawmidi
snd_seq
snd_seq_device
snd_seq_dummy
snd_seq_midi
snd_seq_midi_event
snd_seq_oss
snd_timer
soundcore
Thanks dudes,
James.
--
James Gregory -- http://codelore.com -- [EMAIL PROTECTED]
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] WAN link optimisation
On Wed, 2007-07-25 at 10:54 +1000, James Gregory wrote:
> On Tue, 2007-07-24 at 13:37 +1000, Gavin Carr wrote:
> > Hey sluggers,
> >
> > Anyone have any pointers to open source projects (or features of projects)
> > around WAN link optimisation? I'm specifically looking for a way of
> > duplicating traffic across multiple links to avoid resends on high latency
> > links, but I'm interested in the whole area.
> Don't have references handy, but I expect that you would be able to
> improve performance substantially by tinkering some values in proc.
So following the interest this thread has received, I dug up some
references. Some time ago, Evgeniy Polyakov posted a link to some neat
graphs detailing the effects of different retransmit ("resend" as you
call it) algorithms. These graphs:
http://developer.osdl.org/shemminger/tcp/net-2.6.22/
Are the most illuminating, and some of Stephen Hemminger's presentation
on the subject detail other network performance stuff. Take a look
around:
http://developer.osdl.org/shemminger/
Now, I realise that this does not directly address your interest in
duplicating TCP traffic to mitigate the effect of per-link packet-loss,
but it does demonstrate that the cost of that packet-loss can be
substantially reduced by small changes to the behaviour of TCP. Huge
windows will get you the biggest benefit on high-latency links, as, if
they are huge enough, you can keep the pipeline full enough to hide the
latency of receiving the duplicate ACKs you require to trigger a
retransmit.
I'm sure there's a bunch of commercial stuff you can use, like that
'Fast TCP' thing (which actually sounds pretty cool), but if you want
open-source, I suspect the biggest win you'll see is in just shuffling a
few parameters in your kernel's network stack.
HTH,
James.
--
James Gregory -- http://codelore.com -- [EMAIL PROTECTED]
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] WAN link optimisation
On Tue, 2007-07-24 at 13:37 +1000, Gavin Carr wrote: > Hey sluggers, > > Anyone have any pointers to open source projects (or features of projects) > around WAN link optimisation? I'm specifically looking for a way of > duplicating traffic across multiple links to avoid resends on high latency > links, but I'm interested in the whole area. Hey Gavin, Don't have references handy, but I expect that you would be able to improve performance substantially by tinkering some values in proc. Specifically I'm thinking of selecting a different retransmit algorithm for TCP, and enabling the various 'smart ACK' schemes available these days. All the usual stuff like increasing send and receive buffers and enabling window scaling applies as well. There's some good stuff about tuning linux for trans-continental links out there, which I've previously found with google. HTH, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Etax 2007
On Tue, 2007-07-10 at 23:20 +1000, James Gregory wrote: > On Mon, 2007-07-02 at 21:39 +1000, Dean Hamstead wrote: > > ran perfectly in wine (whatever is current). > > Hey Dean, > > That's awesome. Can you tell us how you did it? I just tried the current > Ubuntu packages from winehq.org, and got this message: In case anyone was wondering, the updates that my computer pulled down this morning from winehq.org appear to have fixed the problem. I've not gone all the way through the process yet, but this problem no longer exists. James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] de-HTMLzing emails?
On Mon, 2007-07-23 at 09:50 +1000, Voytek Eymont wrote: > I want to receive emails on my h/held, but as GPRS rates are horrendously > expensive ($15 to send 240k email that took about 2 minutes) I'd like to > drop attachements and convert any emails to text/plain, what is the best > way to do so ? Don't have a one-step solution for you, but I'd suggest using "links -dump" to do the conversion, and a carefully assembled procmail rule to deliver to a mailbox on a server somewhere and configuring your handheld to fetch mail from that mailbox. Google will probably produce relevant examples. HTH, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Etax 2007
On Mon, 2007-07-02 at 21:39 +1000, Dean Hamstead wrote: > ran perfectly in wine (whatever is current). Hey Dean, That's awesome. Can you tell us how you did it? I just tried the current Ubuntu packages from winehq.org, and got this message: > Error saving data: Access violation at address 0050021C in module > 'etax2007.txt'. Read of address 0008 after entering my TFN. Did you do something differently? Thanks, James. -- James Gregory -- http://codelore.com -- [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Convert HTML files to RSS feed
Dudes, I have a directory of HTML files (which will be updated periodically) that I need to publish as an RSS feed. I've looked around briefly, but can't see anything obvious. Does anyone know of such a tool? Thanks! James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Copy and Paste
All, I'm becoming increasingly frustrated at the copy and paste behaviour of my linux desktops. Sometimes I can press ctrl-v to paste, other times a middle-click will work, sometimes the two behaviours both work but do different things. I don't really see the pattern. I understand that this is because GTK and X use different copy and paste mechanisms. What I want to know is: Is there any way to make this behaviour more consistent? Can I disable one of (either of) these mechanisms, so that it works one way, all the time? Can I synchronise them? What's the best thing to do? Thanks, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Mandrake - Problems with installing RPMs
On Sun, 2006-07-23 at 10:23 +1000, Nicholas Tomlin wrote: > I installed Gnucash onto my machine initially but can't seem to get it to > run, > sure you can click it and it has the appearance of starting for a second but > then it disappears into hyperspace. I then thought to re-install it but ran > into another problem: > It starts on it and then asks me to put the 'Ultimate desktop CD/DVD' in the > drive, to which I go 'WTF - I thought it was already in there?' So I found it a lot easier to get urpmi (the program behind the thing you're clicking on) to pull packages from the network rather than from disks. My recollection is that you need the automounter to work to use disks, and it very rarely did. Let me know if you need help with this; I don't recall the details right now. I remember their being a tool where you could tell it what country's mirrors to use and it would set it up for you (I no longer have a Mandrake machine around to try this on). However, once you set it up to pull packages from the internet, ask it to install again and hopefully it will pull down all the stuff you need. If you still have it doing the same thing, I'd suggest starting gnucash from a terminal. Then it will spit out some error messages if there's a problem starting it. If we can get those messages it would help in diagnosing whatever problem you initially had. HTH, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Dual head too vast
On Mon, 2006-07-17 at 15:14 +1000, Shane Stephens wrote: > Hi James, > > I'm pretty sure Xinerama and DRI won't mix - I seem to recall noticing > something saying that in the xorg log when I tried to do it. Thanks for the reply, Shane. It appears that you're right. But it also appears that using MergedFB will do approximately the same thing, as will using the fglrx "BigDesktop" thing. These pages have some brief info on the subject: http://dri.freedesktop.org/wiki/MergedFB http://wiki.linuxquestions.org/wiki/MergedFB So yeah, it seems that I can get the desired outcome, but I can't use Xinerama. I'm a bit short on time to tinker with it atm, but I'll give the BigDesktop thing another go at some point. That seems to be the method that will give the best performance. I must say, it's kinda silly that Xinerama won't do DRI the way you've got it setup. I'll have to look into it; surely it can't be too hard to fix. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Dual head too vast
On Mon, 2006-07-17 at 12:34 +1000, James Gregory wrote: > What's the magical incantation I'm missing? I'll happily send my > xorg.conf file if it will help, but don't want to prematurely spam the > list with it. I'm not that attached to the fglrx drivers, so if > switching to one of the plethora other ATI drivers will solve my > problem, I'm quite amenable to that. Ugh. So, it turns out that when I'd specified different bus IDs for each of the two cards (lspci shows it as two devices), it decided that Xinerama hinting was a no-go. If you set it to use the same bus ID, it works fine. Still without 3D though. It claims my server is missing DRI, even though I've got the line in there to load it. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Dual head too vast
Dudes, After much wailing and gnashing of teeth I managed to get a new dual-head Radeon video card doing dual-head with the fglrx drivers. I used xinerama since I couldn't work out how to get the ATI BigScreen thing working (I think at some point I made that work, but the only application that could use the second head was the login screen of gdm; I couldn't use it once gnome-session started). But now when I maximise windows, they end up being 2560 pixels wide (ie, spanning both screens), when what I really want is to maximise to only one screen. Currently I get 'goo' and 'gle' one two screens. What's the magical incantation I'm missing? I'll happily send my xorg.conf file if it will help, but don't want to prematurely spam the list with it. I'm not that attached to the fglrx drivers, so if switching to one of the plethora other ATI drivers will solve my problem, I'm quite amenable to that. Ideally I'd like to get 3D acceleration working too, but I'll let that be for now. Thanks! James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] gnome-terminal and bitmap fonts on Ubuntu
On Mon, 2006-06-12 at 00:55 +1000, Peter Hardy wrote: > On Sun, 2006-06-11 at 15:57 +1000, James Gregory wrote: > > Anyway, I've just worked this out. The trouble is that fontconfig on > > Ubuntu is configured to ignore my beautiful bitmap fonts. There's a file > > called /etc/fonts/conf.d/30-debconf-no-bitmaps.conf which has the > > express purpose of disabling those fonts. > > > > Anyway, if you open that up and comment it out, or remove it by whatever > > means you desire, then you can follow the myriad sets of instructions on > > how to get back your old-skool fonts. > > The Debian Way to do this is to run > > dpkg-reconfigure fontconfig Ha! How did I miss that? Thanks dude. So my next question, what's the gnome-terminal equivalent for setting these xterm X resources: ! Uncomment this to use color for the bold attribute XTerm*VT100*colorBDMode: on XTerm*VT100*colorBD: DodgerBlue1 ! Uncomment this to use color for underline attribute XTerm.VT100*colorULMode: on XTerm.VT100*underLine: off XTerm*VT100*colorUL: red1 Which give me reds and blues for the titles of sections in manpages. I'd probably settle for having a better way to read manpages, but it would be better to have that change across the board. Thanks, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] gnome-terminal and bitmap fonts on Ubuntu
I thought this might help someone else, it's been annoying me for far too long now. I'm quite enamoured with the fonts that xterm uses, but gnome-terminal does some stuff that I'd like from my xterms. Alas gnome-terminal was insistent on anti-aliasing a TrueType fixed-width font. Far too slow and unreadable for me. I'd found a bunch of pages instructing me to unpack the .pcf files into my .fonts directory and everything would be fine. That didn't work for me. Anyway, I've just worked this out. The trouble is that fontconfig on Ubuntu is configured to ignore my beautiful bitmap fonts. There's a file called /etc/fonts/conf.d/30-debconf-no-bitmaps.conf which has the express purpose of disabling those fonts. Anyway, if you open that up and comment it out, or remove it by whatever means you desire, then you can follow the myriad sets of instructions on how to get back your old-skool fonts. Hope that helps someone. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Automatic wireless profiles
On Fri, 2006-06-09 at 03:13 +1000, Jeff Waugh wrote: > > > > So these days I'm on Ubuntu and there's this network manager thing which > > can see that /etc/network/interfaces has both wired and unwired networks, > > and it works out which I'm using and configures things appropriately. > > Which is very nice. > > > > But what I want is a fusion of these two technologies. Is there a way I > > can setup Ubuntu to automatically use a wired network if it's available, > > and if not do this search for available wireless networks, setting the > > appropriate key if one is found? > > That's exactly what NetworkManager does. :-) Plug in wired and you're on, > unplug, and it'll use your preferred wifi network (ie. one that you've > chosen before). Is it not doing this for you? Perhaps you're not actually > using NetworkManager, but the installed-by-default network configuration > tools? If so, I strongly recommend NM, depending on which kind of wireless > hardware you have (the Atheros drivers suck for scanning, so they make NM > entirely pointless, though a very recent version works around some of that > breakage, to the detriment of some niceness). Ok, so it turns out I was only kinda using it. I'd added nm-applet, so when I was plugging it in I was getting the little twirly thing that said it was doing DHCP, but if I clicked on its icon it told me there were no networks. So I read the documentation and discovered that I need to remove the contents of my /etc/network/interfaces file for all the interfaces I want it to manage. If I do that, it scans for APs and gives me a list of them that I can configure by clicking on them. I don't yet know if it can automatically pick one of these APs and set the key correspondingly, or if I'll need to click on the icon and choose one each time. I suppose I'll find out when it becomes an issue. But yes, that was most of the problem. Thanks for your help. It's pretty nifty once you get past that hurdle :) James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Automatic wireless profiles
A long, long time ago I used to run Mandrake on my laptop and I had this nifty utility installed that would work out the ESSIDs of the wireless networks nearby and match them to one of a set of keys I had listed in a configuration file and magically configure my wireless card and run dhclient to get an address. So these days I'm on Ubuntu and there's this network manager thing which can see that /etc/network/interfaces has both wired and unwired networks, and it works out which I'm using and configures things appropriately. Which is very nice. But what I want is a fusion of these two technologies. Is there a way I can setup Ubuntu to automatically use a wired network if it's available, and if not do this search for available wireless networks, setting the appropriate key if one is found? Thanks, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Dapper, gnome-panel and LD_ASSUME_KERNEL
On Sun, 2006-06-04 at 22:44 +1000, James Gregory wrote: > Since just the other day I've been logging into my Dapper notebook only > to discover that gnome-panel is "frozen". Doesn't respond to mouse > clicks etc. I straced it and found out that it's stuck on a futex. In case anyone was wondering, I eventually "fixed" this by running gconftool --recursive-unset /apps/panel and killing gnome-panel. That reset it to its default configuration, which works. Presumably I had added some rogue applet which caused it to fail. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Dapper, gnome-panel and LD_ASSUME_KERNEL
Since just the other day I've been logging into my Dapper notebook only to discover that gnome-panel is "frozen". Doesn't respond to mouse clicks etc. I straced it and found out that it's stuck on a futex. On a hunch I made a .xsession file that ran LD_ASSUME_KERNEL=2.2.5 gnome-session So that it would use old-skool threads. Now, that's solved the problem, I can click on stuff and it all behaves as you'd expect, but it's a pretty nasty solution. Anyone know what's going on? Is there an easier way to debug this than downloading all of gnome and building it with debug symbols etc? Thanks dudes, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: Working PCI wireless [Was: [SLUG] PCI Wireless cards]
On Mon, 2006-05-29 at 21:39 +1000, Jeff Waugh wrote: > > > > For the archives, I ended up getting a Linksys WMP-54G and it worked > > straight out of the box. Most impressed. > > Which driver? Dude, I'm not meant to know those things! I just click at stuff and things. I need a pie-chart. STAT! But inspecting lsmod, the most likely culprit appears to be the rt2500 driver. It comes up as ra0. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] PCI Wireless cards
On Wed, 2006-05-24 at 15:12 +1000, david wrote: > I've got a netgear wg511t that worked painlessly with Breezy... plugged > it in and it worked. For the archives, I ended up getting a Linksys WMP-54G and it worked straight out of the box. Most impressed. James. > On Wed, 2006-05-24 at 14:58 +1000, James Gregory wrote: > > Dudes, > > > > I need to buy a PCI wireless network card. I need it to work with Ubuntu > > (dapper) with the least amount of effort possible. Speed isn't so > > important, but I'll get the fastest one that meets my zero-effort > > criterion. What should I buy? > > > > There's this one: > > > > http://www.elx.com.au/item/bfDLI40605 > > > > But the page uses words like "problematic" and "pioneer", which makes me > > wary. Is there a better option? > > > > Many thanks, > > > > James. > > > > > -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] PCI Wireless cards
Dudes, I need to buy a PCI wireless network card. I need it to work with Ubuntu (dapper) with the least amount of effort possible. Speed isn't so important, but I'll get the fastest one that meets my zero-effort criterion. What should I buy? There's this one: http://www.elx.com.au/item/bfDLI40605 But the page uses words like "problematic" and "pioneer", which makes me wary. Is there a better option? Many thanks, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Portmap CPU usage
I run an NFS server for a small number of files from a computer I do
some dev work on. On that computer, portmap is using around 20% of CPU
when I'm not doing anything on these NFS clients, and that figure rises
significantly when I actually want to use it.
Stracing portmap shows that it seems to be opening and closing
the /etc/hosts.{allow,deny} files over and over. This not at all
necessary, since I never change those files and the security of this
particular machine is just not a priority.
Is there a way I can make it stop polling those files (or even just make
it stat them first)? Alternatively, what else can I do to reduce
portmap's CPU usage?
I'm also a little curious as to why it's doing anything at all; I don't
think anything is requesting files. Is there a way I can debug where all
these requests are coming from?
Thanks,
James.
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] My mythtv saga...
On Fri, 2006-01-13 at 13:40 +1100, David Gillies wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > James Gregory wrote: > > On Thu, 2006-01-12 at 09:41 +1100, Michael Fox wrote: > > > >>As people would of last saw I bought a DVICO fusion plus card and soon > >>after ditched it. > > > > Just on this, I got my DVICO card working with the stock kernel in > > ubuntu-dapper just last night. It appears that the bug in the driver was > > eventually fixed. > > I actually got my DVICO dvb-plus working on stock breezy badger last > month. What bug was stopping this card from working previously? (I'd > only tried Knoppmyth with no success whatsoever with this card previously) There were a number of different versions of the card released, which used different tuner chips at different memory addresses. Some of the variants worked, mine didn't. I don't know if all of them now work, but mine does. :) James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] My mythtv saga...
On Thu, 2006-01-12 at 09:41 +1100, Michael Fox wrote: > > As people would of last saw I bought a DVICO fusion plus card and soon > after ditched it. Just on this, I got my DVICO card working with the stock kernel in ubuntu-dapper just last night. It appears that the bug in the driver was eventually fixed. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: pentium M series
On Sun, 2005-12-25 at 09:34 +1100, O Plameras wrote:
> James Gregory wrote:
>
> >On Wed, 2005-12-21 at 16:40 +1100, O Plameras wrote:
> >
> >
> >>#define spin_lock_init(lock)do { (void)(lock); } while(0)
> >>#define _raw_spin_lock(lock)do { (void)(lock); } while(0)
> >>
> >>I am reading this cited macros. This is very very clear.
> >>
> >>It is very very clear what it says, and, i.e., regardless of what
> >>is the value of 'lock' the kernel should do nothing.
> >>
> >>
> >
> >That's exactly right. And if you compile without CONFIG_SMP, that's what
> >gets built into your kernel. You can get away with it because of the
> >clever way in which a CPU does one thing at a time; there is no "true"
> >parallelism.
> >
>
> By the way, is it not true that 'pipelining' that's a feature of x86
> CPU's starting
> with i586 which I have pointed out in one of my previous post is
> (another name)
> implementation of 'parallel' processing ?
I suppose I wasn't entirely clear here. The precise kind of parallelism
I was referring to was concurrent threads of execution. The kind that
requires locking (as you may have inferred from the context).
Just about every CPU I can think of implements some form of
'pipelining', and it's good friends 'super-scalar, out-of-order
execution'. There is an important distinction between these
optimisations and using explicit threads: the former is invisible to the
programmer.
My statement wasn't entirely true, CPUs do many things simoultaneously,
and of import, they can perform a very limited number of instructions in
parallel. But, in addition to evaluating these numerous instructions
concurrently, they are also ensuring that at the other end, it all looks
like it happened serially, but very, very fast.
It's kinda interesting actually, the chip looks ahead some number of
instructions and will identify groups of instructions that don't have
dependencies on each others' outcomes. It then tries to fit those sets
of instructions to the various bits of silicon in the chip (for example,
the integer add unit and the floating-point divide unit will (almost
certainly) be different bits of silicon), those operations will be
performed concurrently and since they are all independent (no one of
this group depends on another's outcome), it won't make any difference.
That's where the all-important "instruction re-ordering" comes in; the
compiler must try to order instructions to make it possible for the CPU
to identify such groups of instructions.
Note however that isomorphic re-orderings of code will have no effect on
the outcome of executing it, so are somewhat irrelevant to a discussion
of locking.
> This means that more than one
> instructions may be
> executed in one clock cycle. This is implemented by using a bus
> interface unit
> (BIU) and an execution unit. Experts on Intel Arch may confirm the
> truthfullness or
> falsehood of this assertion. (I'm not an expert, I just know by
> researching).
There's a heap of 'units', dude. Modern CPUs are insane.
> With pipelining, the CPU overlaps instruction fetching and decoding with
> instruction execution, i.e., while one instruction is executing BIU is
> fetching
> and decoding the next instruction. So, assuming you're willing to add
> hardware
> you can execute more and more operations in parallel.
Yes and no. I suppose you could construct a chip whose explicit purpose
is "to pipeline", but that's not really why hardware engineers started
doing it. It's more about making the most of the silicon you've got. You
don't want bits of silicon (that cost money) to be sitting there idle.
Pipelining just lets you make sure that more of your silicon is busy
more of the time.
> So, in this way there is true parallelism in x86 arch.
I suppose it's a matter of interpretation. It was extremely intentional
that I chose to refer to the type of parallelism in UP-x86 as bereft of
"true parallelism". You can of course choose to interpret those words
however you will. Having made explicit exactly what level of parallelism
UP-x86 provides, the only really important thing to realise is that it
is not a sufficient level of parallelism as to have any effect on
locking. UP kernels don't crash because they don't have locking. They
would if they had what I have referred to as "true parallelism". I'll
happily adopt another term for it if you can suggest a better one.
Merrily,
James.
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: pentium M series
On Wed, 2005-12-21 at 17:58 +1100, O Plameras wrote: > Spinlock is shown for UP. > > What is spinlock for SMP ? I omitted it because there's substantially more code to do it. The short answer (which I'm only half sure of, I'm sure someone else here can confirm or deny my allegations) is that it disables pre-emption, takes out a bus lock and does all the standard stuff. It's a binary semaphore, as per Dijkstra's original formulation of it. Here's a wikipedia article on semaphores: http://en.wikipedia.org/wiki/Semaphore_%28programming%29 Here's the obscure assembly code that does the hard work: #define spin_lock_string \ "\n1:\t" \ "lock ; decb %0\n\t" \ "jns 3f\n" \ "2:\t" \ "rep;nop\n\t" \ "cmpb $0,%0\n\t" \ "jle 2b\n\t" \ "jmp 1b\n" \ "3:\n\t" which you'll find in include/asm-i386/spinlock.h. HTH, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: pentium M series
On Wed, 2005-12-21 at 16:40 +1100, O Plameras wrote:
> #define spin_lock_init(lock)do { (void)(lock); } while(0)
> #define _raw_spin_lock(lock)do { (void)(lock); } while(0)
>
> I am reading this cited macros. This is very very clear.
>
> It is very very clear what it says, and, i.e., regardless of what
> is the value of 'lock' the kernel should do nothing.
That's exactly right. And if you compile without CONFIG_SMP, that's what
gets built into your kernel. You can get away with it because of the
clever way in which a CPU does one thing at a time; there is no "true"
parallelism.
If you compile *with* CONFIG_SMP, those definitions get preprocessed out
and some other macros get compiled in that do things like disable
interrupts, bus lock and do atomic writes and so forth, with the net
effect of implementing a spin lock.
If you install the -smp kernel on Ubuntu today, you get the version
compiled *with* CONFIG_SMP, and the code I cited will be compiled out
and substituted with this devious spin-lock implementing spin lock code.
Otherwise you don't, and (unsurprisingly) you get UP support, with no
spin locks, which you've verified for us with your insightful analysis
of those macros.
Jeff explained that right now Ubuntu ships only with the non-CONFIG_SMP
kernel on the CD due to space constraints *and* that subsequent releases
of Ubuntu have a solution to this UP/SMP problem in a single kernel
image. If I understand correctly, it's implemented with some runtime
function-table rewriting, which means you would only incur the cost of a
function call, rather than the bus-lock etc.
I hope that's clear.
James.
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: pentium M series
On Wed, 2005-12-21 at 15:23 +1100, O Plameras wrote:
> James Gregory wrote:
>
> >/*
> > * If CONFIG_SMP is unset, declare the _raw_* definitions as nops
> > */
> >#define spin_lock_init(lock)do { (void)(lock); } while(0)
> >#define _raw_spin_lock(lock)do { (void)(lock); } while(0)
> >
> >
> >
>
> Is this euphemism for "in ubuntu SMP is not implemented" ?
The piece of code I was citing comes from a region protected by
#ifndef CONFIG_SMP
And as I noted, the equivalent region which is compiled in when
CONFIG_SMP *is* set, actually implements a spin lock.
So I suppose you could think of it as a euphemism for "when SMP is
disabled, SMP is disabled".
James.
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: pentium M series
On Mon, 2005-12-19 at 14:40 +1100, Ian Wienand wrote:
> Mostly I think the sub-architecture is passed via different flags to
> gcc, which can try to optimise the code. Talking with people from SGI
> who look into that sort of thing, the benefits of even a much better
> compiler fall into noise compared to things like lock contention.
Right. Which is what I've found too. Which led me to wondering: Jeff:
how do you plan to do this UP/SMP kernel efficiently? The kernel
implements spin_lock as a macro that gets compiled out when building for
UP:
/*
* If CONFIG_SMP is unset, declare the _raw_* definitions as nops
*/
#define spin_lock_init(lock)do { (void)(lock); } while(0)
#define _raw_spin_lock(lock)do { (void)(lock); } while(0)
(from include/linux/spinlock.h on 2.6.10)
(the first line is not so exciting, but the second one is pretty
important).
The corresponding code for when CONFIG_SMP is set actually implements a
spin lock, along with all the consequences you'd expect.
Does Ubuntu have a magical solution to this, or do you just execute the
extra code on UP?
Yours,
Curious in Woolloomoolloo.
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] drivers
On Mon, 2005-11-07 at 16:06 +1100, Christopher JS Vance wrote: > I've been asked to get a particular touchscreen (eGalax usb) going > before Tuesday next week on Breezy. > > The manufacturer has actually provided some Linux software in the way > of binary bits plus some sources for a number of old rpm-based > distributions (RH<=9, FC<=3, Mdk<=10), all for XFree86 and a 2.4 kernel. Yuck. deb -> alien is surmountable with alien. Xorg -> XF86 has a good chance of still working -- that's just normal C linkages, not wacky like the kernel. 2.6 -> 2.4 is problematic. The module loaders are fussy about what they'll load at the best of times. You might get it to load with 'insmod -f' or similar, but you might also just crash your computer. Or it might do nothing, or it might just fail to load. As horrible as it sounds, I'd be looking at using one of their supported platforms unless someone knows of a more friendly driver for this piece of hardware. Is it this thing: http://www.egalax.com.tw/eg/drivers.htm Looks to me like the driver packages they have include source code. Have you tried building it to see what happens? James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Using Debian Alien
On Tue, 2005-10-18 at 07:54 +1000, Richard Hayes wrote: > Dear list, > > I found a Mandrake RPM I wanted to use and since it was a new project I > downloaded a Mandrake CD 10.1 and installed it on an old box. > > So I was have been swimming in RPM Dependancy custard for many hours. urpmi can probably just solve that problem for you. Might be easier. > So, rather than fight with Mandrake I throught I would use Debian with Alien. > > I used apt-get to install Alien but I could not find any instruction for > using > Alien I usually invoke it as 'alien -d -c -k', which will get you a suitable deb file. It won't help however, since your aliened deb package will be bereft of the dependency information you'd need to get it to work. James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Driver query
On Tue, 2005-10-04 at 12:26 +1000, Michael Fox wrote: > Anyone able to confirm if the Dvico Fusion HDTV card drivers are now > merged into recent 2.6 kernels? ie. no need to grab the files and > compile into a kernel. > > About to embark on playing around with one and all the howtos I have > found are out dated, so was wanting to confirm if the process about > the driver can be skipped. According to the site, they have been for some time. I'm running mine on the Ubuntu Breezy kernel. It works very well, but there's a small bug[1] in that kernel (which I suspect comes directly from upstream), which you'll need to fix in order to make it work. I don't know if the problem is solved in 2.6.13. Try the breezy kernel, see if that works (and I'd be very interested if it does; I've only got the one card to test my changes on). If not, I can make the debs for my modified kernel available. As for Myth, I've been unable to make it work, but I didn't try very hard (it's less trivial than just 'aptitude install mythtv'). HTH, James. [1] http://bugzilla.ubuntu.com/show_bug.cgi?id=5773 -- % Will all this regular expression scariness go away in % Perl 6? Yes, definitely; we're replacing it with a completely different set of scariness. -- http://www.perl.com/pub/a/2003/07/01/regexps.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Portfw: iptables or tcp proxies
On Tue, 2005-10-04 at 08:10 +1000, Carlo Sogono wrote: > We have been using tcp proxies (simpleproxy/tcpproxy) to forward ports > from our firewall to internal machines. Recently I have changed this > setup to regular iptables port forwarding. My understanding is that > since tcp proxies are daemons that accept connections from the > Internet and make connections to internal IPs, they would be a lot > slower than iptables (kernel-level routing?). Are there any advantages > to running tcp proxies instead of using port forwarding? That's a hard question in general. One thing that a "TCP proxy" does is to ACK packets as soon as they get to the router, rather than when they reach their endpoint. This is a desirable property if your gateway forwards to a high-throughput but also high-latency link (satellite for example). It does break TCP semantics though. TCP layer proxies are also able to insulate your network from "broken TCP packets". The problem is that unless you run a well-written TCP proxy, you then open up your firewall to attack. If you're just using this to forward to SSH, then I'd be using packet-forwarding. If it's web, I'd probably run Squid or similar in front of your web-server. But straight "TCP proxies" don't offer that many benefits on their own unless you have very specialised needs (assuming you keep your kernels patched on the machines in your network). HTH, James. -- "Now, there are no problems – only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Qemu Questions
On Thu, 2005-07-21 at 08:08 +1000, Peter Rundle wrote: > James Gregory wrote: > > > The user-net option will allow your guest system to make *outgoing* > > connections. You can't (without some tinkering) use it to run an > >externally accessible webserver for example. > > For my application that will suffice. > > Just need a little clarification about the network config, > I understand that the guest gets an 10.0.2.x ip from qemu's "virtual > dhcp" server and it's default gateway is set to 10.0.2.2 which is... > (I'm guessing) qemu creating a "virtual gateway", does Qemu then Nat the > outbound packets from the guest to the Hosts ip address? I.E Qemu is a > user level application running on the host which has access to the > hosts networking services? That's more or less right. qemu makes userspace sockets to connect to whereever, just like any other application. It doesn't let you send raw ethernet frames out (or even raw packets); you'd need to be superuser for that. > > I don't understand; how can you tell that there's an hourglass there if > > everything is black? Can you point us to a screenshot? > > Just the cursor is black. The rest of the display is the normal colour. > When you launch an application the busy/hourglass cursor appears but it > is also black (windows nt4 guest). Not to worry but curious bug. If I > launch the guest without the loadvm (i.e boot from scratch) the cursor > colour is normal. Ahh. I misunderstood. I'm guessing it's a bug in the video bios for qemu's "virtual video card". If it were linux as your host, setting SoftwareCursor would no doubt fix the problem, so I guess you can either: - Find the Windows NT equivalent of SoftwareCursor - Get another video bios (I think there's a few that work) - Get a video driver for Windows NT that plays nice with the vga bios you have. HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Qemu Questions
On Wed, 2005-07-20 at 08:54 +1000, Peter Rundle wrote: > James Gregory wrote: > > > I have actually had this working in the past, but I remember it being a huge > > amount of effort. I don't know if you can get away with it for what you > > need to > > do, but the user-net option is much simpler to get going. I suggest you try > > that; it will at least tell you if your guest OS is setup correctly. It will > > also let you run it as a normal user. > > Err ok I thought that user-net option didn't actually give the PC > network connection. The documentation on qemu that I found so far is > pretty thin, just enough to get an expert going, not being one of those > I'm stuggling a little. So does the user-net option actually give you > network connectivity and if so why would anyone bother with the other > method? The user-net option will allow your guest system to make *outgoing* connections. You can't (without some tinkering) use it to run an externally accessible webserver for example. user-net lets me fire up a debian virtual machine, set up networking with dhcp (qemu has its own dhcp server for its "virtual network"), and pull packages down with apt. The other options do have their utility, but for most people, user-net is the right choice. > > > Also, which version are you running? The 0.6 version had a bug with > > networking, > > though I don't recall the details. > > 0.7 ok, that should be fine. > > > I've seen that before, presumably because it doesn't try to redraw until > > something writes to its video memory region. It generally works ok once > > there's > > been some activity on the guest OS. Is this not the case for you? > > No moving the mouse etc, even the busy "hourglass" cursor is black. It > all works it's just black, not that there's anything wrong with being > black ;-) I don't understand; how can you tell that there's an hourglass there if everything is black? Can you point us to a screenshot? HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Scroll wheel speed
I noticed today that I'm spending a truly incredible amount of time operating the scroll wheel on my mouse. Can anyone tell me how to speed it up? 3 or 4 times as fast as it is now would be awesome (by which I mean each 'click' of it will scroll 3 or 4 times the distance it currently does). I looked in the mouse preferences thing and googled around a bit and couldn't find anything. I'm running Ubuntu Breezy. Thanks, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] SQL Brain teaser...
On Wed, 2005-06-01 at 19:59 +1000, Grant Parnell - EverythingLinux wrote: > I'm trying to do some metrics on how long it takes to process an order in > our system based on time between printing the picking slip and bagging the > goods for dispatch. The aim is to determin the maximum performance if we > were to have one guy do nothing but pick & pack orders. > +---+-+-+--+ > | ordno | pickslip_printed| docketdate | packmins | > +---+-+-+--+ > | ordno | 2005-06-01 14:32:16 | 2005-06-01 14:34:47 |3 | So the issue is cases like this one where three orders are processed in parallel: > | ordno | 2005-06-01 15:12:27 | 2005-06-01 15:27:26 | 15 | > | ordno | 2005-06-01 15:12:28 | 2005-06-01 15:30:25 | 18 | > | ordno | 2005-06-01 15:12:29 | 2005-06-01 15:21:53 |9 | So, if we want to consider the time each order actually takes to process, and we're allowing parallel packing we need to make some assumptions about what's going on with that parallel processing. Let's make this assumption: in overlapped time periods, it is acceptable to use the docketdate to demarcate time spent on different orders. So the analysis you're looking for then is not how much time can directly be attributed to each order, but rather how much time is spent on each order, without necessarily knowing which order took that time (you don't have enough data there to do direct attributions even if you wanted to). If we can assume that then you can make some headway on the problem -- you can flatten your data-structure to a time-line of events and measure the time between events. Further, you can also ask the database to ignore time when no order processing is occurring. (excuse my SQL here, it's been a few months) select eventdate, eventname, eventdiff from (select pickslip_printed, 'printed', 1 from orders) union all (select docketdate, 'docketed', -1 from orders) order by eventdate asc; Which discards the relationship of times to specific orders, and allows analysis of just the time elapsed between events. In addition, the eventdiff column will allow you to work out how many orders are outstanding at any point in the timeline. I'd make that into a view, and then build another view that evaluates the sum of the eventdiff column up to that row for every row. Whereever that sum is > 0, there are orders being worked on. Once you get to that point, the elapsed time on an order will be the difference between some of the adjacent rows, which is similarly easily calculated. I say "some of", because of course the one second that elapsed between those orders being "printed" does not represent packing time. If you think about it, you can solve that problem but that gets a bit fiddly (you'd need to differentiate between print-print, print-docket and docket-docket scenarios. The latter two would represent order processing time, which is doable, but you need to keep inter-row state, which is slightly tricky in SQL), so you're probably better off employing some kind of heuristic given what you know about the data (eg that you don't believe an order could be packed in < 30 seconds or similar). Not bullet-proof, but it's a start. HTH, James. (recovering SQL abuser) -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] -Wshadow
On Wed, 2005-05-25 at 07:36 +1000, Erik de Castro Lopo wrote: > When I'm developing, I also add -Werror. Always! > > In the near future, I will probably also add -Wswitch, -Wfloat-equal, > and -Winline. We added -Winline recently in order to solve some build problems on pre-historic RedHat. The problem is that it means I need to remove inlines which are probably ok in the context they were initially intended, but for whatever reason aren't ok now. Now, if you have a great deal of faith in the compiler's ability to optimize then that's all fine. The trouble is that I simply lack that faith in gcc's ability to optimize code (especially on P4). Hmm. Which I've just realised means that I was operating on the assumption that gcc would generate multiple copies of the code if there were places it could be inlined and other places where it couldn't. I don't actually know if that's true. Interesting discussion. James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Videos in presentation software
On Thu, 2005-05-05 at 12:30 +1000, James Gregory wrote: > Hi all, > > I need to give a presentation which will include some videos. Currently > they're avi files, but I'm just as happy to make them mpeg or whatever. I ended up using the 100build of OpenOffice 2 beta. I used the instructions on this page: http://www.ucolick.org/~zingale/OOo_movies/ It worked ok but dropped a lot of frames in the first video I showed with it. Don't miss the step about altering your classpath *inside* OpenOffice; That drove me mad for quite a while. It also seems pretty fussy about what codecs it will use; but less fussy than doing the same thing in PowerPoint. Another oddity is that it behaves differently with different JDKs. Sun's 1.4 JRE opens up a new window for the video while the 1.5 one will embed the movie correctly but it only works once. Anyway, you *can* do it. I thought others here might find the information useful. James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Videos in presentation software
Hi all, I need to give a presentation which will include some videos. Currently they're avi files, but I'm just as happy to make them mpeg or whatever. I've tried inserting them in various tools, with varying levels of success: * On Ubuntu (Hoary) it crashes openoffice (stable version, whatever's in Hoary). * On Mandrake it inserts the video but won't play it (I get a white box in the middle of the slide) (1.1.4 I think. The current version in cooker) * On Mandrake with the 100 build of the 2beta version, it pops up a dialog telling me "The format of the selected file is not supported." (and I've tried with quite a few different file formats). * I just can't find the relevant option in the KOffice presenter thing. Can someone point me at a tool that will do what I want? Will MagicPoint display video files? Easily? What other options do I have? Many thanks, James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Mac OS CD
On Sat, 2005-04-16 at 19:53 +1000, Kazik Malenczak wrote: > I using Mandrake 10.1 and wish to have a flirt with PearPC but I cant > get the MacOS cd to be read. I have all the HFS and HFS+ stuff i think i > need. Do i need anything special in fstab to be able to read or should > auto do the trick? The kernel also has the HFS stuff compiled. > Any help much appreciated It doesn't use the host's filesystem code to access your CD drive. It presents the IDE drive to the client operating system directly. In order to make it work, you need to setup its configuration file to point to your CD-ROM device file. My recollection is that the documentation was relatively clear on the subject, but it has been a while since I used it last. I recommend making an ISO image of your OSX CDs, and working from that. It's likely to be faster. This page: http://pearpc.sourceforge.net/getstart.html has some details on getting access to your CDs. HTH, James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] libcrypto, libssl needed for ProFTPd
On Tue, 2005-04-12 at 21:43 +1000, Voytek wrote: > > > Quoting Voytek <[EMAIL PROTECTED]>: > > >> apache/php/mysql/mail on P2 320MB > > > > Hrmm. The PHP element there makes that sound like a mildly difficult > > upgrade. > > why is that ? Because PHP has had a number of incompatible language changes between versions. RPM will be able to manage the job of replacing the php .so (or executable), but it is likely that there will be code modifications that you'll need to do in order to keep your site running. There's workarounds for that of course (you *can* run multiple versions of php on the same machine, but it's a real pain), but you'll have to do some legwork either way. > > > Which MTA are you using? > > Postfix with MySQL Postfix won't present problems. I don't remember how straightforward it is to upgrade MySQL. Can anyone here comment? I'd take a mysqldump of the database before you do anything. HTH, James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] libcrypto, libssl needed for ProFTPd
Quoting Voytek <[EMAIL PROTECTED]>: > > Perhaps you can tell us about the system and we can help you get it > > upgraded? > > What do you use this box for? > > James, thanks. > > apache/php/mysql/mail on P2 320MB Hrmm. The PHP element there makes that sound like a mildly difficult upgrade. Which MTA are you using? Getting apt setup as below is probably a good first step. > > > I'm curious though, why are you asking yum to install dependencies you've > > manually specified and not asking it to install proftpd and fetch all of > > its > > dependencies for you? > > > > yum install proftpd > > > > seems easier. > > it sure is easier, (and, quicker too) but: > > > # yum install proftpd > > Cannot find a package matching proftpd > > I guess I'd need point yum elsewhere than fedora site ?? Yep, I'd say so. freshrpms has packages for proftpd: http://ftp.freshrpms.net/pub/freshrpms/redhat/7.3/proftpd/ I'm not sure if freshrpms supports yum, but they definitely support apt-rpm. I might be being optimistic, but I think if you install this package: http://ftp.freshrpms.net/pub/freshrpms/redhat/7.3/apt/apt-0.5.5cnc5-fr0.rh73.2.i386.rpm and potentially edit /etc/apt/sources.list to point to their packages for redhat 7.3, you should be able to do this: apt-get update apt-get install proftpd to do what you want. I'm unfortunately not near an RPM based box atm, so I can't check precisely what the apt package I've referred to does. The only variable is editing /etc/apt/sources.list. Give it a go and post back if you have any issues. HTH, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] libcrypto, libssl needed for ProFTPd
Quoting O Plameras <[EMAIL PROTECTED]>: > So, what is the mechanism for it to know if there is or not. RPM keeps a database that records what files are provided by each package. Along with that information are details such as MD5s of the files, permissions, modification dates etc. If a package has a file dependency, RPM will not check whether or not the file exists on the filesystem, it will check to see if it has been provided by a package in its database. If the database says that it exists, it is assumed to exist even if that is not actually the case (and if it is not the case, something is wrong with your computer). My recollection is that this was not always the case, but it has been at least for the 4.x versions of RPM. It is for this reason that building a library and installing it yourself will not satisfy a library dependency for an RPM; even though if you forced the installation you could coerce it to work. I got caught on this concept for quite a while before I appreciated all the magical things that package management does for me. In short: file dependencies are evaluated exactly as per package dependencies. > > And if there is, how will the program loader find where > to load the libraries required ? Ahh. That's a separate question. Happily, there are some standards regarding the structure of RPMs that make these kind of problems straightforward. There's a few directories where (properly built) RPMs will install ".so"s and those directories are known about by the loader. /usr/lib and /lib are amongst those directories. As part of the post-installation scripts triggered by any properly built library RPM, ldconfig will be executed to update ld.so.cache accordingly. Hope that clears things up, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] libcrypto, libssl needed for ProFTPd
Quoting Voytek <[EMAIL PROTECTED]>: > > > Edit your /etc/ld.so.conf and insert the the directory path > > excluding the filename. RPM doesn't inspect ld.so.conf in walking dependencies. This will have no effect on the installation. > (hhhm, perhaps Eric was right telling me to give it up...?) I'm inclined to agree that 7.3 is a little old to be expecting updates on. Perhaps you can tell us about the system and we can help you get it upgraded? What do you use this box for? I'm curious though, why are you asking yum to install dependencies you've manually specified and not asking it to install proftpd and fetch all of its dependencies for you? yum install proftpd seems easier. What was the source of your proftpd rpm anyway? At any rate, this package: http://rpm.pbone.net/index.php3/stat/4/idpl/776260/com/openssl096-0.9.6-23.7.i386.rpm.html claims to provide the files you're after. It might be a solution to your imediate problem. I generally search on rpm.pbone.net for the filename when these dependency questions come up. They have quite an extensive index. HTH, James. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Wine & setup
On Tue, 2005-04-05 at 14:39 +1000, Alan L Tyree wrote: > On Tue, 05 Apr 2005 14:09:06 +1000 > James Gregory <[EMAIL PROTECTED]> wrote: > > > On Tue, 2005-04-05 at 13:09 +1000, Alan L Tyree wrote: > > > I have a windows based CD that I would like to install and use with > > > Wine. The setup.exe prog crashes. > > > > > > What's the best way to proceed? is there anyway to extract and > > > install without using the setup.exe program? > > > > There's a number of ways; it's highly dependent on how the setup.exe > > program works. However, I happened to be looking into something for > > entirely separate reasons just last night and discovered this: > > > > http://support.microsoft.com/default.aspx?scid=kb;en-us;257718 > > > > Which might help you out. IIRC, the MSI installer program > > (instmsi.exe) does run under wine, so getting an MSI file should be > > enough to get you up and running. > > > > All that said, there might be an easier way depending on your app. > > What wine version? What program? What type of installer is it? > > (nullsoft, installshield, etc). > > I don't know what kind of installer it is: here is a list of the files: > [EMAIL PROTECTED]:~$ ls /cdrom > setup.exe > stdole.tlb Those two files make me suspect that it's InstallShield (which is pretty likely; it's very common). InstallShield installers don't work properly right now because of some issues with type-marshalling (the .tlb file is the giveaway; that's the marshalled types) and dcom. Fortunately it's a common problem, so there are some partial solutions. I believe if you install dcom98 in your fake windows partition from the .exe available here: http://www.microsoft.com/downloads/details.aspx?FamilyID=08b1ac1b-7a11-43e8-b59d-0867f9bdda66&DisplayLang=en and copy a stdole32.tlb file from a 'real' windows partition (presumably to somewhere under 'windows' in your fake windows directory; you'll need to check up on that), it will work. If you search for 'installshield wine' on google, you'll find a bunch of useful information on it. The other option is a utility called 'i6comp', which is floating around. It just extracts the files from an InstallShield installer. It is in turn a windows application, but will run out of the box with wine. The other option is preferable though, because it will actually set the various registry entries your program will need. It's likely you'll need to do extra tinkering if you go down the i6comp route. HTH, James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Wine & setup
On Tue, 2005-04-05 at 13:09 +1000, Alan L Tyree wrote: > I have a windows based CD that I would like to install and use with > Wine. The setup.exe prog crashes. > > What's the best way to proceed? is there anyway to extract and install > without using the setup.exe program? There's a number of ways; it's highly dependent on how the setup.exe program works. However, I happened to be looking into something for entirely separate reasons just last night and discovered this: http://support.microsoft.com/default.aspx?scid=kb;en-us;257718 Which might help you out. IIRC, the MSI installer program (instmsi.exe) does run under wine, so getting an MSI file should be enough to get you up and running. All that said, there might be an easier way depending on your app. What wine version? What program? What type of installer is it? (nullsoft, installshield, etc). HTH, James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Evolution - a PITA?
On Fri, 2005-04-01 at 15:44 +1000, Howard Lowndes wrote: > Is anyone else finding Evolution to be a real PITA? No, I think it's a great piece of software. > I thought v1.4 was bad enough with its regular crashes, but I think v2 > is worse with its cycle hogging, and the compose window stalling for up > to a minute. Hm. I run it on my laptop all the time (which I have throttled to 600MHz so it doesn't make so much noise) and have found it very snappy. Particularly when switching between mailboxes -- I found mutt's constant re-indexing of mailboxes a real pain in that situation. It does use a fair bit of RAM though; perhaps you're just seeing delays from swapping. The only thing I miss from 1.4 is the RSS reader. > If anyone can suggest an easy way to export the address book then I > would like to look at thunderbird instead. If you highlight all your contacts and right-click, there's an option in that menu for 'Save as VCard', which spits out a .vcf file. I believe that's a format that Thunderbird knows how to read. HTH, James. -- My love burns for you A Thermonuclear rose Now lets go make out -- http://www.solardeathray.com/rose.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] How do I mount an audio Cd ?
On Sun, 2005-03-13 at 01:32 +1000, QuantumG wrote: > Jeff Waugh wrote: > > >Windows is lying a little bit, to give you a nicer interface. Audio CDs are > >not like data CDs, and cannot be mounted. > > > > > > From a purely philosophical point of view, what would be a good reason > for not have a kernel module that mounts audio CDs by interpreting the > red book format? Seems kind of silly to have code at the application > level doing this low level interpretation. Well, I actually recall there having been a kernel module to do precisely this. In fact, this page: http://www.ii.pw.edu.pl/~borkowsm/cdfs.htm agrees with me. The issue is that the format of an audio CD isn't quite like other random access filesystems that you'd normally deal with in a kernel module. This page: http://www.xiph.org/paranoia/faq.html#play Has some details about it. Now, that in and of itself doesn't mean you *can't* do it in the kernel, but I'm personally of the opinion that it makes more sense to leave that stuff in userland. There are some userspace filesystems that will do this though. KIO's audiocd: plugin does this really nicely. GnomeVFS has some code to do it in the standard distribution, but it's turned off by default (at least it did the last time I was digging through gnome-vfs code. Good times.). I'm actually pretty sure that's how Windows does it as well. But you were after a philosophical answer, so here goes: Developing stuff in the kernel is harder than developing it in userland (as but one example, I'm stuck using C in the kernel, but I can use the obscure language of my choice in userland). The kernel's job is to provide primitives that make it easy for application developers to get this stuff working. Now, there might be a convenience angle to mounting an audio CD (though I'm not sold on it), but I don't think it really serves as a better application development primitive than a library in this case. So it shouldn't be in the kernel because it can just as easily be elsewhere. HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] security different on vfat compared to ext ?
On Wed, 2005-03-02 at 13:03 +1100, Rod Butcher wrote: > Hello sluggers, I've moved an Apache intranet I run on my home office > from an ext3 to a vfat partition. I did this by copying the DocumentRoot > using nautilus and changing the conf files. Apache now returns "You > don't have permission to access /bookmarks.htm on this server" to the > browser. > I don't have this problem if I copy the webserver directory to another > ext3 partition - seemingly proving that I'm changing the necessary conf > info. > So - is there something different about ext and vfat security, > necessitating some more sophisticated directory copy process ? The problem is that Apache will be checking for file attributes that VFAT simply doesn't support. There's options in there somewhere to turn off ownership checking and checking that the file is/isn't executable. I don't know them off the top of my head though, so you'll need to read the manual for it. Why do you need to do this? There might be a better way to achieve what you're trying to do. HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] C newbie seeks directions
On Thu, 2005-02-17 at 14:32 +1100, Rod Butcher wrote: > Hello Sluggers, I'm having to teach myself some C so I can deal with > debugging problems with C modules used by perl (my primary interest is > the perl scripts, but I'm tired of feeling helpless when C programs > won't build or just die). > > I've found an online university course tutorial which covers basic data > types, operators, functions, prototyping, structures, pointers, > malloc :- > http://www.cs.cf.ac.uk/Dave/C/ > It's dated 1999. Should this be enough, any major changes since then, > any recommended tutorials out there ? There hasn't been too many changes since then. I believe C99 is still the accepted standard standard (can anyone verify that?) > Also - am I OK just working with a text editor like Gedit, or do I > really need to use some API to do things properly ? I assume you mean IDE. If you're happy with Gedit, then that's great. I like vim personally. Two tools I find invaluable in navigating C code (and are largely editor agnostic) are ctags and cscope. If you're on debian you want the exuberant-ctags package. Then you run 'ctags -R' on your C source to build a database out of it, and anything you want to see the definition of (functions, structs etc), you put the cursor on and hit ctrl-] and you'll be jumped straight to it. ctrl-t takes you back to where you were. cscope is useful essentially as the inverse of ctags, and there's vim bindings for it too, but it's perfectly usable as a stand-alone app. Presumably other-editor-experts can tell you the equivalents in other editors. > Recommended newbie-friendly C mailing lists ? slug-chat? > Anything else I should study to do this properly ? Probably, but most of it is just practice. HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DVB card suggestions [Was: Digital TV Tuner Cards On Linux?]
On Mon, 2005-02-14 at 15:42 +1100, O Plameras wrote: > James Gregory wrote: > >Be careful with this. I've got one, and it's non-trivial to get going > >(well, it's non-trivial to get going if you want to keep the guarantees > >your package manager gives you about the authenticity of the code on > >your system. It's apparently straightforward otherwise). I did try to > >make a set of packages to make it work, but it uses a different version > >of the v4l2 code than what is in the mainline kernel (at least as of > >2.6.9. I'm not sure about now). > > > > This is fixed by running 'DVB-Build.sh remove-symbols'. This particular > script > reconciles Version Number conflicts. It installs a second, parallel version of v4l2 and effectively disables the one that comes with your kernel. It does so by renaming the symbols provided by the old kernel modules. I guess if that works for you, then that's fine. Having attempted to turn Chris' drivers into a kernel patch, I can tell you that there's some pretty fundamental differences between the two versions. The approach of "let's just clobber some symbols" makes me pretty nervous about the whole thing. If it built cleanly against the upstream version of v4l2 I'd have no problem with it. I realise he's provided a work-around, just be sure you understand what you're doing and make sure you're ok with the risk of stuff breaking now, or in the future. > C Pascoe has tested these drivers on stock kernel 2.6.9. > > I have tested it on 2.6.9-1.667 and 2.6.10-rc3. > > >Unless you're more of a kernel build > >ninja than I am, you won't be able to do a clean build of it and you'll > >need to use the rather dodgy method described on Chris Pascoe's website, > >which involves mangling the module dependencies file. > > On Fedora Core 2 or 3 the default kernel has all the pre-requisite > requirements > for 'CONFIGS'. So, no need to re-compile the kernel. Well, as previously established, it has all the pre-requisites except the right version of v4l2. Hope that clears up what I meant. James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DVB card suggestions [Was: Digital TV Tuner Cards On Linux?]
On Mon, 2005-02-14 at 14:23 +1100, Michael Fox wrote: > Appreciated. > > I might get the divico fusion one mentioned around here... and the > price might not be as cheap as the Twinhan, but hey.. if it works I > don't mind if its a little more. Be careful with this. I've got one, and it's non-trivial to get going (well, it's non-trivial to get going if you want to keep the guarantees your package manager gives you about the authenticity of the code on your system. It's apparently straightforward otherwise). I did try to make a set of packages to make it work, but it uses a different version of the v4l2 code than what is in the mainline kernel (at least as of 2.6.9. I'm not sure about now). Unless you're more of a kernel build ninja than I am, you won't be able to do a clean build of it and you'll need to use the rather dodgy method described on Chris Pascoe's website, which involves mangling the module dependencies file. I'm running mine under Windows XP until such time as upstream uses the same version of v4l2; Seems like too much bother otherwise (to me). HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] uploading directory tree recursively
On Sun, 2005-02-13 at 23:13 +1030, [EMAIL PROTECTED] wrote: > Any tips on how to upload this 4 GIG of (self-composed and therefore valued) > music? I'd be using rsync for this. If your other linux box has rsync installed and it's running ssh then this should be a snap. Something like this: rsync -e "ssh -c blowfish" -avz my_music_dir/ '[EMAIL PROTECTED]:/home/me/other_music_dir/' Note the trailing slashes on both directory names. The manpage will give you the details, but it operates slightly differently without the slashes. The '-c blowfish' part just makes it go faster. You could quite happily substitue '-e ssh' with no ill side-affects. HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Visual Studio builds on Linux ?
On Thu, 2005-02-10 at 14:27 +1100, Jason Rennie wrote: > Hi all, > > We have a problem at work. We have developed some code under M$ Visual > Studio 2003 (bleech) and we need to get it to compile under linux (and > solaris and a couple of others). So this is just "normal" C, with no dependencies on the Windows API? Or is this code that requires a substantial chunk of Windows to run? When you say you need it to compile on Linux, is that to imply that you need it to *run* on Linux? If it's just compiling, then I heard a rumour that the "express" version (the compiler, without the IDE) of VS worked on Wine. If it's a case of turning your Windows app into a Linux app, winebuild might be the place to start. It generates a Makefile that uses wineg++ et al to build and link your program against winelib. It may or may not be able to do what you want, but it's worth a look. It's part of the standard wine tree. HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] diffstat, Ctrl-R, pushd/popd
Hey, this is a fun game! On Wed, 2005-02-09 at 22:33 +1100, Angus Lees wrote: > At Wed, 9 Feb 2005 09:42:31 +1100, Ben Leslie wrote: > > 1/ Ctrl-R history searching > > > > When using the shell you pretty quickly work out that pressing up will > > search > > backwards through you history, however it tooks me ages to find out that > > you > > could search backs through the history by typing Ctrl-R and a search string. > > One of my favourite is M-. (thats alt or maybe windows key and > full-stop) I've put this: bind "\C-b":vi-bWord bind "\C-f":vi-fWord in my .bashrc to let me use ctrl-f/b to move forward and backwards words. It was recently pointed out to me though that you can actually do this with alt-f/b without any .bashrc modification at all. Also, this: alias screen='screen -e^\`\`' makes the 'screen key' ctrl-~, which means I don't go crazy every time I try to go to the start of a line with ctrl-a in a screen session. Keep this stuff coming. It makes my life easier! James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Two concurrent Gnome Desktops?
On Mon, 2005-02-07 at 09:24 +1100, Peter Rundle wrote: > I also never managed to solve the sound lock out problem. My system is > running > esd, and in theory esd sould manage/mix multiple audio out requests, And it does. But it won't open up access to your sound card for all users unless you ask it to. If you tell esd to be 'public' and listen on a tcp socket, this problem will go away. By default, however it listens on a unix domain socket, which essentially gets file-like permissions that are probably blocking out your other session. The other way to do this is to tell esd to close the connection to the audio device when it's not using it. Then, when you switch users the device will be unlocked and it will work like you'd expect (similarly, the first esd will be able to re-open the device after control has been relinquished by the second one). The mechanism setting the permissions of your audio device is almost certainly pam_console. If it bothers you, you can turn it off and fall-back to the "old" way of doing things by putting /dev/dsp in the audio group etc. And please, if you've got issues with GNOME (or anything else), take them to a sensible forum. Like their bugzilla. James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Exporting applications from windows?
On Sat, 2005-02-05 at 19:08 +1100, Karl Bowden wrote: > Is it possible to have an application running remotely on a Windows > machine but appareing locally on a linux machine? > I have been playing arround with rdesktop and a win2k machine. But I > have only been able to export a complete desktop. > What I am looking for is something more similar to the way X > applications can be forwarded over a ssh session. > If I am missing something obvious please clip me arround the ears, but a > feature like this would much ease the transition to linux at my place of > work. In *theory* you should be able to take the GDI implementation from Wine, drop that into the directory that your application is installed (the Windows equivalent of LD_PRELOAD), set an environment variable, make an offering to the god of dynamic linking, and have it turn the GDI API calls into X data that you can display on any X server. In practice though, I don't think anyone's got that to work successfully. It's worth digging around for, or similarly for people using other GDI replacements to do the same thing. What Windows apps do you need? Maybe using Wine directly would be an option? HTH, James. -- "There is no I in TEAM but there is an i in Ninja" -- http://www.ninjaburger.com/sekrit/ signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Controlling address space layout
On Tue, 2005-01-18 at 02:45 +, Dave Airlie wrote: > > dynamically linked program in linux? Alternatively, are dynamically > > linked libraries always going to be in the same place? > > once upon a time maybe.. with all these prelink and random patches > floating about it really depends on what kernel you are running.. That functionality can be turned off, I don't remember how, but it is definitely optional. I'd suggest taking a look through the wine-devel archives, this seems to come up every time there's a new RedHat release. It's probably also worth taking a look at the code that bootstraps wine; it does some of this stuff. Unfortunately I don't know enough about it to say if it's what you're after or not. HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Visual image difference command line tool
On Tue, 2005-01-18 at 08:38 +1100, [EMAIL PROTECTED] wrote: > If this doesn't exist as an obscure technique in the gimp or some such I'd > be surprised. The standard technique is to subtract one image from the other. Gimp's layers functionality can do this, and I'm pretty sure you can do it with ImageMagick too. Check out the convert manpage; I don't remember the details. HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] libgd dependencies prob
On Thu, 2005-01-13 at 15:20 +1100, Voytek wrote: > where should I go from here ? > > - leave the current mrtg ? > - force install ? > - go and have a cup of tea instead ? I'd suggest letting the computer work through the dependencies for you. Which distribution is this? If it's RedHat, I'd suggest looking into yum or apt-rpm. Happy to help out if you hit issues with either of them. James. -- "being bad ass is encouraged. being a black hat hacker that can take shit down and f shit up is cool and hot and sexy. being a grey hat and pointing out the flaws is hot too...white hats are even hot if they do it up right. the more you learn, the more bad ass you are at what you can do... and just maybe girls will think it's hot and want to hook up with you." -- http://www.catschwartz.com/archives/2004_10.html signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Visual Basic port
On Fri, 2005-01-14 at 16:33 +1100, Erik de Castro Lopo wrote: > Looks like you need a GUI and the available compiled languages > with GUI-ability for Linux (that I know of anyway) are: > [...] There's also C# with its GTK bindings (or SWF I guess; you'd get Windows compatibility for free then). I would have suggested it earlier, but I honestly don't know if it's any good. Has anyone had any experience with it? Gotta learn C# some day soon. re: clicky designers -- glade and libglade will give you the same ability to design UIs with any of the languages mentioned thus far. You'd still be lacking an IDE to glue it all together though; that's one way in which Qt-Designer is a nice choice. I believe Anjuta does a similar thing for other languages (I don't use it though, so I don't know which ones). If you're after a low-effort port, I'd say gambas is your most likely candidate. Alas I've not used it either. HTH, j. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Postgresql case-insensitive mishmash brain twister.
On Thu, 2005-01-06 at 13:08 +1100, [EMAIL PROTECTED] wrote: > Nope but a nice try... Same SQL query both servers. Different result. > > Case insensitive searching is what I'm after as a default. I'm sorry, that query is far too white-space challenged for me to be able to parse it, so I can't be much help in correcting your query. I will say that if you want to do a case insensitive string match, 'ilike' is the operator you're looking for. If you're seeing different string comparison behaviour between machines, I'd be checking the encoding type of the data. Unfortunately I can't remember how to do that, so you'll need to look it up. It's done on a database level, IIRC. HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Webcams - which to buy?
On Sun, 2005-01-02 at 21:25 +1100, Elliott-Brennan wrote: > Apologies for the delay in saying thanks James. > > Dean has mentioned that I should look at a capture card instead. I think > I'm definitely going to need some help in the coming months! :) If you choose the right one, a capture card is a snap to setup. You'll need an extra camera though, which is a bit of a pain. I just remembered though that my previous place of employment bought a few of the Creative webcams. I don't remember which particular models, but they worked quite well. Capture cards are certainly nifty, but if you don't need that functionality they're probably a little overkill. Also, I just did a google search for "webcam linux" and found these two relatively pertinent links: http://alpha.dyndns.org/ov511/cameras.html http://www.linux.com/howtos/Webcam-HOWTO/devices.shtml The first one lists model names and numbers and tells you about driver support. You could print that out and go to your local electronics store and try your luck. A strategy which I found worked well when I was looking for USB MIDI widgets. HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] URPMI's not working
On Fri, 2004-12-24 at 06:59 +1100, Scott Clements wrote: > Hey Sluggers, > > > > I am very new to Linux and I am probably making the mistake, I am > using Mandrake 10.1 Official and trying to add new media to my URPMI > conf. However when I try to addmedia from sever different mirrors > around the globe I get the following error messages after my input. > > > > As root I type: > > > > urpmi.addmedia main > ftp://mirror.pacific.net.au/linux/Mandrakelinux/official/10.1/i586/media/main/ > with /media_info/hdlist.cz That's ok. I generally don't have a slash before the media_info, but I think urpmi can work it out either way. > and I get > > > > retrieving source hdlist (or synthesis) of "main"... > > > ftp://mirror.pacific.net.au/linux/Mandrakelinux/official/10.1/i586/media/main/media_info/hdlist.cz > > ...retrieving failed: curl: (7) Failed to connect to 0.0.0.10: Invalid > argument So the real problem is here. Your computer has resolved mirror.pacific.net.au to 0.0.0.10, which is an unusual thing for it to do. Does your ISP require you to use a proxy? If so, try setting that with the http_proxy environment variable, so it would be something like this: http_proxy=http://www-cache:3128/ urpmi.addmedia main http://mirror.pacific.net.au/linux/Mandrakelinux/official/10.1/i586/media/main/ with /media_info/hdlist.cz or use ftp_proxy for ftp mirrors (notice that I changed your ftp to http there). Even if proxies are required, that's still weird. Try adding --wget to your command line; your curl might be borked. If all that fails then tell us a bit about how your network fits together and I'll see if I can offer more helpful advice. James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Webcams - which to buy?
On Thu, 2004-12-30 at 23:23 +1100, Elliott-Brennan wrote: > Hi, > A relative has just moved O/S and I'm looking at buying a webcam (I have > Mandrake 10 Official with the latest kernel). I gather that the Logitech > Webcam should work. I'm wondering if anyone is running one (or another > one) and what their experience has been (good, bad, indifferent). I bought a logitech one a while ago and wasn't able to make it work. The current generation of Logitech cams have got a different chipset to the ones that were quite common a few years ago. The new ones are simply too much work to be worth it (but if you really want to do it, someone *has* written a mostly working driver for them. It's just a pain to get going). Sorry I don't have any more helpful info there. Maybe the guys at elx have something you can use. Also, if you're intending to do video-conferencing -- I bought a Logitech USB headset to do the same thing. The headset itself worked quite well with linux and was almost no effort to get it up and running. For some reason though, it had real issues when I tried to use it with Skype. Anyway, if you've got a working sound-card I'd recommend trying out a non-USB headset. See if that works for you first. HTH, James. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Windows on Linux
On Tue, Dec 14, 2004 at 01:27:54PM +1100, Alan L Tyree wrote: > If this is a serious question Yeah, totally. I see WINE as a very important piece of software, and it's always good to get more people using it. > -- I don't know. It is packaged up with Butterworths (Lexus/Nexus) > Australian Legal Research Library. I was told by Butterworths people > that they were instructed to use it instead of a simple HTML because > their parent company (at some time or another) had acquired ownership > of Folio. Ok. So there's no trial version (or similar) available, then? > You can get it to run on Windows XP, but it requires that the old > installation be removed (at least a major part of it) and then the new > one installed. Honest, that is the instructions that they send with it. > This has to be done with each monthly update. Hmm. Sounds like the sort of thing you'd want to automate. That's basically the procedure that modern package management systems implement, but they automate the process. > That actually worked for a while for me with Wine, but > when I switched to Debian Testing from RH8, it choked up and I haven't > had the time to fool around with it too much. Ahhh. There was some discussion about Debian's WINE packages on the list recently. My recollection is that the Debian packages had some problems. The packages cited in this post: http://www.winehq.com/hypermail/wine-devel/2004/12/0004.html might work better for you. > I will have another look at it in the next day or so. What kind of > debugging information might be useful? A description of what you do, in what way it does not work as expected and a copy of any messages that get written out to the console would be a fantastic starting point. Also tell me if you're going from a "clean" .wine directory and anything else about the state of the machine when you run it that might be relevant. James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Windows on Linux
On Tue, Dec 14, 2004 at 12:57:37PM +1100, Alan L Tyree wrote: > > > Folio Views is so bad that > it doesn't even work properly on new versions of Windows. It's just that > it is the only piece of Windows software that I have any reason to run > and (not surprisingly) it chokes Wine unless everything is just so. Awesome. Where can I get a copy of it? j. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Windows on Linux
> > I've been watching wine for years and I feel that it's just never > > going to make it. At one stage I managed to get Word and Excel working > > under it but then they changed to the new system with the dosdevices > > directory and I haven't been able to get them working since. Also even > > though Excel was "working" I couldn't get the Visual basic stuff to > > work, which is the only reason I'd use Excel, to deal with a spread > > sheet with a VB macro in it. However I hear that codeweavers > > cross-over office seems to be more reliable (don't know what the > > codeweavers secret is as it's based on wine). The Codeweavers guys are doing some awesome work, and yeah, I always recommend that people use one of the commercially backed versions of WINE atm. The way it works (AFAIK) is that Codeweavers keep their own tree, and work on that in order to get support for the apps they've got targetted in their next release. After the release happens, they submit their patches back to mainline WINE and AJ goes off applying patches (which is a pretty time consuming process). So mainline WINE is generally about a year behind Codeweavers work, and at the moment it's also lacking the nice interface for getting stuff installed[*]. There's a lot of structural changes going on at the moment that will ultimately result in a better codebase. They're not without migration issues though. WINE also hits issues every time the threading model changes in the kernel. We've seen a lot of such changes in the last year or so, but it looks like things have settled down since then. Mer, so the reason I'm writing is twofold: 1. Tell us about the issues and I reckon we can get them solved. 2. I think they will get there, and I think the changes that are happening now are a vital step on the road to "there". > > I have to say that I feel the same about Wine. I run a legal application > that uses Folio Views (Terrible software!) and it is a real pain to get > in running every time that Wine changes. Since the latest dosdevices > change I haven't been able to get it going at all. There's a lot of debate about issues like this on the wine-devel mailing list. There's also versioning issues in the .so that does most of the work, which they're trying to form sensible policy on. They are, in my opinion, doing the right thing. They're trying to move away from the massively abused configuration file that has been with WINE since its inception (as far as I know). In the long term, you can expect this to make it easier to get things to work, rather than harder; but there are definitely migration issues. The only thing that changed in the dosdevices patch was that the drive definitions were moved out of the config file and turned into (easier to manage) symlinks. If you can tell us more about the problems you're seeing, I'll try to help you get it working again. James. [*] Just on this: I started hacking on some stuff to let you install Windows software with RPM a while ago, but I never got enough time to make it really useful. If you are interested in this, let me know what you need to get running on WINE and I'll tack it onto my todo list. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] jfs on rh73 ?
On Mon, Dec 13, 2004 at 02:06:52PM +1100, O Plameras wrote: > fdisk will see it as 'type 83 - linux' not as a filesystem. > > There is a way to ascertain what filesystem is the partition but can't > remember. > > What I'd do is do a series of mount command and see the results. If there's any doubt about what sort of file-system it is, mount with -t auto; that will try all the available filesystems. I thought we were pretty sure this was JFS though. > >>60 seconds later, the prompt still hasn't come back..? > > > >Ok, that's a problem. I'd guess you've got a broken filesystem. Have you > >got a backup of the disk? If not, I'd suggest taking an image of it and > >then using jfs_fsck to see if it can find anything. > > That means the filesystem in the partition is not 'jfs'. Well, it doesn't necessarily mean that. The kernel could be blocking for any one (or more, I suppose) of a number of reasons. What makes you think that this block is the wrong filesystem selection? James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] jfs on rh73 ?
On Mon, Dec 13, 2004 at 01:01:15PM +1100, Voytek wrote: > > > > > > If all these indicate JFS is loaded, then proceed and mount your > > JFS partition. > > what do these need to autoload or survive reboot ? I was under the impression that mount would attempt to load the right module when you attempt to load a filesystem of that type. I could be wrong on that, but I've never had to do anything to mount exotic filesystems. > but, 'fdisk' still doesn't see it, should it ? Uhh, there's no such thing as a 'JFS partition'. It's just a 'normal' partition with a JFS filesystem on it. In much the same way that my XFS partitions just show up as 'Linux'. This isn't a problem. > when I try this: > # mount -t jfs /dev/sdb /mnt/hd > > 60 seconds later, the prompt still hasn't come back..? Ok, that's a problem. I'd guess you've got a broken filesystem. Have you got a backup of the disk? If not, I'd suggest taking an image of it and then using jfs_fsck to see if it can find anything. HTH, James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: Goals
On Sun, Dec 05, 2004 at 03:16:44PM +1100, Rod Butcher wrote: > I've thought about the "rpm hell" bit before. I notice that some > commercial apps (like Opera) seem to bundle everything required into a > single tarball and say "just install this". Do they achieve this by > statically linking in all the required libraries, Yep. That's what they do. > and is this what is done in Windows ? My recollection of most Windows installation procedures is that they ship with all the dlls necessary for that application to run. You can see evidence of this if you use the windows packaging system; it will tell you about orphaned libraries and so forth. > I feel that it IS asking too much of non-technical users to "install x > and y before installing z", especially when installation of x.0.2 > gives the message "conflicts with x.0.1" or requires installation of v > and w. We techos tend to forgot the little glitches that occur which > spell badhair day to a non-technical user. There IS QA in the Free / > Opensource world. But I would agree that the Linux software install > process, logical as it may seem to a computer professional, is not up > to the needs of the non-technical user. I personally see all of these things as solved problems. I can't remember the last time it's been more than a few mouse clicks to install a piece of software I wanted. That said, one of the questions I dread from new Linux users is "how do I install X?". I don't dread it because it's a difficult question to answer, but because there's not one answer. The fact that a third party (your distributor) has taken on the responsibility of ensuring that this new software works for you is both a strength and a weakness of Linux. The fact that there is competition in the area of software integration means that the current generation of Linux distributions work really nicely with all the software you're likely to install on them. The downside is that the answer to questions like "I downloaded this file and I double click on it and I just ended up with more files, help!?" always ends up being a dialog rather than "double click it and cross your fingers". People expect software provinders to be installation providers, which I suppose is not unreasonable, it's just unfortunate. I think that is the problem that needs to be addressed. James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: Photoshop in Linux WAS [SLUG] Newbie -when is next slug meet?
On Tue, Nov 30, 2004 at 11:26:23PM +1100, Rod Butcher wrote: > > Whatever people seem to have a chunk of experience > > with first more or less defines what they think is the best interface. > I agree, hadn't thought of that.. I cut my teeth on Waterloo script on > mainframes, and struggled to adjust to M$Word's graphical wysywig idea - > I still feel comfortable with markup languages like HTML. > But meanwhile back at the farm, my .png image's background remains > resolutely opaque. It consists of black text on white bg. As instructed > I selected layer, new layer, transparent. No dice. This can't be the > full story. How/where do I tell it that "listen, Gimp, whenever you see > white space in this here image, I want it to be transparent". ? From > memory, in PS I specified an index value for transparency. In general PNG doesn't work that way. It can do indexed transparency, but the advantage of using it comes in the fact that it has an alpha layer. The best way to use it is as someone else here pointed out: Make a new image with a transparent background and just don't put stuff where you want it transparent. IE has some issues displaying PNGs with alpha layers, but there are workarounds. Can we see some of these images with the resolutely opaque backgrounds? James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: Photoshop in Linux WAS [SLUG] Newbie -when is next slug meet?
On Tue, Nov 30, 2004 at 06:40:45PM +1100, Craige McWhirter wrote: > The difference in configurations and layouts between > 95/98/xp/NT/2000/2003 is incredibly bloody annoying. If I had one > constructive thing to say to Microsoft it would be "*STOP MOVING SHIT!*" > especially as I'm dealing with all those listed OS's :) Stuff moves around between versions on Linux systems as well. Upgrading Postgresql is an example that leaps to mind. Why do you find the UI revisions in the Windows operating systems more frustrating? As for myself, my impressions have been that I have a lot of difficulty getting my work done on Windows machines, but I have a lot of trouble explaining to others how to do things on Linux machines. I've got gripes about particular issues on both systems, but I might rant about that on slug-chat rather than wasting everyone's time here. James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] C++ newbie terrors - the program that won't compile
On Thu, Nov 25, 2004 at 08:30:53AM +1100, Nicholas Tomlin wrote:
> Hell sluggers,
Right on.
> ** error message:
> [...]
>
> ** I'm not worried about the antiquated header, The syntax is the issue. I've
> tried all sorts of mods to the syntax and cannot get it to go away, the
> program is a 'cut and paste' straight out of a text book, but it does not
> work! (while student is programming; writeln f..k).
The header isn't antiquated, you don't put the .h on C++ includes. The
.h variant doesn't use namespaces, if you do
#include
your header warning will go away. You'll then get a new error, because
you haven't either specified the std namespace for all the std stuff
you're using, nor specified that you want the namespace imported
(assuming you're running a vaguely recent version of g++). So, you can
either do this kinda thing:
std::cout << ...
or put this at the top of the file somewhere:
using namespace std;
Have you intentionally dropped the arguments from main()? Anyway, you
probably want something like this:
#include
using namespace std;
int main(int argc, char *argv[]) {
int x = 5;
int y = 7;
cout << x + y << " " << x * y << endl;;
return 0;
}
> I assume this is normal practice, to publish inaccurate and unusable
> programs so some schmuck spends hours sorting it out...
The error you were getting was a missing <<. I'd double check that. The
issue with your header files is probably the age of your references.
There was a long time where g++ would just ignore namespaces and there's
a lot of reference material that still ignores this fact.
HTH,
James.
--
"Now, there are no problems only opportunities. However, this seemed to be an
insurmountable opportunity."
- http://www.surfare.net/~toolman/temp/diagram.html
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] guessing compile variables from binary rpm
On Fri, Nov 19, 2004 at 12:29:46PM +1100, Rod Butcher wrote: > Sluggers, can anybody tell me how to find out what libs, compiler > options etc went into building a binary in an rpm ? Specific example is, > I'm trying to find out how Mandrake compiles audacity to get it to work > with wxwindows and gnome themes, I can't. Given just the binary RPMs I don't know of a general way. Some code will tell you how it was built (PHP is an example that springs to mind). In general, the best thing to do is to inspect Cooker CVS and take a look at the spec file that created the RPM. There's often numerous patches as well. In the specific case you've mentioned I'd suggest that they've done something to wxWindows build flags rather than audacity though (if they've done anything. It just sounds like GTK is doing its GTK thing). James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Mandrake non-usb scanner permissions woes
On Tue, Nov 16, 2004 at 09:30:46AM +1100, William Chivers wrote: > Searching for this problem in Slug and elsewhere locates plenty of > information about this problem and USB scanners, but nothing for > parallel scanners. IIRC there's also some library thing that does this and doesn't need a specific /dev entry at all. Someone else will probably know more about it than I do though. > Should I locate the offending file and change its permissions? If so, > what is the file? Searching file:/ for "scanner" did not find anything > relevant. There is no scanner entry in /dev. Or should I somehow add > users to the scaner group or vice-versa? I suspect the answer is to alter the pam_console config to change these permissions for you when you login. Take a look at /etc/security/console.perms. The format is relatively clear. I couldn't tell you what the name of the device is though; I've never used a parallel port scanner. > > I had this problem with Mandrake 9.1 and 9.2 and just ran the scanner as > root, thinking that Mandrake might eventually sort this out... I suspect it's a pretty uncommon configuration these days; it probably doesn't get much testing. It might be worth your while to file a bug about it. HTH, James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] group membership
On Tue, Nov 16, 2004 at 11:58:24AM +1100, [EMAIL PROTECTED] wrote: > On Tue, Nov 16, 2004 at 10:21:00AM +1100, Phil Scarratt wrote: > > [ .. ] it wasn't as simple to answer as I thought. The question: can a > > group be > > listed as a member of another group? > > No they can't (you're right) > > It's one thing that w__ has over linux/unix. > That said, there are number of ways to get the > effect that you want -- depending on exactly you > want. > > Restrictive parent dir perms, ACLs ... Is it possible to do it with LDAP as your user database? My recollection is that the various modules that make the pwent stuff work with LDAP allow you to specify arbitrary database queries to determine the answer to questions like "am I in this group?". Is it possible to mangle those queries to an extent that multiple levels of grouping could be encapsulated in LDAP group objects and through judicious use of magic it could be made to operate as if the single group relationship was multi-layered? Not perfect by any means, but it'd probably be passable for file permissions and similar. James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Most valuable free/OSS software that doesn't exist?
On Mon, Nov 15, 2004 at 12:28:47PM +1100, Jamie Wilkinson wrote: > This one time, at band camp, O Plameras wrote: > >1. Given a specific server in the network, can remember exactly previous > >configurations > >for 3 generations (parent, grand parent, and great grand parent) along > >with each file in > >each of those generations. > >2. Given a specific server in the network, SysAdmin must be able to > >revert back to that > >generation along with each file in that generation through an operation > >or a set of > >operations availble in the systems. > > those last three are tough problems, but should be theoretically possible > with either full dumps of the system before and after changes, or using some > sort of delta algorithm. I think isconf does somethign close to this though > -- at least in the forward direction. > > I don't have that requirement for rollbacks though, I test on a small set of > machines before rolling out new changes, changes are always small and > manageable. It's not a perfect method but it works. Current versions of RPM allow you to rollback to previous versions of software. I've never had a need to do it though, so I can't say if it works as advertised or not. Your cross-platform requirement makes this moot of course, but it gets you a little closer to a solution... It does mean that you need to trust RPM for all your software management, which is a decision you need to make. > > >3. Must be able to effect change of Operating Systems from older version > >to a newer one. > >Because of item 2. above the reverse will be possible. > > cfengine and isconf are two tools that spring to mind for this task. I'm > using cfengine on a production network of 30ish machines with good results. I've also heard of people storing all of /etc in version control for this purpose. In my opinion it would be unnecessary if you kept your cfengine stuff in a source control system, but it would give you that absolute confidence that you could roll back and forward, even if cfengine failed. > >5. Must be able to effect change of Applications (delete, add, or a > >combination) for a given > >server in the network. > > cfengine again; servers on our network that are part of the, say, ftp server > class, have their installed packages list checked, and packages installed. > I don't do the reverse, because whilst I can guarantee that I need to > install and configure a service on a machine, I can't guarantee that a > machine that isn't part of a class isn't supposed to have that service > configured. At least, not yet (lots of auditing involved) RHEL also lets you do centralised upgrades. Of course, I'm a total amateur at this stuff, so I couldn't tell you if it lets you do the reverse operation or lets you remove stuff. Jamie? Actually, I know Mandrake has a 'parallel installer' application that ships with it; it uses SSH to install upgrades on a small army of Mandrake machines. Now, I'm not recommending Mandrake for production server use here, but I'd be surprised if other distros didn't have equivalent functionality. Does anyone know if this is the case? > >6. Etc. that is to do with any management of changes in the server of > >the network. > > Um. Probably "ssh" is the tool you're looking for there :) HTH, James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: Preventing attacks
On Tue, Nov 09, 2004 at 03:31:50PM +1100, Toliman wrote: > and it is 'relatively' secure, in that it would hopefully > take a p4 a few hours to brute force... more likely in minutes. How long is 'a few hours'? I didn't think things were that dire. Are you talking about a straight brute force or some kind of known-plaintext attack or what? James. -- "Now, there are no problems only opportunities. However, this seemed to be an insurmountable opportunity." - http://www.surfare.net/~toolman/temp/diagram.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
