Re: Programming question: sizeof struct?
On Fri, 9 Jul 1999, Joop Stakenborg wrote: struct foo { char text[3]; int num; }; sizeof would return 6 and not 5. So it's obvious that the compiler is placing a pad byte between text and num to align num. I want it to stop! That's odd. On most architectures, I would expect sizeof to result in at least 7 because int's are at least four bytes long. It is generally bad form to write code that, like this appears to do, relies upon the rules of packing and alignment. Of course, you didn't write it so you're stuck with it, but such code is generally less portable than doing things another way. Coding up a simple example, I see that, without any options, the sizeof(struct foo) results in 8. Adding the -fpack-struct option, causes the sizeof(struct foo) to result in 7. Adding the line #pragma pack(1) before the definition of the struct causes sizeof(struct foo) to result in 7. This is with what I have installed (it says egcs-2.91.60 on the computer at the house, but I get identical results on an older system running gcc 2.7.2.1) #include stdlib.h #include stdio.h struct foo { char text[3]; int num; }; int main() { printf(sizeof(struct foo) = %d\n, sizeof(struct foo)); return EXIT_SUCCESS; } I believe that, based upon your entire message, there are three possible conclusions that can be reached about your difficulty. 1: You may be using an architecture where the char type (upon which the typedef is based) is two or more bytes long. 2: You may be using a different C compiler than I expect. 3: You may be misinterpreting the output from the compiler. (Suppose struct foo is defined as you give it, but you're actually looking at sizeof(foo) where foo is completely different. That sort of thing can drive you crazy.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: /usr/include/linux and /usr/include/asm?
) because on some computers sizeof(double) is 8 and 8 is an acceptable value for FOPEN_MAX. Yes, it works, at least some places, but it's stupid. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: /usr/include/linux and /usr/include/asm?
On Tue, 29 Jun 1999, Carl Mummert wrote: It's not the symlinks, it's the contents of /usr/include/*.h that's the problem. They are the problem, but they cannot be fixed. Since the GNU C library is portable to various kernels and hardware platforms, it has to get its information about the underlying system from somewhere. Now, this is getting nonsensical. First, you tell me that the reason that there are no symlinks is because the details of the underlying system can interfere with the operation of the libc. Now, you tell me that the reason that you can't apply a proper fix the first problem is because sometimes libc has to get information about the underlying system. If you delete the symlinks and replace them with directories containing files that don't have anything to do with the kernel that's running you have eliminated any possibility of libc-based programs getting information about the underlying system. That means that the libc itself must not have access to information about the underlying system on those computers with the symlinks removed. This technique works, in most cases, because the libc doesn't have any business knowing what the kernel innards look like. Since most programs talk to libc rather than using kernel-specific interfaces, it works for most people. Unfortunately, I sometimes do kernel-specific programming and I think that this shows poor design and a complete lack of thought. I honestly expected better of you. (Well, them. I mostly mean the glibc maintainers.) Making my life a little tougher based upon the idea that I'm more able to deal with it than Joe Random Luser is NOT ACCEPTABLE to me. It shouldn't be acceptable to you, either. The files in /usr/include/sys (which, as you say, are the bulk of the problem) have no business referring to any files in /usr/include/linux or /usr/include/asm unless they need things that can change if the kernel changes. The proper approach would be to separate out the libc header stuff that's truly system-dependant (they won't kill libc-based programs because otherwise you'd need a new libc for every different kernel version) and put them in new header files that don't have any libc-killer stuff You can do that without changing the meaning of any of the linux include files if you use the approach outlined by Plauger in the C USER'S JOURNAL several years ago. However, I expect I'm the only one who thinks that's the proper approach so, how's this for a solution: Give the /usr/include/asm and /usr/include/linux directories up as lost causes. Instead, define new directories. Say /usr/include/kernel-asm and /usr/include/kernel-linux. Make THEM symlinks to the appropriate directories in the kernel source tree. I would do this on my own computer, but I don't really get any benefit from it unless everyone else, particularly those who develop modules and whatnot, decides to use the same convention. Therefore, we then attempt to convince people to actually use those directories instead of /usr/include/asm and /usr/include/linux for the good reasons that we've discussed. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: New windows
On Tue, 22 Jun 1999, Revenant wrote: On Mon, 21 Jun 1999, Christian Dysthe wrote: Instead I am again presented with the taskbar and popup menus that hasn't really changed for years. Is this to conquer new users, or is it becuase Linux users secretly have missed the look and feel of MS Windows? Well, this seems like the ideal opportunity for me to advertise my ignorance: What exactly is the difference between a window manager and a desktop environment? A disclaimer: I'm not part of the GNOME or KDE development effort. I don't really know where they're going so all this is just speculation. A window manager manages windows. (Duh!) A desktop environment does other things. Look, the window manager does things like lets you position and resize your windows and crud like that. The desktop environment allows you to interact with the system. Whereas the window manager knows about icons only so far as to allow you to interact with iconized programs, a desktop environment knows about icons that do things like launch applications. Is this a giant step backwards? Hasn't this been done in Windows for years? Well, yes, if you limit yourself to Windows, there really isn't much more there than having your UI be your file manager. However, OS/2's Workplace Shell (what I sincerely hope GNOME, at least, is trying to achieve) shows that what you CAN achieve is a whole lot more than a simple file manager. With an object-oriented window manager, those icons on the desktop become code that standalone applications can interact with in well-defined ways that are moderated by the environment. (Actually, they're moderated by the object broker, but that's a distinction without a difference for the end user.) What you end up with is something that's like OLE (now ActiveX) but where the underlying system participates in the object model. This can be really incredibly neat because you can build componentware from it. An example might prove useful: One of the applications that I was working on back when I still worked on OS/2 applications was a satellite tracking system. Under Windows, and regular X, such an application would normally read a file with the satellite data (called keplerians) and then present the user with a list of kinds of things he might want (AOS/LOS predictions, real-time tracking, and so forth.) Even if the internal structure of the program is MVC (Model-View-Controller) the application still is a single, monolithic program. It can't be embedded or extended without knowing the details of the source. With the WPS, what you do is create desktop objects for satellites and ground stations. These desktop objects include the data and code needed to determine the position and motion of an object given the time and are created through some sort of object creation process. When you write your application, you then don't need to include any satellite tracking stuff. Instead, you simply interact with those desktop objects (usually, but not always, using some sort of drag-and-drop interface) and do whatever it is you want to do. The environment does things like moderating the interaction and generating menus and such so that those objects are first class objects in the system. It also keeps track of the state of the system so you can do something like declare one of the ground stations as the default station since that's where you expect to be most of the time. If the guy who defines the objects does his job correctly, and the satellite and station objects are both derived from the same superclass, other objects can be defined that also work with the original programs. An example might be an object that is like a station object, but whose position is read from a GPS receiver rather than being fixed in space. That means that someone else who sees a better way of presenting the information but who doesn't know beans about the universal variable formulation can create a better tracking program (better in the sense that it fills some need better, has a better display, or whatnot) and it'll work just as well, in terms of tracking satellites, as the one written by the orbital mechanics guru. As long as the interface is published, it's automatic. Back before IBM (in effect) pulled the plug on OS/2, the development tools were getting in to the act so that the applications using the objects could be done in REXX or languages similar to Visual BASIC(tm). This is truly nifty and is truly a goal worth working towards even if it results in systems that appear to be superficially similar to the the craptacular stuff that come out of Redmond, WA. That's why it's the sort of thing _I_ would expect to see them working towards. I wonder what they expect to achieve. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: PPP problems
On Wed, 2 Jun 1999, K.Y.Lo wrote: I have configured the CHAP/PAP chatscript to connect ISP with dynamic ip. dont have no clue what does 'last messge repeated 11 times' mean? Just what it says. The last message (in this case, this message) Jun 3 01:48:33 griz pppd[197]: sent [LCP ConfReq id=0x1 mru 552 asyncmap 0x0 auth chap md5 magic 0x9df3beb5 pcomp accomp] Was repeated 11 times. I config PPP for so long . I cant getting thru ISP. It looks to me like the ISP isn't set up to start doing PPP right away. What happens when you dial the system with a terminal program and log in that way? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Groups
On Wed, 2 Jun 1999, Oleg Krivosheev wrote: come on, that's not very good idea to edit groups manually Why not? It's just a delimited text file. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: ISPs
On 3 Jun 1999, Colin Marquardt wrote: Starting with pppd-2.3.7, you don?t even need the DNS addresses. Search for the option usepeerdns in /usr/doc/ppp/README.gz and take a look at /usr/doc/ppp/examples/scripts/ip-up.local.add This only works if the other end is set up to give you the DNS addresses. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Leased pppd LCP ConfReq trouble...
On Wed, 2 Jun 1999, Alexey G. Khramkov wrote: I'm configuring leased line to my ISP. ISP has a Cisco router. When I try to connect the pppd hangs up the line after 10 (by default) sended LCP ConfReq packets. No other messages in /var/log/ppp.log with debug and kdebug 7 options in /etc/ppp/options. snippage My /etc/ppp/options: asyncmap 0:crtscts:lock:modem:netmask 255.255.255.252:noipdefault: -detach (for test reason):-pap:-chap(ISP doesn't use any for leased): debug:kdebug 7 (I don't sure it's work, but anyway):noipx My /et/ppp/options.ttyS1: crtscts 194.84.224.26:194.84.224.25 netmask 255.255.255.252 -chap -pap modem persist defaultroute I look very suspiciously at the IP address configuration, the netmask configuration and the -chap, -pap part which, I note, you have both in the command line that calls pppd and the /etc/ppp/options file. I don't know how it works at your ISP (obviously) but Brokersys requires that all incoming callers authenticate, even the permanent connections. (We don't run any nailed up circuits, so literally anyone can call the incoming lines.) Your local:remote IP addresses don't apparently match the netmask. (At brokersys, the addresses for the local and remote would be in different subnets---the PPP connection makes a natural boundary for transferring between subnets and this maximizes the number of available addresses for the end user. Both of the addresses you give are inside the same subnet.) You might try replacing the line 194.84.224.26:194.84.224.25 with one that reads : (That's just a colon all by itself on the line.) Or delete that line and add the options ipcp-accept-local and ipcp-accept-remote. Connect: ppp0 -- /dev/ttyS1 sent [LCP ConfReq id=0x1 asyncmap 0x0 magic 0xbb31 pcomp accomp] It sure would be nice to know what the ConfReq packets look like. You might try disabling magic number negotiation (-mn) and compression (I think it's -pc) and see what that does. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Is there a PPPoE client for Linux?
On 1 Jun 1999, Arcady Genkin wrote: My ISP is going to implement that crap for my ADSL connection. They'll of course be only supporting WinXX. Should I start looking for another ISP? PPP over Ethernet is a truly stupid idea whose purpose is to allow DSLAM manufacturers to sell equipment to both the telephone companies and ISPs. (Why would you want to provide access to a packet-switched network over a circuit-switched network simulated on a packet-switched network? Ick!) If there are any alternatives, I'd suggest switching. I was unable to find any information at DejaNews or AltaVista about support it, though. I think the answer is no. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Wierd PPP Problems
On Sun, 16 May 1999, Peter Ludwig wrote: You misunderstand. I'm not telling you or Mr. Hoover to ask the ISP what the problem is, only what a disconnect code (or however it's done on their system) is. However, I would most certainly agree with Mr Hasler about the fact that most ISP's will say sorry no-do-linux and that's it. I myself was connected with one of the big ISP's here in Oz, and well, they decided that not supporting linux was going to be a big thing for them. You obviously run an ISP, so... how's this one, would you answer this question if it was asked of you? When I'm connecting to the server my ppp connection requires the remote-ip address. What is the remote IP address for your system for dial-up accounts? They decided that they could not provide me with that information... How's a guy supposed to connect ANYTHING other than win95/98 to their system? Of course they couldn't tell you. That's cause there is no way for them to know this in advance! Any set up that REQUIRES this information is badly broken. Fortunately, Linux's set-up doesn't require that information. In fact, you don't need any more information to set up dial-in Internet access with Linux than you do with Windows-9[58], but it helps if you have some clue about how this stuff works. From my perspective, your question, as posed, has no answer. Until we bring the 5396 on-line, at which point it will be five, we have four access servers answering calls. Each one has its own IP address. One of those (ernestine) is for permanent connections, but the other three (bertrand, grover, and cleveland) are assigned to callers randomly. So, you call and ask which IP address my access server is going to have and, unless you're paying for a nailed-up connection, I have to answer I don't know (which is true, how am I to know which AS is going to answer the call? Tea Leaves?) and you go away mad because you don't understand that you don't need the information that you ask for. BTW - It was dynamic IP addressing, with the remote end not providing it's IP address for the connection, so pppd would not work properly... Horse Puckey! I've been doing dynamic-IP with Linux (in both directions) for the better part of four years, now, and it is painfully easy to set up. Simply tell your pppd to accept whatever addresses it is given. (man pppd is your friend.) That works for both local and remote addresses. Just let pppd figure out what's going on and everything works like magic. If you try to tell my equipment what address you have, it'll hang up on you. (I hope! Keeping a call alive even though it had no prayer of working was one of the bugs that Ascend was supposed to fix.) To make dynamic IP addressing work, you simply don't tell pppd anything about either IP address and, as long as you have the defaultroute option set, it works just fine. At least it always has for me. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Wierd PPP Problems
On 13 May 1999, John Hasler wrote: This works best if you can call the dial-in number and talk to support simultaneously. If he can't do that, then he's got the wrong Internet provider. It will no doubt startle you to learn that some of us have only one phone line (and only one local ISP as well). It will? THAT assertion certainly startles me! How do you know what I expect unless I tell you? As a matter of fact, I normally assume that a user normally has only a single telephone line. Even if I had been under the impression that most people had as many telephone lines serving their house as I have serving mine, the long hours of troubleshooting connection difficulties with those people who have only a single telephone line would have relieved me of that misconception pretty quick. That's why I pointed out that it was best instead of saying something like you should call while you're attempting to connect. Many people I know of will call on their cellular phones (assuming they have them---not necessarily a good assumption for someone running a bargain-basement ISP like Brokersys) because it makes the process so much easier. As to the single ISP issue, I know of no localities that are served by one Internet provider that aren't also served by one or more of the national providers. Some of those national providers even offer 800-number access for those areas without local POPs. Of course, I have no way of knowing if Mr. Hoover did this, but asking on this list about ISP connection difficulties without consulting the ISP first is probably not the best way to go. I'm sure that this is not true of your company, but almost all ISP's would tell Mr. Hoover We do not support Linux and refuse to listen to his problem. You misunderstand. I'm not telling you or Mr. Hoover to ask the ISP what the problem is, only what a disconnect code (or however it's done on their system) is. Since that is independent of the operating system you use, it doesn't matter what you're running. You should NEVER expect telephone technical support to solve your problems, you should only expect them to provide the information you need to solve your own problems. (If they happen to be able to tell you what the problem is, that should be viewed as a happy windfall.) Anyway, if they won't tell you why your connection was last dropped and you can't find another ISP, then you're stuck with increasing the debug level on pppd and undergoing the mind-numbing task of digging through mounds of low-level output. (It helps if you have a good grasp of troubleshooting techniques. For example, I'd start with attempts to connect through minicom, though, just to see what it looks like.) As a matter of fact, here at Brokersys, although we use Linux exclusively for our servers (and more than a few workstations) we provide only limited support for the Linux systems of our dial-in callers. (We have a package that includes a pap-secrets file and a script that runs pppd and instructions to modify those files so that they'll work for your userid, and that's about it. There are no step-by-step instructions, like we have for Windows-9x.) This is because there is a general assumption that a Linux user has some idea what to do with the five critical pieces of information and because there are 46-dozen different ways of setting up PPP access through Linux so no useful set of step-by-step instructions would be applicable to everyone. However, you can call us up (the number's in the signature) and ask us what our dial-in number is (281-774-7100 for modems, 281-774-2741 for ISDN callers) and for the address of our DNS servers (dial-in users should use 192.168.136.1 and 192.168.136.2) and what the smtp server name is (mail.brokersys.com) and the pop3 server name is (mail.brokersys.com) and the Web server name (www.brokersys.com) and what caused your last disconnect (cause code 185 or far end disconnect, probably, if you're having enough trouble to actually call) and we'll tell you. Why not? That's hardly secret information. I'd even bet that you can get that sort of information out of the big national services, too. I've worked with some of them working out reverse-DNS problems on a couple of occasions, so I know they have people who work on just that sort of thing. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: What group does Apache run as?
On Wed, 12 May 1999, Stephen A. Witt wrote: I'm trying to install Jitterbug, which is a bug tracking system, and I ned to be able to figure out which group apache runs as. You don't find out what group apache runs as, you TELL IT in the /etc/apache/httpd.conf file on the line that begins Group. I understand that on Debian installations, it's usually www-data. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Wierd PPP Problems
On Wed, 12 May 1999, Wayne Topa wrote: Quoting John Hasler([EMAIL PROTECTED]): Chris Hoover writes: I'm having some strange problems with ppp on my dial out server. Everything has been working fine for quite some time (over 51 days of uptime on 2.2.2), and then tonight I started having problems connecting. Does anyone know what is going on, and how to fix it? At a guess, your ISP changed something. Does it fail every time, or intermittently? What sort of authentication are you using? What does your chat script look like? John is right, its your ISP. I used to have a problem like that with Netcom, twice a year. Their Support group kept telling me I was doing something wrong. My 'fix' for 2 years was to put my ppp-on into a 'Netcom' script that repeated the ppp-on attempts. It would take anywhere from 1 to 12 attempts before I got connected. The problem would last for a week or two and then, after they got so many complaints, they would fix it. Speaking as the owner of an ISP that has fielded his share of such complaints: Assuming that your ISP is using remote access servers, what you describe is a classic symptom of a telephone trunking problem. Either a port card has gone bad and you're getting controlled clock slips or you're signal is going through an old 1A switch that needs maintenance. In either case, it has nothing to do with the ISP (although it also has nothing to do with your system, either.) You should feel lucky that they get fixed after a couple of weeks, for SWBT hasn't to my knowledge EVER fixed a trunking problem in Houston. They certainly have never fixed one that we attempted to report. If the authentication server fails, NOBODY is able to log in and that gets your attention real quick. Usually, it takes about 20 minutes to fix, most of that being the time it takes to get a hold of someone who has root-level access and an understanding of what needs to be done. If there's something wierd about your setup, then it won't ever connect at all no matter how many times you retry. The original poster should contact his ISP (CALL THEM! nothing frustrates us faster than getting an email that talks about problems connecting except for the emails we get that talk about problems getting email) and see what the failures look like from their end. This works best if you can call the dial-in number and talk to support simultaneously. If he can't do that, then he's got the wrong Internet provider. If the modem is an internal modem, he should be prepared to power-cycle his computer because some modems don't like long uptimes. (It's a lot easier to powercycle external modems, which is one reason I recommend them.) Of course, I have no way of knowing if Mr. Hoover did this, but asking on this list about ISP connection difficulties without consulting the ISP first is probably not the best way to go. A critical piece of the puzzle is WHY the connection failed and for that information you'll need to go to the other end. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Life at 4 bogomips
On Sun, 9 May 1999, Pann McCuaig wrote: On Sun, May 09, 1999 at 04:50:30PM -0700, George Bonser wrote: I guess I won't be compiling any kernels on that box ... well, maybe ONE just to see how long it takes. :) If memory serves, my first linux box, a 386SX-16 with 4MB took about 5 hours to compile 1.0.9. ISTR that compiling kernel V0.11 took about 5 minutes on the old Northgate 4MB 386-20 I was using in 1992. Of course, the kernel's kind of grown since then. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: setting dns values
On Mon, 26 Apr 1999, Wayne Topa wrote: Quoting Micha Feigin([EMAIL PROTECTED]): Where do i write the dns values for dialup connections, and how do i specify deferent dns values for deferent connections ? Thanx Nameservers are put into /etc/resolv.conf like: AFAIK only the first 2-3 entries are used. Any nameservers (even if not your ISO,s) are used. It would seen to be faster if they matched your ISP tho. I know of no way ( or need) to switch based on which ISP you connect to. I see no noticable difference here using nameservers from other ISP's. Danger, Will Robinson! Some ISP's (the one I run, leaps to mind) use RFC-1918 addresses for utility functions like the nameservers for dial-in users. This is so we don't have to reconfigure all of the dial-in users when we renumber the rest of the system. Unfortunately, that means that you can't use them from outside Brokersys. Also, you can't just use any old nameserver values because they sometimes change. If I were to set up those Mindspring DNS server values, they would work until Mindspring decided to change their DNS numbers. Since I'm not a Mindspring customer, I wouldn't know until it stopped working. All-in-all, it's best for all concerned if you use the DNS server addresses that are associated with the ISP you are connecting through. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: UMAX 1220S, SCSI card (436P?)
On Mon, 26 Apr 1999, joost witteveen wrote: So, does anyone know if the UMAX Astra 1220 S also works with a normal SCSI card? (If so, I'll just buy one of those). I use the 1220S with an Adaptec 1542 that I've had forever. It seems to work okay. I had to get a cable that adapts between the Centronics connector on the Adaptec to the 25-pin D-connector on the scanner, though. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: [OT] Whitch adaptec SCSI card?
On Mon, 26 Apr 1999, Ries van Twisk wrote: Dell is currently offering a PowerEdge 1300. 18Gb Ultra-2/LVD SCSI HD. Adaptec AIC-7890 Ultra-2/LVD SCSI Controller (inergrated) DDS-3 DAT-Drive 12/24GB SCSI-3 I will call them about the of they can offer my an other SCSI Controler. I haven't talked to the support engineer yet. Hopefully I know this by tomorrow. The controller is integrated into the motherboard, so it's unlikely that you'll get them to change it. However, that SCSI controller is (I believe) what Dell delivers in their Poweredge 2200's that I install Debian 1.3.3 on without any problems. I'm sorry, I wasn't paying attention. Are you having problems or are you anticipating them? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
gated
Anyone using gated of any variation with any of the 2.2.x kernels? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: ISA vs PCI Modem
On Sat, 24 Apr 1999, Greg Scharrer wrote: I am thinking about buying a 56k modem. I have a 28.8k modem. I know not to buy a Winmodem. I have seen ads for ISA and PCI modems. Is one kind better than the other? Does the type of board slot affect capability or performance? I can't imagine that ISA vs PCI would have that much effect on the system load even when communicating at full speed, assuming that you've got the buffering set up properly on either one. (In fact, my recommendation would be to get an external modem, which makes the point moot.) My comment (and I'm speaking as an owner of an ISP who has taken his share of your service is crappy because when I use my $0.59 modem I can't connect with it faster than 49,333 BPS calls) is that if you want good performance, you should get a decent modem. That means that if you buy USR (I don't, which is a long story) you get a Courier rather than a Sportster. I have heard really good things about Microcom and I've personally had good luck with DataRace, not that I would be offended if you didn't take my advice. I don't buy modems, any more, because 28.8 is fast enough for everything but connecting to the Internet, and I don't use a modem to connect to the Internet. (I'm running ISDN now, but I'll be installing DSL in a month or so.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Disk geommetry, was Re: Kernel Upgrade: Why?
On Thu, 22 Apr 1999, Ookhoi wrote: Well, there was a discussion here about a benchmark Linux vs NT, and some people here said that the preformance of Linux could have been affected by the fact that Linux was near the center, and NT on the outer side. Probably not the reason. And the data on the outer side passes the heads much faster than the data on the inner side. But then, there is much more data on the outer side, and a piece of data on the outer side will go round in the same amount of time as a piece of data on the inner side.. I am not aware of any disks that use a higher density recording format for the outer tracks than they do for the inner tracks. As far as I am aware (and I really haven't paid much attention to such things since ST-277's were state of the art) the bit density of the outer tracks is LOWER than the bit density of the inner tracks. That's because the outer tracks are physically larger, but they hold the same number of bits. Not that it matters. The whole disk spins as a single unit so even if there were more bits on the outer tracks, you'll still wait the same amount of time (on average) for the sector you want to come around. Read on, and I'll explain. So, is there an advantage if whe put for example swap at the outer side of a disk? NO! Look, the access times for disk are dominated by two times, the time to seek to the correct track and the time to wait for the data to come around again on the disk. The time it takes the data to come around on the disk is, on average, one half of the time it takes for the disk to go around once. That's independant of everything else and is a fairly short time, anyway. The time it takes to seek to the correct track depends upon where you're seeking from and where you're seeking to. Obviously, if the heads happen to be at innermost cylinder, it will take longer to seek to the outermost cylinder than if the heads were in the middle or toward the outside. So, for higher performance in a situation where you're too cheap to add enough RAM, you'll want the swap file near where the heads are likely to be. You can also turn that around. Seeking to the middle from either extreme is likely to be faster than seeking to the other extreme. (This works for both average and worst-case times.) Predicting where the heads are likely to be takes some doing, especially on systems with effective disk caches, but you can take some educated guesses. The middle of a disk is a better guess than either extreme, but isn't necessarily the best guess. If you spend a lot of time reading and writing (especially writing) files from a particular partition, you might want to put the swap file near that partition on the principle that the heads are likely to be near there anyway, so it should reduce the amount of time waiting for any given swap. An extreme example of this would be where you dedicate an entire drive to a (fairly small) swap partition. That's how the news servers I use do it. For something less extreme, I kind of like the recommendation made by OS/2 gurus: Their advice was to put the swap file in the most used partition on the least used drive. You might try something like that where you put the swap partition in the middle of a disk that isn't used for very much. In short, my recommendation for boosting the performance of a computer that uses a significant amount of swap is to add RAM to the computer. HTH. HAND. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
RE: IDEA:Offical Debian Support Team?
On Tue, 20 Apr 1999, Madel, Kurt wrote: I believe that support for Debian is very important and is something that should be investigated. However, I believe that it may be difficult to combat an over-commercialized distribution if you start pushing for professional 24X7 commercial support. The kind of support that corporate business requires is of the commercial variety. Actually, corporate business has resources in place to do INTERNAL support. The sort of telephone camp on hold support lines that have become what people think of as technical support are something that corporate business can do without if their internal support personnel have the tools they need to do their jobs without it. I believe that the so-called information revolution is going through its third phase. The first phase was when computers were large, rare, and expensive. The second phase started with the rise of the microcomputer and the introduction of the ISV. The third phase started with the rise of the Internet and the introduction of freeware of commercial quality and a substantial feature set. People keep talking about how Linux, and other freeware, are unsupported like what they get from Microsoft can be characterized as support. However, I view poor support as a problem of the second phase and as soon as we can escape second-phase thinking (such as I've got to have a toll-free telephone number that I can call to harass the vendor) the issue of support will simply disappear. After all, who needs support when you (or your hired expert) has the source? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: QPopper question?
On Mon, 19 Apr 1999, Clint Rhodes wrote: I am having some strange problems with my qpopper. I have several usres who are not able to check their email, and I also have several (including my office) who ARE able to check their email. Any suggestions would be helpful! Things to check: Quotas (qpopper copies the mail spool file so if the mail spool file is more than half the size of the remaining quota, it'll fail) Password/username combos (telnet localhost 110 is your friend! The relevant commands are 'user', 'pass', 'list', and 'quit') General permissions problems. (As I recall, qpopper copies the spool file to .userid.pop so that must not exist or must be owned by the user.) The tasklist. (sometimes the mail client fails without cleaning up and you'll need to kill the qpoppers left hanging.) I can't recall if qpopper pays attention to expirations or not, but that might be something else to check. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: ftp has stopped working
On Tue, 13 Apr 1999, Richard E. Hawkins Esq. wrote: one of our systems is now refusing incoming connections, generating the error, ftp: connect: Connection refused This has survived through a reboot. Can anyone give me a hint? Check your inetd.conf and make sure that inetd is running. That's the message you get when there's no ftpd running on the socket that the client tries to attach to. (It's usually 21. You probably should also check the client to make sure it isn't doing something stupid.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: setting up DNS
On Sun, 11 Apr 1999, Lev Lvovsky wrote: hrm, maybe there's still something that i'm missing...let's say I want to register www.linuxnewbie.com (just an example), and have it point to my machine (silver168). The ADSL provider provides me with a static IP, so I figured this'd be possible (plus a friend of mine with the same setup did it)... You do NOT register www.linuxnewbie.com. You register linuxnewbie.com. Then, you set up your DNS records to include an address for the host name www.linuxnewbie.com. At Brokersys, we give all of our computers real names, like weck.brokersys.com and the use CNAMEs to map the logical names, like mail.brokersys.com, www.brokersys.com, and so forth, to the real names. However, there's nothing at all preventing you from having multiple A records that resolve to the same address or even just having a single A record to name www.linuxnewbie.com. (There are even circumstances under which the same name resolves to multiple IP addresses, but I won't go into them, here.) Also, it's possible to register a domain and have a zone file with no addresses in it. You might do this to park a domain. (Or, you might just have no zone file at all in that circumstance.) If you can, please have your ISP help you set up your domain. They've likely done it a lot more than you ever will and they'll be able to make sure everything is set up on your end (and theirs---they'll likely be providing secondary name service for you) before you send the application in to the NIC. It shouldn't be too much trouble for them to fill in the application for your domaind and to send you a properly formatted hosts file for the domain you want to set up so that you can use it as a basis for customization. (I don't know anyone in the ISP business who has written a DNS hosts file from scratch. Everyone I know got their first one from someone else and copies and modifies it to suit their changing needs.) One of the things that annoy us about some of our users is they occasionally send incorrectly filled-in applications (if we run the primary nameserver for a domain, we insist on being TPOC and we insist that customers be all the other POC's for that is the intent of the NIC) to the NIC and tell the NIC that we're providing name service for domains that we've never heard of. In any case, you WILL want to inform them if you need their secondary name service. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: setting up DNS
On Mon, 12 Apr 1999, iodine wrote: On the topic of DNS anyone kind enough to forward a copy of a named.conf file for bind that shows me how to setup the domains I host on the dns server as seconardy. named.boot contains (cut to show example of interested part) secondary domainpri\domain.zone I just need to know how I changed the named.conf on a primary dns to suit a secondary dns setup, and then forced it to only look at 1 master for changes? Really? I just used named-bootconf.pl that came with the updated bind. It generated stuff like this for my secondary domains: zone fakedomain.com { type slave; file fakedomain.com.domain; masters { 192.168.17.233; }; }; -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: setting up DNS
On Mon, 12 Apr 1999, iodine wrote: Really? I just used named-bootconf.pl that came with the updated bind. It generated stuff like this for my secondary domains: zone fakedomain.com { type slave; file fakedomain.com.domain; masters { 192.168.17.233; }; }; yeah I know but this is for a Slackware machine, and I don't have a Debian machine setup any more :( No offense, but what does Slackware vs Debian have to do with it? The machine from which I cadged my example was originally installed with Slackware 1.2 in 1995. (At least I think it was 1.2. It's been a long time and I've long since lost track of the disks I actually used.) In any case, the machine in question has, to my knowledge, no .debs on it except those that may be temporarily stored while in transit. I installed an updated bind from a tarball months ago and that tarball included the named-bootconf.pl program to which I refer. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: What DO you lose with Linux ???
On Sun, 11 Apr 1999, Jiri Baum wrote: If you have messages that MUST not get into hostile hands, I suggest reading a good cryptography text (sorry, I'm not keeping up with it these days; is `Applied Cryptography' by Bruce Sterling... ITYM Bruce Schneier. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Runiing processes. What are these?
On Fri, 9 Apr 1999, Christian Dysthe wrote: I was looking at running processes. Found these: 199 2 S0:00 /sbin/getty 38400 tty2 200 3 S0:00 /sbin/getty 38400 tty3 201 4 S0:00 /sbin/getty 38400 tty4 202 5 S0:00 /sbin/getty 38400 tty5 203 6 S0:00 /sbin/getty 38400 tty6 Why are all these running, and do they all have to run? I use Xisp to connect to my ISP. Those are there so that you can log on to virtual terminals 2-6 (when you're not logged-in on tty1, there's a getty running on tty1.) No, you don't necessarily need them, but most people find them convenient. If you want to turn them off (and I recommend you leave at least one running) read up on /etc/inittab. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: too big packages?
On Sat, 10 Apr 1999, virtanen wrote: Anyone knows a cheap way to get debian packages, which are too big for a floppy to a computer, which isn't networked and has got only a cdrom driver and a floppy disk driver? Is it possible to 'pkzip' deb-packages in a dos-machine to make them smaller and 'unzip' them in the debian-machine? A number of compression utilities have the ability to split archives into floppy-sized parts. The one that I have used most recently is RAR, which comes in both Linux and DOS flavors. (I got a copy from a cohort of mine, so I don't know where you would find it.) Note that I was using RAR for its multivolume capability, not its compression. I have also been known, in the past, to write short C programs to split a file up into parts that can be sneakernetted on floppy and then can be COPYed or cat'ed back into the original file. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Windows 49er Bug
On Thu, 8 Apr 1999, Richard E. Hawkins Esq. wrote: Isn't there something similar in Linux? Except the limit is a bit over a year? A bit more than that. The date rolls over in 2038 on 32 bit unices. Wrong limit. There's a counter in some MS-written systems (which includes OS/2, as I recall) that counts milliseconds since the computer booted. Since it's a 32-bit counter, it rolls over in about 49 days. Nasty things can (but don't necessarily) happen when it rolls over. ISTR that there is a similar counter in Linux that rolls over in about 400 days, but I've been having trouble remembering things lately. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Changing root password for sudo?
On Fri, 9 Apr 1999, William R Pentney wrote: I have changed my root password, and su takes notice, but sudo still uses the old password. How can I change the sudo root password? Do I edit etc/sudoers? To use sudo, you don't have to demonstrate that you have permission to act in some ways as root, that's what the sudoers file is for. All you have to do for sudo is prove that you are the person who is in the sudoers file. Therefore, the password you enter to use sudo is the one that you use to access the system, NOT the root level password. That is, in fact, the whole point behind sudo: You can have staff do some administrative-level tasks without passing out the root password. I hope that's clear? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: router @isdn
On Fri, 9 Apr 1999, Ronnie Cooley wrote: Hello I have installed debian2.1 I get pings back when i ping my router and when i ping my isp from debian. But when i start BitchX it will do nothing but blink at me. Any ideas. the puter that has linux on it is networked to another computer both hooked to same router i can get on the net in windows on both machines but not when i am in debian on the second computer, You don't give me nearly enough information to determine what's wrong, so you'll need to do some detective work. (FWIW, the computer I'm typing on is a Debian 2.1 system connected through a Pipeline 75 to the computer that runs the mail client, so it's not something inherent in Debian.) What services have you attempted to use besides IRC? Does ftp/telnet/http/whatever else work or not? If everything works from Linux, then you've probably got a BitchX configuration problem of some sort. (Not necessarily, though. It might still be a routing problem.) In any other case, then I need to know if the name resolution is functional: When you ping, do you have to enter an IP address, or can you use a name? If entering a name never works, then edit resolv.conf(5) to add the proper name server addresses. Otherwise, it's likely you have a routing problem. Make sure that all three pieces (the two computers and the router) have an IP address, that those IP addresses are all different, that those IP addresses were all assigned to you (or are RFC-1918 addresses which are different from those that your ISP may use) and that the netmasks are correct. If the same computer with the same IP address works under Windows-95 and not under Linux, then the problem is almost certainly that the netmask or default gateway is set incorrectly. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: parallel computing, rsh, and DHCP environment?
On Thu, 8 Apr 1999 [EMAIL PROTECTED] wrote: I'm trying to set up 2-4 Debian boxes at school to run in parallel under lam (MPI) and have several problems frustrating me. The machines are currently Not recognized by their names (eg. dasher.myschool.edu ) but rather only by IP address. Is it possible to get the systems recognized by their names? IP addresses change under this system at times, so names make it easier ... Of course it is! If the LAN was mostly static, then I'd suggest using the /etc/hosts files on each computer, but if things are going to change with any frequency at all, then you'd probably want to set up a name server on one of the computers and set up the others to refer to that so that you don't have to change the configuration of each computer when you change the address of one of them. If you're hooked up to the Internet, then ask your ISP for help. If not, install named on one of the computers, get a copy of the cricket (_DNS and BIND_ published by O'Reilly Associates) and go to town. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: parallel computing, rsh, and DHCP environment?
On Fri, 9 Apr 1999 [EMAIL PROTECTED] wrote: In [EMAIL PROTECTED], on 04/09/99 at 08:52 AM, Jonathan Guthrie [EMAIL PROTECTED] said: Of course it is! If the LAN was mostly static, then I'd suggest using the /etc/hosts files on each computer, but if things are going to change with any frequency at all, then you'd probably want to set up a name server on one of the computers and set up the others to refer to that so that you don't have to change the configuration of each computer when you change the address of one of them. OK, but if the addresses change occasionally, then it would still be a mess even with setting up a DNS, yes? (given that the hosts file needs IP addresses?) Well, the advantage of using DNS is that the required changes are limited to the file on the master DNS server. That means there is only a single place that needs to be changed. Of course, DNS servers work best when their addresses are fixed and known. Perhaps I'm simply stuck with making the best of a bad situation ... :( Out of curiosity, do DHCP-administered addresses change only when a lease expires without renewal? (These machines are not allowed to stay up constantly.) That depends upon whether or not you (as opposed to some MIS department) are running the DHCP server. Normally, computers that need names get fixed IP addresses and the DHCP server is configured accordingly. That means that a given computer would ALWAYS have the same IP address. If you don't have that sort of control over the network addresses, then you'll need to do some sort of dynamic updates of a name server. If you can't get a fixed IP address of at least one computer under your control, then getting everything to work can get ugly. It's not particularly complicated to write a program to handle the dynamic IP allocation under any circumstances, I've written similar address discovery code for my current day job, but the more control you have over the infrastructure, the easier it is to do. (You comment about the machines not being permitted to stay up 100% implies to me that you don't have a lot of control over the infrastructure.) I'm certainly not aware of any available application that can do what you need. However, someone else might be. Dynamic DNS is not exactly unheard of. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: What DO you lose with Linux ???
On Sat, 3 Apr 1999, Andrew Holmes wrote: Wouldn't it be nice if POP mail boxes could be set to automatically bounce messages over a certain size as soon as they arrive at the ISP :-) That would be the MAX_MESSAGE_SIZE configuration parameter from sendmail.cf, now, wouldn't it? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: using Hauppauge WinTV under Debian
On Thu, 1 Apr 1999, Andrei Ivanov wrote: I have one windows machine at home among many which I'm trying to transition to Debian but it has one card that I don't think will work with Debian. It's my Hauppauge WinTV card it's a TV/Radio tuner which works under all the winblows versions Just because it works with Windows doesnt guarantee it's work with Linux. It's a WinTV card. Basicly, logic suggests it's a Win-hardware, and therefor you can not use it with Linux. With the invention of winmodems and winprinters and suchlike, it has become common to say that these windevices will only work under Windows. That is not technically correct. There is nothing magical about Windows, it's just software. There's nothing magical about these devices, they're just hardware that uses I/O and memory mapping as well as IRQ and DMA mapping to do their work. It may not be possible to get programming information about winmodems and it may be necessary to implement a variety of protocols (V.8, V.80, V.42, V.42bis, and whatnot) in the driver, but it is, in principle, quite possible to use winmodems under Linux, and they're the hardest of those listed to do, or so I believe. I think that it may be about time for somebody to look into Linux drivers for winmodems, at least, if only just to see what the task entails. I know that people have been working on Windows-specific printers for some time. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: kernel unable to boot -- can't install
On Wed, 31 Mar 1999, Christopher Burrows wrote: i recently got a motherboard, and i am unable to boot linux on it. this is very frustrating, and i am looking for someone to help me, if at all possible, diagnose what is going wrong. here is what happens: i am trying to boot off the slink rescue disk (1.44), and initially the welcome screen appears. i hit enter to boot. then i see the loading. business for the usual amount of time. once this is through, the kernel begins to boot and report its status -- but it gets no more than a screen-length into that process when the machine reboots. i am unable to tell what the last step is _before_ the machine reboots, because it happens very very quickly. I got a motherboard that did this a while back. I did a DejaNews search and discovered that this was fixed, at least in my case by updating the BIOS. I had to boot DOS to do the flash upgrade, but it fixed my problem. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: PostgreSQL
On Sun, 28 Mar 1999, Sami Dalouche wrote: I wanted to install postgreSQL on a machine without X but postgresql needs xlib6g tk... Is there a package without X Tk .. support ? Why do we need X ? I run PostgreSQL on several computers. The X stuff is there for the X client and the tk stuff is there for the tk support. I think. It's been a while (couple of months, anyway) since I installed pgsql from .deb files, so I don't really recall, but I think that the .deb files installs everything, including the kitchen sink. You don't have a limited no-X version. I've run (and continue to run, my main database server has no X on it) PostgreSQL on computers without X, but I installed it from the tarballs from postgresql.org. ICOCBW. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: A question about compiling
On Sun, 28 Mar 1999, Jayson Baird wrote: I feel really moronic asking this, but what is the package that allows you to use make. Is it GCC or something else? I don't know about you, but _I_ would expect make to be in the make package. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: what is flex ?
On Sat, 27 Mar 1999, Will Lowe wrote: What is flex ? Flex is the gnu clone of lex, one of the original unix lexical analysis tools. A decent book on compiler construction could probably explain it better. Compilers: Principles, Techniques, and Tools by Aho, Sethi, and Ullman (1984, Addison-Wesley, also called The Dragon Book because it's got a picture of a dragon on the front) is probably the standard compilers text. Actually, I found that the book _Lex and Yacc_ published by O'Reilly Associates is quite readable (I read the whole thing in two days) and is a good introduction to things both lexish and yaccish (including bison.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Where is /etc/rc.d/rc.local on Debian?
On Sun, 28 Mar 1999, Marek Habersack wrote: You are NEVER going to find a SRPM of HP FireHunter, or any other commercial software pre-packaged for Red Hat. Well, such software won't make it into Debian then... I guess... I don't care if FireHunter is not part of Debian, but you have to consider this: Not every piece of software that I'm going to want to run on my Debian systems is going to be part of the Debian distribution. In the case of conventional (for lack of a better word) commercial software, it is CERTAINLY not going to be part of the Debian distribution. However, I may want to run it. Now, I can usually fiddle with it and, after a while, get it to work. However, not everyone who may want to run the software can and even _I_ would prefer to have it install cleanly without effort because, in most cases, I'd rather not have to mess with it. Trying to figure out why FlagShip doesn't seem to want to install correctly takes time away from the work I really need to be doing, which is why I bought the software in the first place. Now, I'm ornery enough to usually fight my way through the difficulties, heck, I even got Stronghold to work with PHP 3.6, but that ornerieness won't last forever. Eventually, I'll decide to not install software just to avoid the trouble. It occurs to me that the people who are trying to get the distributions to all agree on the same directory structure may be going about this in the wrong way. If you can get the ISV's to support a single directory format, then RedHat, at least, will use it. (The process of doing this is, of course, left as an exercise for the reader.) This is because it is in the ISV's best interest for there to be a standard, but it is in RedHat's interest for there to be no standard at all. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: The GNU thing
On Wed, 24 Mar 1999, Brian Servis wrote: *- On 24 Mar, Jonathan Guthrie wrote about Re: The GNU thing I can't get xemacs-20 to suspend with CNTRL-Z on my computer so I either have to use multiple virtual terminals or do everything under X. Do you know what the problem is? It is a bug in libgpm that is tickled when xemacs is compiled with gpm support. See bugs 20356, 20398, 22651, 23686. Indeed it is. The workaround that I have implemented is I kill gpm before starting emacs. I find the text mode mouse to be pretty useless, anyway, so I'll probably just set it up so that gpm never starts again. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Problems installing Debian on a 486
On Wed, 24 Mar 1999, G. Crimp wrote: The A20 line has something to do with working around a bug in memory addressing that first showed up in 286's I think. I don't really know much about it, but on my 486, their is an option in the BIOS to set the line. You might try changing this option in the Bios at boot and see if it helps. The original 8088 had 20 address lines called A0 through A19. This allowed it to directly access 1024k of memory. (It was split 640k/384k as RAM/ROM. It could have been worse, my understanding is that the original plan was to split it half and half. Oh, and the 8086 had 19 address lines called A1 through A19. I digress.) Anyway, on the 8086's and 8088's you could access the bottom 65520 bytes of RAM by setting the segment register to the very top of RAM and using an offset larger than the amount of RAM above the start of the segment register. Anyway, along comes the 80286. It has, in effect, 24 address lines (A0 through A23) for a total allowed memory of 16,384k. That broke those programs that relied upon the memory wrapping around. Since they all ran under DOS and since DOS was limited to 1024k, PC manufacturers put a control in which would not pass the A20 line through to the RAM, which had the effect of simulating the behavior of the 8086/8. That's what the A20 gate is about. You could run some DOS programs with the A20 gate disabled that you couldn't run with it enabled. You, of course, want to run a protected-mode operating system, where relying on tricks like that simply cannot work. Since many people also wanted to run protected mode software, (to do things like loadhi and with extended memory and suchlike) the mboard manufacturers made that gate configurable. That's the whole story, to the best of my knowledge. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: The GNU thing
On Thu, 25 Mar 1999, Marcus Brinkmann wrote: I think this statement is absolute truth. Linux wouldn't exist without GCC, but it certainly could without textutils or shellutils and suchlike. interesting opinion. Maybe you never tried to dpkg --purge --force-depends --force-essential those packages. So you're saying that the FSF is the only possible source of things like ls, rm, mv, etc? And here I thought that the BSD folks had written their own. Gosh, the responsibility of being the only folks on the face of the earth who are CAPABLE of producing a make or a sed must be absolutely awesome! Where do I go to bow down to their feet? My point is not that nobody uses those systems, but that anyone with even marginal skill can write those programs. GCC is harder to replace. But then, some sort of textutils you need. In this regard, I find Linus' comment utterly bullshit and closed-minded. OTOH, I would not hesitate to call Linux insignificant, compared with the GNU projects. Reminds me of the joke: Q: What's the difference between Linux and GNU? A: Linux has a kernel that boots. While that isn't strictly true any more, it was true for a long time. It is still true, if you aren't willing to run beta code. If the FSF is full of such SH programmers, why couldn't they cobble together something in a couple of months? I did it years ago. (The answer, of course, is that they don't know how to write the essential stuff first and let the rest just happen. That's how Linus got so far ahead without a dedicated following or taxpayer support.) One more thing: You talk about missing perl because it depends upon cat. If you are unable to write a cat (it's about 20 lines of C, fer crissake!) or an ls or an rm then you don't deserve an opinion about whose work in necessary and whose isn't. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: curious syslog line
On Tue, 23 Mar 1999, Ryan Yeske wrote: Mar 23 15:44:01 mustard modprobe: can't locate module char-major-6 this line keeps showing up in my syslog file, has been since somtime moving from Debian hamm--slink and kernel 2.0.36--2.2.3. does anyone know what this might be? Well, according to the contents of my /dev directory, char-major-6 is lp[0-3]. So, it looks to me like you don't have your parallel printer module set up properly. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: what is net-pf-19?
On Tue, 23 Mar 1999, Gregory T. Norris wrote: Since it doesn't seem to be hurting anything, I've disabled it for now by adding alias net-pf-19 off to /etc/modutils/aliases... still, I'm curious about what it was trying to load. According to /usr/src/linux/include/linux/socket.h, something called Acorn Econet uses protocol family 19. I've never heard of it, but experimental support for it is in kernel 2.2.3. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: The GNU thing
On Wed, 24 Mar 1999, David B. Teague wrote: The importance of compilers was one reason I chose to license Linux under the GNU Public License (GPL). The GPL was the license for the GCC compiler. I think that all the other projects from the GNU group are for Linux insignificant in comparison. I think this statement is absolute truth. Linux wouldn't exist without GCC, but it certainly could without textutils or shellutils and suchlike. the Emacs editor is horrible, for example. While Linux is larger than Emacs, at least Linux has the excuse that it needs to be. Emacs is a religion. Some hate it, some love it. Few are neutral. I love Emacs. I respect those who use small editors and restart the editor for each compilation error. Okay, so emacs is a religion, I can deal with that. I use emacs the editor quite a bit. However, I can't get xemacs-20 to suspend with CNTRL-Z on my computer so I either have to use multiple virtual terminals or do everything under X. Do you know what the problem is? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: The GNU thing
On 24 Mar 1999, Gary L. Hennigan wrote: Jonathan Guthrie [EMAIL PROTECTED] writes: | Okay, so emacs is a religion, I can deal with that. I use emacs the | editor quite a bit. However, I can't get xemacs-20 to suspend with | CNTRL-Z on my computer so I either have to use multiple virtual terminals | or do everything under X. | Do you know what the problem is? When you're running it in a terminal window you can't get it to suspend when you hit Ctrl+Z? Yes. Is your suspend character set to Ctrl+Z? Yes. (I also verified this by doing the stty -a, just to be sure.) The symptoms (gosh, wouldn't have been nice for me to include the symptoms the first time) are different from those I would expect for a different, or missing, suspend character. xemacs stops taking input, but won't start the shell. I can hit ^C a couple of times and get xemacs' attention, but the only thing it'll let me do is abort the edit and dump core (and it won't do the core dump because the default core size, as set by ulimit, is 0, and I never remember to change that before I run emacs.) This is under Debian 2.1 as most recently released, kernel V2.2.3, and xemacs20-nomule-20.4-13. (FWIW, it also does this if I use the mule executable, not that it should make any difference.) Is it possible that the .emacs or the .xemacs-options file has something in it that could cause this behavior? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Why can't I force the NIC driver to load???
On Wed, 24 Mar 1999, Robert W. Mudge Miller wrote: My NIC (a Winbond W89C940 type) is not detected at boot time because its PCI ID is unknown (as predicted in the Ethernet-HOWTO, Sec 3.4). I am running Caldera OpenLinux Lite V1.1, Linux kernel v2.0.29. I am following the advice given in the HOWTO to bypass the PCI probing. I enter linux ether=11,0x6100,eth0 at the LILO boot prompt but still no NIC driver gets loaded. A computer w/o a NIC is well nigh useless. Any thoughts, anyone? Use the source, Luke! Patch the ne.c file to detect the proper PCI ID stuff and then use that long enough to download a later kernel. The later kernels detect a larger assortment of PCI network cards. If you don't know how to do that, it would probably be faster/easier for me to go where you are and do it myself, since you're in town and all. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Network configuration
On 22 Mar 1999, Shaun Lipscombe wrote: Yes this is true for stuff that you dont use night and day, like iso9660 support for instance. I would like to see a good reason for not compiling NIC support right in, like how much will that enlarge the kernel by? 100k ? The total sizes of the *.o files that are in /usr/src/linux/drivers/net that seem to be associated with my PCI ne2000 card is about 20k. I usually install all the drivers for stuff that I know I'm going to use and make modules for stuff (like the floppy and CD-ROM drivers and FAT and ISO9660 file systems) that I need occasionally, but not all the time, and for stuff (like the IPv6 stuff) that I think I might want to play around with at some point, but which I don't really have a good excuse for installing now. Also, some of the smallest kernels (and complete installations) I've ever installed were for routers and such. User machines, which tend to have sound cards and joysticks, and whatnot, tend to have larger kernels, in my experience. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: sound
On Tue, 23 Mar 1999, Eric Ravelomanantsoa wrote: In tere anyway to compile your kernel without have to use modules to enable sound If you have a non-pnp card, yes. For example, I've played with the same card as Doug for a while, and just gave up after a while. Found a different sound card, SB16. It wasnt pnp, and after a quick kernel recompilation it works fine. I compiled the driver into kernel. How did you do that, since make [menu/x]config of 2.2.x never suggest the Sounblaster stuff. Oh? Then how did I do it, then? I'm glad you asked, so I'll tell you. (I never use make xconfig, so this assumes you're using menuconfig or that they aren't different.) Go into Sound and select Sound Card Support. Select OSS Sound Modules. You should be given the option of choosing 100% Sound Blaster Compatables and, if you tell it to build it in to the kernel, it will give you the option to set the parameters for the card. Chunk O' Cheese, eh? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: when installing the kernel-source
On Mon, 22 Mar 1999, Ed Cogburn wrote: P.S. Some programs I've tried since going to the 2.2 kernel no longer have working sound. Sound in the mirrormagic game is now broken, for example, but things like GNOME and emusic still work. I can't get any midi players to work with by SB AWE-32 (not PnP) and 2.2.3. Of course, I really haven't tried all that hard. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: when installing the kernel-source
On Mon, 22 Mar 1999, Bob Nielsen wrote: I can't get any midi players to work with by SB AWE-32 (not PnP) and 2.2.3. Of course, I really haven't tried all that hard. Timidity works for me (SB16, not PnP, not as module). I'll give it a shot. awedrv (which is what I was using with 2.0.36) appears to play the file but no sound comes out. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: more ram and larger harddrive then the bios can take
On Mon, 22 Mar 1999, Hanish wrote: If we have a software which directly talks to the HardDisk controller, then It should be able to curcumvent this problem. However the boot partition will most probably have to be in the 8.4GB limit as it is the BIOS that load the initial part of the OS. I wouldn't be surprised if you can't have a boot disk larger than 8.4GB, but you can have other disks that are larger. I tried this earlier this year (on an old 486-50 motherboard) and wasn't able to get it to work. Linux and the BIOS had different ideas about the hard drive's geometry and they interfered with each other. I wound up getting a different motherboard, which it really needed, anyway. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: dependency problems
On Tue, 9 Mar 1999, John Cuson wrote: someone last week told me about the pc magazine article that was placed on zdnet and led to the spate of discussion we've just seen. my response at the time was it's not for weenies. that was, of course, before i began setting up a series of libraries to support a gnome application and ran into my current problem...how does one resolve a dependency problem that results because gnome-bin won't install because libgnome31 isn't installed, and libgnome31 won't install because gnome-bin isn't installed? can someone offer me a hint here? When I upgraded to slink a couple of weeks ago, I had a similar problem in dselect. (It seemed to be caused by a lack of free disk space.) I wound up deleting the request for both of the packages and everything that relied upon them and then starting over. I suppose another approach would have been to retrieve the packages manually and force an install. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: I can't believe this
On Thu, 11 Mar 1999, Richard Lyon wrote: Logically it may be better to spend some money on an os which doesn't require specialist training to understand. Right. Which one might that be? (I've used dozens of OSes and I haven't yet come across one like that.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: help me to undertand GMT time!!!!
On Thu, 11 Mar 1999 [EMAIL PROTECTED] wrote: In a message dated 3/11/99 11:31:55 AM Central Standard Time, [EMAIL PROTECTED] writes: One question: is midnite 2400 hrs or hrs? Or does it matter? Midnight is hours; there is no 2400 hours - after 23:59:59 it changes to 00:00:00. Midnight is 2400 hours. After 24:00:00 it changes to 00:00:01 I'm as authoritative on this subject as you are, so who's correct? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: realaudio with 2.2.2
On Mon, 1 Mar 1999, Joey Hess wrote: Use rvplayer.deb installer. It has the wrapper built in. Problem solved, painlessly. ...but only if you're using slink. I was already using hamm's installer and it didn't have that file. I upgraded yesterday and all is as you describe. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Permissions of the sysadmin
On Tue, 2 Mar 1999, Torsten Landschoff wrote: The superuser can get in, but root can always get in. Preventing that requires rethinking the concept of root. Hopefully root will stay as it is. I hate this message Access denied I get everytime I want to do something on the NT server in university (yes, I am the admin...) Oh, I wasn't advocating rethinking root I was just pointing out that you cannot effectively prevent root from accessing anything you want to. (In fact, it can be damn tough to prevent access to certain users if you allow them to sudo stuff.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
RE: modem
On Wed, 3 Mar 1999, Pablo Cernadas wrote: I need help to choise a modem that works on debian can somebody tellme about? External. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Help..can't delete broken links, directories et all!!
On Sun, 28 Feb 1999, Roddie Rod wrote: Anyway, now I a gig or more of broken symlinks, directories and files. Most are in /usr/lost+found but some are in /usr/lib/* and /usr/man. Problem is I can't delete them. I have tried to rm -f, delete using midnight commander and renaming the files! Nothing works is tells me operation not permitted. I've tried to chown and chmod nothing! Sounds like that partition has been mounted read-only, possibly because of errors that fsck can't fix automatically. What happens when you type mount? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: realaudio with 2.2.2
On Sun, 28 Feb 1999, Bob Nielsen wrote: I've been able to get it to work with some sites (which use video) and not others. Does anyone know if Real is going to have an upgrade soon? G2? The problem is that the real player relies upon broken sound system drivers. There is a library that emulates the broken driver called rpopen.tar.gz. Have you tried that? -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: realaudio with 2.2.2
On Mon, 1 Mar 1999, Bob Nielsen wrote: The problem is that the real player relies upon broken sound system drivers. There is a library that emulates the broken driver called rpopen.tar.gz. Have you tried that? Do you have a URL for this? Oh, I have no idea where it came from. ISTR that I found it a couple of weeks (perhaps a month) ago using a DejaNews search. A quick check of the source (it's GPLed) shows that the author doesn't take credit for this work, so you can't get it that way. However, since it's only 8k long and since it's GPLed, I put it on my personal Web site. http://www.brokersys.com/~jguthrie/rpopen.tar.gz There are no links to it or anything, you have to just know it's there. While you're at it, you can check out my 100 bottles of beer on the wall page at http://www.brokersys.com/~jguthrie/beer.php3 -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: LCP EchoReq over and over again
On Sat, 27 Feb 1999, Luiz Otavio L. Zorzella wrote: I'm having this problem for quite some time. I keep on seeing this message about LCP EchoReq on /var/log: Feb 26 06:35:12 localhost pppd[23168]: sent [LCP EchoReq id=0x4e magic=0x30c5fb97] Feb 26 06:35:12 localhost pppd[23168]: rcvd [LCP EchoRep id=0x4e magic=0x91dc4c0d] Feb 26 06:35:42 localhost pppd[23168]: sent [LCP EchoReq id=0x4f magic=0x30c5fb97] Feb 26 06:35:42 localhost pppd[23168]: rcvd [LCP EchoRep id=0x4f magic=0x91dc4c0d] many similar entries deleted As you see, it happens once every minute. Is this a problem? How can I make it stop? It is not a problem. It is caused by the ppd on your end checking to make sure the other end is still alive. I used to do similar checking when I was using Linux boxes as terminal servers. It was right handy because I could use the lcp-echo-failure command to determine whether the PPPD was talking to another computer or a modem with a stuck DCD bit. (I had to stop because of occasional problems with magic number negotiation. I don't recall the circumstances better than this.) I don't know if this works (and, no offense, but I'm not going to dig through the source for pppd to find out---do it yourself if you're of a mind to) but I would suspect that setting the lcp-echo-interval value to 0 might turn it off. According to the man page (you did read the man page for pppd, didn't you?) lcp-echo-interval sets the interval between echo requests if no other traffic is on the line. It apparently defaults to 60 on your machine. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Debian and Redhat - are most linux users missing the point?
On Sat, 27 Feb 1999, Tom Pfeifer wrote: Ed Cogburn wrote: As Deb becomes bigger, attracting more users, with some of them becoming developers, Deb's weaknesses such as the install problems will be addressed as well (am I the only one who likes dselect? :-) ). No, there's at least two of us :-) I think dselect, especially in combination with the apt access method, is terrific - it just takes some time upfront to get used to it. Waay back when, I used Slackware. Then, one of my users suggested that I try Redhat. He raved about the ability that RPM files give you to upgrade your system automatically. So, I tried it. I installed Redhat on a system and it worked okay, so I installed it on another. On the second system, I chose not to install X initially. Later, I changed my mind and went looking for the text-mode package selection utility so that I could install X. I couldn't find one. Now, I KNOW that there is one (for the install runs such a beast) but I couldn't find out what it was called or determine any references to it. The only Redhat package selection utilities I could find were X based. Since I didn't have X installed (installing X, after all, was why I was trying to find a text mode installer) I wound up having to run RPM for each of the parts which meant I had to figure out what all the parts were. It was not a fun process. However, there were other distributions and one of the CD-ROMs that I had had a copy of Debian on it. Debian doesn't force me to install X on my computer. Even if I don't install X initially, I can still select and install packages on my computer. So, I freakin' LOVE dselect. Don't ever get rid of it. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Std. Serial ports beyond 115K?
On Tue, 23 Feb 1999, Dimitri Patakidis wrote: I want to connect an external ISDN to my system using the MB's standard serial ports. However, everything's telling me I can only push 115K through these serial ports - far less than the theoretical max of an ISDN TA that does V.42bis. Any comments/suggestions on overclocking ;-) built-in serial ports to 230 (or 460) Kbps? How good are you at making changes to multilayer PC boards? The limitation of 115.2Kbps is a limitation of the bit-rate clock on PCs. You cannot go faster than this without having a higher bit-rate clock. However, with a higher bit-rate clock you can get speeds of up to 1.5Mbps. The problem is that you can't change that clock without modifying the circuit that generates it, which is kind of tough if the serial port is built in to your motherboard. Note: I realize that HS serial ports/boards are available, but I don't want to shell out the money if I don't have to Note2: If you can recommend a good external ISDN TA (North America) or an internal that supports DialD, I'd love to hear from you. The best way to get the most bang for your ISDN buck is to use an ISDN router. ISDN IN --- ETHERNET OUT No RS-232 translation in between. Speaking as the owner of an ISP that has to support this sort of thing, I'd like to encourage Mr. Traas to get a router. It's a whole lot easier to make a router work than an ISDN TA. However, if he doesn't want to spend the $50 to use a high-speed serial port, he sure isn't going to want to spend $500 extra to get a router. My recommendation would be to look for someone who is using an old, say, Pipeline 25 that wants to move up. I believe my company got one for $100 six-eight months ago. That's less than a decent ISDN TA is going to cost and it does routing. Not that I'm wedded to Ascend equipment, or anything, I just have more experience with them than other things. You can also use the Netgear or the Cisco or the 3com. I would recommend you stay away from the Danube and its ilk. They are not pleasant devices to set up. You'll need to do some work to find used equipment, though. Ebay is a definite option, here. Drop me a private line and I'll give you the Web addresses of a couple of people that deal in used equipment that you may not be aware of because they mostly sell to ISP's. You will need to get something that knows how to do NAT (like masquerading, only different) if you want to use it with a regular dial-up account. (With Ascend gear, it's mostly a matter of making sure you have a recent enough firmware revision, and you can update the firmware for most models. I don't know about the others.) -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Std. Serial ports beyond 115K?
On Tue, 23 Feb 1999, Kevin Traas wrote: My big thing is that I want/need Dial- Bandwidth-on-Demand support - very big requirements. I don't think these little ISDN routers provide the level of control that I want. (If you know different, please let me know.) All of the little ISDN routers I've ever worked with support those. That's basic functionality. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Apache and public_html directories
On Tue, 23 Feb 1999, Paul Miller wrote: Is the file /home/mark/public_html readable by the user or group that Apache runs as? I set the home directories of my users up to have permissions of 710 and membership in the group that Apache runs as (wheel on my main computer, www-data on every Debian installation I've ever seen) and the public_html directory has permissions of 775. This means that Apache can get into everybody's public_html directory, but no users can. This does not have to be set up this way. The permission on the users home directory should be a minimum of 701. The public_html directory the same. The files within the public_html directory need to be a minimum of 604. I'm not sure what you're objecting to, but I definitely don't agree with your conclusions. 701 gives everybody search access to the directory which means that if a user happens to know or can guess the location of a file farther down (like /home/id/public_html/index.html or, much worse, a file ending in .htaccess) and that file happens to be world readable (which 604 does) then that file can be read by anyone who has shell (few people) or ftp (many people) access to my system. Not a good thing. By giving the execute access to a privileged group to which the Web server belongs, I have a system that is much more resistant to tampering. Also please note that there is no effective difference between the 711 and 701 permissions or between the 644 and 604 permissions since the access given to the group in the former cases is the same as that given to everybody in the latter cases. I personally do not feel comfortable giving users membership in the wheel group as I have always believed it to be researved for those granted some super-user rights. With the above file permissions, you can leave home directories and such at the default group, or whatever grouping scheme you wish to use. Why would I give users membership in the wheel group? Most of my staff doesn't have membership in the wheel group! All I said was that the Web server is running in the Wheel group. As I stated in my message, the purpose of this particular configuration is to keep users out of each other's directories, so putting them in a group that has any access to anyone else's home directory (or setting the home directory so that everyone has some access, as I explain above the two are equivalent) is a bad idea. On the contrary, if I put the home directories in a privileged group and give that group only limited access to those directories then my users' directories are as safe as I can make them without enforcing some kind of permissions policy upon them. (One can never trust users to do the smart thing.) The Web server can get in, but, since it doesn't have read access, only if it knows the path it wants to reach, which is of course public_html and the subdirectories underneath. If ordinary users don't have any search permissions, it doesn't matter if they have privileges underneath or not, they can't get to underneath. The superuser can get in, but root can always get in. Preventing that requires rethinking the concept of root. My question about the permissions on the files was only to verify that the Web server had permission to read them. If that isn't the case, then it will return the error that was given. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Web email packages question
Quoting Randy Edwards [EMAIL PROTECTED]: What packages are available that would allow me to get to my email from the web server on my system? Run -- don't walk! -- to http://www.tdyc.com/~rkrusty/Debian/ and grab Ivan Moore's IMP setup. When I did an overview of all the web-based systems IMP stood out as at least one of the best, if not the best of the bunch. Ivan has a really slick deb created of it. As a result of your message, and at the request of one of my users who wanted WWW access to his email account, I have installed IMP. While it seems to work okay for most application, it has problems with mailboxes that have large numbers (somewhere between 500 and 700 messages) in them. To be sure, those problems seem to be with php3 rather than imp (I got segfaults on the Web server process before reducing the size of my inbox) but that's something to be aware of. (BTW, I'm running it on a copy of Stronghold 2.4.1 that I patched so it could use PHP-3.0.6. The amazing part is that it all works.) - This mail sent through IMP: http://web.horde.org/imp/
Re: GRiD install
On Sun, 21 Feb 1999 [EMAIL PROTECTED] wrote: I've been trying to install linux on a grid laptop and when it probes the ide controller ( I think, its after the line starting with hda, identifying the harddrive, one more line comes up then) the system reboots. Anyone know how to fix this? This sounds a lot like the problem I had with a Pentium motherboard I tried to use recently. The problem turned out to be an old BIOS that (as I recall) wasn't properly initializing the keyboard controller or some such. I upgraded the BIOS and it started working. This was a desktop computer, but that sort of thing might be something to look at, anyway. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: Apache and public_html directories
On Sun, 21 Feb 1999, Mark Ciciretti wrote: How do I set up public_html directories for users? I installed the 1.3.4 package of apache and apache-common. I then created a public_html directory under /home/mark/. When I enter the URL localhost/~mark/ I get Forbidden You don't have permission to access /~mark/ on this server. Is the file /home/mark/public_html readable by the user or group that Apache runs as? I set the home directories of my users up to have permissions of 710 and membership in the group that Apache runs as (wheel on my main computer, www-data on every Debian installation I've ever seen) and the public_html directory has permissions of 775. This means that Apache can get into everybody's public_html directory, but no users can. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: compiling debian source
On Mon, 22 Feb 1999, Stephen Pitts wrote: Specifically, I want to recompile VIM to use the X version, but I don't want to upgrade to Slink and have to do the job all over again. There must be a safe 'Debian' way? I had a look at dpkg-source but that didn't help much. This might not answer your question..but: What's wrong with VIM as it is? On my system, gvim brings up an X version of VIM. Like the man said, (well, implied,) it doesn't work with hamm. Not on this computer, anyway. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA
Re: sybase rpm
On Wed, 17 Feb 1999, Daryl Williams wrote: has anyone tried installing the sybase rpm package? we are needing to install a professional database and i would love to put it on a debian system if possible. if anyone is willing to share their experiences doing this i am willing to learn Just out of curiousity, what is so magical or professional about sybase? -- Jonathan Guthrie ([EMAIL PROTECTED]) Happy PostgreSQL user.
Re: lilo
On Wed, 17 Feb 1999, Chen Xu wrote: When I installed FreeBSD I installed boot manager. So when I boot, I can have chices F1 for DOS and F2 for FreeBSD. And F5 for second disk, which is not working yet. Question: How can I configure Lilo to boot Debian directly from HD? Will the installation remove the boot manger I already have? Is there any way that I configure my boot manager(boot easy) to boot Debian from HD without installing Lilo? If you're trying to do something complicated, you need to know what you're doing. I just skimmed your message so my information may be incomplete or incorrect. You have been warned. The approach that is normally taken to have lilo coexist with a boot program is to install lilo in the boot sector of the boot partition and then tell whatever booter you're using that that partition is bootable. Do NOT put LILO in the master boot record for the disk (unless you're installing Linux on the second disk and it's needed by FreeBSD's boot manager for some reason) as it will overwrite the FreeBSD's boot manager. Works like a champ with IBM's Boot Manager and System Commander. -- Jonathan Guthrie ([EMAIL PROTECTED]) Brokersys +281-895-8101 http://www.brokersys.com/ 12703 Veterans Memorial #106, Houston, TX 77014, USA