phunni

What's the difference between postfix and procmail? Some people seeme to be using them together and others seem to be using them in an either/or kind of situation...

I just bought a shiny new mouse - an mx100 from logitech. Seems really great so far - although I am using a very generic setup for X (xorg). I've googled around for ways to get the best out of the mouse and not found any really clear instructions. A few plaes have mentioned evdev (which I have available as a module), but I haven't been able to do anything with it. Is anyone able to point me in the right direction at all?

Thanks guys - this at leat gives me confidence to get started - knowing that I'm not installing things which I don't need. Speaking of which - isn't clamav a virus checker? Do I need one of those. While I do have a windows workstation on the network, it never picks up email - that's only ever done on my linux desktop Someone mentioed me today that apache might be able to deliver mail to my desktop. I already have it on the server, simply to serve the "you can't access that page - it's naughty!" kind of stuff. Is this a red herring - expecially since I'm going to have to install procmail anyway...

quick kernel question: - If I compile it direcctly into the kernel and not as a module and then run make clean && make bzImage with out the normal modules modules_install amd I right in thinking that my modules will not be overwritten and that I don't need to recompile them?

I checkd in the kernel logs (why didn't I think of that earlier?) and found: Dec 19 13:56:02 kirstine kernel: Unable to load NLS charset cp437 Dec 19 13:56:02 kirstine kernel: FAT: codepage cp437 not found So I reckon I need to take a look at that :D Thanks richarddoe - I'll let you know how it all works out...

I have a server which currently firewalls/proxies for the computers at home. I want to set it up so that my local mail clients can pick up their mail from it via smtp and have it invoke spamassassin on the mail before it becomes available to the local client. I've seen a few recipies ( including one here somewhere) about this which talk about procmail/postfix and that kind of thing. They also talk about delivering to mbox files which is where my confusion starts. I'm assuming that the expectation there is that the mail will be colected from the mbox file locally and not via smtp - am I correct in thinking that? Also are either procmail or postfix (or anything else) suitable as smtp servers? I can't see any reason to send mail via this server as my main concern is simply getting spam filtering of my local desktop and onto the server. Thoughts please...

Never mind - I figured it out. I had to set the following in squid.conf: httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on Seems to be working now. One more question though - if I set my firewall to pass ftp requests to squid, wil lthat just work or will squid need further configuration..?

I've just installed squid and it seems to be working. Ideally, Iwant transparent proxying so I tried to set my firewall to forward everything to the squid port. This resulted in almost everything I requested being refused with an "invalid URL" message. If I manually specify the proxy - it seems to work, but I don't wan't to do it that way if I can help it... plus if I run: squidclient http://mandrakeusers.org > test I am told that the connection to localhost:3128 is refused... This might be a problem as I eventually want to run dansguardian as a filter so it'll be connecting to the proxy from the same machine... Any suggestions as to how to configure this? I've specified the visible and unique hostnames as "server" since that it the computers network name...

I know it's not another linux distro - but it's still UNIX so... The servers at work have gone down 3 times in the last fortnight and the big wigs are getting annoyed (understandably). It's been discovered that the last two times it happened a core OS file disappeared just before everything went belly up. It has therefore occured to one or two people that we might be being hacked (probably some cheeky computer science student... ) and the question has arisen as to what might be a suitable intrusion detection system. Tripwire has been suggested - but apparently it's now commercial (although why that should matter...) so I'm trying to hunt down a potential solution. Any suggestions?

any ideas at all...?

Odd, I thought I had already replied to this - but I can do it again... After recompiling (gave me a good excuse to go to 2.6.9 as well...) I have everything that iphitus has - except that ntfs is a module not compiled in. I don;t think that's relevant though as none of the media I'm having trouble with are ntfs - and I did check just in case!

To play devil's advocate a little bit here - he does, if you read the rest of the article, suggest that you should use Linux if it's the best solution for a particular context. Obviously he probably doesn't think that's likely to happen that often because he's biased. But then, I'd expect the CEO of any company to be biased in favour of their products...

I believe that Scotty is referring to a file manager called gentoo rather than anything to do with the distro of the same name... Sorry - can't actually help with your problem, but just thought I'd bring some clarity. A couple of things though - since you ask for file manager suggestions I would suggest mc (midnight commander). Also do you have any other error messages? There isn't much to work with in the output you've given

Thanks - that's sorted it :D

I just installed cygwin on my work machine and through a little tweaking managed to destroy my cygwin.bat file Can someone please send me theirs? my username @gmail.com is as good an email as any...

I'm using a custom vanilla kernel from kernel.org. I have vfat available as a module, but it doesn't make any difference when it's loaded (obviously it won't work when it's not loaded!)

result of fdisk -l: Disk /dev/sda: 249 MB, 249823232 bytes 16 heads, 32 sectors/track, 953 cylinders Units = cylinders of 512 * 512 = 262144 bytes Device Boot Start End Blocks Id System /dev/sda1 * 1 952 243630+ 6 FAT16 This only works after I try and mount it from /dev/sd/<some weird name> I am still told that it can't be mounted - same with my floppy, I get the same error even if I specify msdos

from a brief scan of the man page - possibly. Is this going to involve the use of regular expressions?

I have a project that a shell script is going to deploy on a server for me ni both a testing and a staging context. Some paths, db connection strings etc, will be different for my development environment and boththe staging and testing environments. How can I, in a shell script (for bash) do a simple find and replace for text? Is it possible?

I have tried that - in fact the memory stick isn't in /etc/fstab so I have to do that to mount it. I only use the fstab entry for my floppy which returns the same error.

no idea about the unhide - I'll remove it, and see what happens. I haven't changed any hardware - but I recently did an OS reinstall and it's been broken ever since. My clie memory stick has always apepared as /dev/sda1 - although only after I tried to mount it as /dev/sd/<some weird filename> and failed: at that point sda1 appeared

Every time I try to mount either my floppy drive or my sony clie memory stick as vfat (it has worked in the past) I get the standard error: mount: wrong fs type, bad option, bad superblock on /dev/floppy/0, or too many mounted file systems At first I wondered if vfat was simply not supported by my kernel, but I looked in the config and it's there as a module: vfat. I loaded the module and I still get the same error. Any ideas why this might be? Edit: my fstab for the floppy drive: /dev/floppy/0 /mnt/floppy vfat user,noauto,unhide 0 0

turns out I am using sh - how can I change this to bash?

OK - so Solaris isn't strictly another Linux distro and nor is the windows box I work from - but you get what I'm trying to say... I'm currently using putty to remotely connect to our Solaris server. It all seems very comfortable and Unix like - just like using the clie at home on my linux box. Except that I can't use tab completion. Is this possible using putty? Does the fact that I'm connecting to Solaris casue any complications?

In Word, you can control-click on the contents and go there. In the same document in OO this doesn't seem to work - nor is there any obvious way of using any links that might be there...