November 2001 - WLUG - wlug.mailman3.com

Raw IO Howto

by Skip Gaede

Folks, I'd like to be able to queue read/write request to my Zip drive and bypass the buffering that goes on in the kernel. From what I've read, you can do this by binding the raw device (/dev/raw) to /dev/hdc1 for example. When I do that, I get a message that /dev/raw/raw1 has ben bound to char major 22 1. When I try open("/dev/raw/raw1",O_RDWR|O_CREAT,S_IREAD|S_IWRITE); it fails. Can someone tell me how to do this please? Thanks, Skip

Nov. 9, 2001

Star Wars: ASCII Art-oo

by jeff

http://www.joereiss.net/starwars/ascii/ enjoy... -jeff -- Q: What do you get when you cross a mobster with an international standard? A: You get someone who makes you an offer that you can't understand!

Nov. 9, 2001

you are invited

by jeff

ok...so it's like 2 am and i'm having a typical bout of keyboard induced insomnia... i would like to invite all the members of wlug to a little meeting that mwblug will be having tonight (wednesday) at the tin alley grill which is a little pub located right on route 9 in framingham. it's open-ended stuff and it's mostly to get new and old peeps acquainted and talk about the usual...over a few pops and lotsa larfs... please check out site www.linuxbusca.com for more details and we hope to see maybe one or twenty three of you there... :) peace and happy hacking... -jeff p.s. meeting will be 'round 7 pm... look for a scary bald guy with tattoos...that's me... :)

Nov. 7, 2001

Re: [Wlug] Question

by Keith Wright

> From: Theo Van Dinter <felicity(a)kluge.net> > > On Tue, Nov 06, 2001 at 12:34:37AM -0500, Keith Wright wrote: > > DNS has nothing to do with Netscape. Start by trying to ping. > > ??? Netscape has to use something to resolve hostnames... Of course. What I meant was that dorking around with the Netscape menus is not the way to fix it. Netscape also needs a graphic display, but if someone asks: ``I am trying to get Netscape to work, but I am stuck in text mode'', I will say graphics mode has nothing to do with Netscape, you must fix the X window system. > A few notes: <cut> Good points. Pierce, let us know when it is fixed, or what new problem comes up. -- -- Keith Wright <kwright(a)free-comp-shop.com> Programmer in Chief, Free Computer Shop <http://www.free-comp-shop.com> --- Food, Shelter, Source code. ---

Nov. 7, 2001

RE: [Wlug] Unique ID

by Peter Nikolla

I would like to thank everyone for there input. Peter N. -----Original Message----- From: wlug-admin(a)mail.wlug.org [mailto:wlug-admin@mail.wlug.org]On Behalf Of Theo Van Dinter Sent: Tuesday, November 06, 2001 3:06 PM To: wlug(a)mail.wlug.org Subject: Re: [Wlug] Unique ID On Tue, Nov 06, 2001 at 02:52:12PM -0500, Bill Mills-Curran wrote: > <opinion> > Also, there are valid reasons for licensing and selling/leasing some > products. [snip] > Although I believe in the open source > model, I don't think it's all-encompassing. Some products just need > income in order to live. > </opinion> Absolutely, open-source software is great but there is plenty of room for commercial stuff out there. Licenses are one thing, locking software to a particular machine is another issue entirely. I have no problem with people charging me money for products that are worth it (heck, I *buy* my RedHat CDs whenever possible), but I really don't like the one license/one machine model. There was some talk a while ago about making open-source licensing code, ala flexlm. Essentially there's a daemon somewhere that says "there are X licenses available", and the application can be installed in a billion different places. However, the licensing code will only allow the application to run in X places simultaneously. It's the "book licensing method" -- you can give the book to anyone you want, but it can only be used by one person at a time. It would also be possible to say "unlimited licenses", and then use the licensing system as an audit system to find out how many users are actually using the application, etc. I think the end of the discussion was that open source and this licensing type of system are contrary to each other. Personally, I think it'd be a fun thing to work on if I had the time. -- Randomly Generated Tagline: There are two types of Linux developers - those who can spell, and those who can't. There is a constant pitched battle between the two. (From one of the post-1.1.54 kernel update messages posted to c.o.l.a) _______________________________________________ Wlug mailing list Wlug(a)mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug

Nov. 6, 2001

Re: [Wlug] Unique ID

by Bill Mills-Curran

On Tue, 6 Nov 2001, Theo Van Dinter wrote: > Date: Tue, 6 Nov 2001 15:05:34 -0500 > From: Theo Van Dinter <felicity(a)kluge.net> > Reply-To: wlug(a)mail.wlug.org > To: wlug(a)mail.wlug.org > Subject: Re: [Wlug] Unique ID > <snip> > There was some talk a while ago about making open-source licensing code, ala > flexlm. <snip> I've worked with flexlm & actually had to port the source. It turns out that the license server needs to be node-locked. I think that all licensing models (except maybe the MicroSoft "everyone contacts the MS big-brother server" model) require node-locking at some level. Flexlm locks the server to a node, which doles out licenses as necessary. Bill

Nov. 6, 2001

Re: [Wlug] Unique ID

by Theo Van Dinter

On Tue, Nov 06, 2001 at 02:52:12PM -0500, Bill Mills-Curran wrote: > <opinion> > Also, there are valid reasons for licensing and selling/leasing some > products. [snip] > Although I believe in the open source > model, I don't think it's all-encompassing. Some products just need > income in order to live. > </opinion> Absolutely, open-source software is great but there is plenty of room for commercial stuff out there. Licenses are one thing, locking software to a particular machine is another issue entirely. I have no problem with people charging me money for products that are worth it (heck, I *buy* my RedHat CDs whenever possible), but I really don't like the one license/one machine model. There was some talk a while ago about making open-source licensing code, ala flexlm. Essentially there's a daemon somewhere that says "there are X licenses available", and the application can be installed in a billion different places. However, the licensing code will only allow the application to run in X places simultaneously. It's the "book licensing method" -- you can give the book to anyone you want, but it can only be used by one person at a time. It would also be possible to say "unlimited licenses", and then use the licensing system as an audit system to find out how many users are actually using the application, etc. I think the end of the discussion was that open source and this licensing type of system are contrary to each other. Personally, I think it'd be a fun thing to work on if I had the time. -- Randomly Generated Tagline: There are two types of Linux developers - those who can spell, and those who can't. There is a constant pitched battle between the two. (From one of the post-1.1.54 kernel update messages posted to c.o.l.a)

Nov. 6, 2001

Re: [Wlug] Unique ID

by Bill Mills-Curran

On Tue, 6 Nov 2001, Theo Van Dinter wrote: > Date: Tue, 6 Nov 2001 09:42:20 -0500 > From: Theo Van Dinter <felicity(a)kluge.net> > Reply-To: wlug(a)mail.wlug.org > To: wlug(a)mail.wlug.org > Subject: Re: [Wlug] Unique ID > > On Mon, Nov 05, 2001 at 08:00:11PM -0500, Peter Nikolla wrote: > > I am tiring to right a locking mechanism for my program. I wanted to find > > something unique on a Linux/Unix system that I could read to lock my code to > > that machine. Does anyone know of a call I can make to get an unique ID > > like a disk ID, or system ID? > > I'll agree with Keith on this one, it's fairly impossible to do if you're > doing software-only, and probably isn't a good idea anyway. All you'll > end up doing is ticking people off as they change system information > around on you to bypass the "copy protection". Even semi-devious > tricks like locking on root filesystem UUID (which I think would be > cool to play with once or twice but not actually use for real code) > can be easily fooled by something like tune2fs, if the underlying filesystem > has UUIDs at all (ext[23] and XFS do, but does JFS, reiser, etc?) > There are valid reasons besides software license locking to want to know if you are on a specific machine. Sometimes utilities just need to know, and, considering that a machine may be aliased, it can sometimes be tricky to get good info. <opinion> Also, there are valid reasons for licensing and selling/leasing some products. I used to work for a company which sold a very high-value, small customer base mechanical engineering product. This product would never exist as an open source product because it took too much work to make the product. Although I believe in the open source model, I don't think it's all-encompassing. Some products just need income in order to live. </opinion> Bill

Nov. 6, 2001

Re: [Wlug] Unique ID

by Theo Van Dinter

On Mon, Nov 05, 2001 at 08:00:11PM -0500, Peter Nikolla wrote: > I am tiring to right a locking mechanism for my program. I wanted to find > something unique on a Linux/Unix system that I could read to lock my code to > that machine. Does anyone know of a call I can make to get an unique ID > like a disk ID, or system ID? I'll agree with Keith on this one, it's fairly impossible to do if you're doing software-only, and probably isn't a good idea anyway. All you'll end up doing is ticking people off as they change system information around on you to bypass the "copy protection". Even semi-devious tricks like locking on root filesystem UUID (which I think would be cool to play with once or twice but not actually use for real code) can be easily fooled by something like tune2fs, if the underlying filesystem has UUIDs at all (ext[23] and XFS do, but does JFS, reiser, etc?) Typically, software tries to lock on hostid, but that's manually changable on every platform I can think of, and since PCs don't have the extra chips for hostid, it's generated in software for Linux/BSD/etc for extra-easy changing. The hostid on some platforms (SPARCs specifically here) are some algorithmic version of the main ethernet controller's MAC address. But that's changable as well via ifconfig. Why are you trying to do this for anyway? Unless your software is so critical that its functions can't be done elsewhere, I think you'll find that most people will see the "lock" business as a reason to not buy the software. -- Randomly Generated Tagline: "Flourescent lights are generating negative ions." - Today's BOFH Excuse

Nov. 6, 2001

Re: [Wlug] Question

by Theo Van Dinter

On Tue, Nov 06, 2001 at 12:34:37AM -0500, Keith Wright wrote: > DNS has nothing to do with Netscape. Start by trying to ping. ??? Netscape has to use something to resolve hostnames... > network program) works as if by magic. You need to put your > nameserver IP address into /etc/resolve.conf thus: > search free-comp-shop.com > nameserver 208.218.130.4 > nameserver 208.218.130.5 > your names and numbers will vary. Also try ``nslookup'', and > post error messages. A few notes: 1) /etc/resolv.conf is the filename. Put your name servers in there as above "nameserver <IP>", multiple name servers on multiple lines. You can add the "search <domain>" and/or "domain <domain>" bits if you want, but it'll only help you if you don't use FQDNs (fully qualified domain names). 2) Check /etc/nsswitch.conf, and find the line starting with "hosts:". nsswitch.conf (Name Service Switch) controls how the C library does resolving of various things (passwd, group, services, hosts, etc.) Make sure it looks something like "hosts: files dns", unless you're doing NIS/NIS+/etc on your network. 3) Tools like "nslookup", "host", and "dig" will help you figure out if DNS can resolve the names that you're searching for, but they won't tell you if the system uses DNS (see #2). Tools like "ping" will resolve via the C library, so you'll know if the host can be looked up that way. -- Randomly Generated Tagline: "What do you do then? I touch that button and pray really really fast." - Lennier on Babylon 5

Nov. 6, 2001