From owner-freebsd-doc Thu May 23 09:27:59 1996 Return-Path: owner-doc Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id JAA26173 for doc-outgoing; Thu, 23 May 1996 09:27:59 -0700 (PDT) Received: from host.igs.net (root@host.igs.net [206.248.16.2]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id JAA26118 for ; Thu, 23 May 1996 09:27:36 -0700 (PDT) Received: from ttyA11.ott.igs.net (ttyA11.ott.igs.net [206.248.17.47]) by host.igs.net (8.7.4/8.6.12) with SMTP id MAA20596; Thu, 23 May 1996 12:27:16 -0400 (EDT) Message-Id: <199605231627.MAA20596@host.igs.net> From: cskinner@bml.ca (Chris K. Skinner) To: csdayton@midway.uchicago.edu, doc@freebsd.org, support@cdrom.com Subject: Re: FreeBSD 2.1 Documentation and Installation of "Everything" to 2.1 Gig drive. Date: Thu, 23 May 1996 16:27:12 GMT X-Newsreader: Forte Free Agent 1.0.82 Sender: owner-doc@freebsd.org X-Loop: FreeBSD.org Precedence: bulk At 12:25 PM 5/22/96 CDT, "Soren Dayton" wrote: >Hii [snip] >>file extensions / file types How can ya tell what these >>header codes mean? ls -la|more will show if it has an executable file mode bit set, but little else can be gleaned from it. I know I said that I had little unix in my brain, but I know SCO had a file header and a header check utility from which file type info could be determined. I dunno if this is the same with FreeBSD? If there is such a thing what is its name? [snip] >>Sysinstall comments... >you need a lot more site specific information for DNS than you can >sensibly do in sysinstall. Same for the others. I have been able to put together a buggy DNS database and configuration that I think works? an nslookup finds all of the hosts that I need, but it comes up with an initial message saying something about the default NS: ------------------------------------------ unix# nslookup cskinner Server: unix.bml.ca Address: 0.0.0.0 Name: cskinner.bml.ca Address: 10.0.0.2 unix# ------------------------------------------ The address of unix.bml.ca when a lookup is done is found at 10.0.0.1 in its configured address. Is the address that's coming out above 0.0.0.0 a normal part of the display or is my setup still a bit out of kilter? Mostly from the 5 to 10 files that I threw together to fashion the DNS 'named' configuration in /etc/namedb directory I think that if I knew more unix and less Dos/Windows crap that I could make a program that could ask the user the following questions 1. Do you wanna make (a) an internet connected host or (b) a local tcp/ip LAN configuration. 2. If (b), input the size of the handfull of stations that you want to connect with your host (1-5) Others may be added by extending the generated configuration file. 3. If (a), do you want to make this a primary or secondary DNS? (P or S) 4. If (a) continued, If (P), then .... else (S) .... I had access to my internet provider's configuration files and saw nothing too complicated. My Linux Net Admin's Guide explains some stuff but not in detail, but maybe I haven't read far enough yet. I haven't seen the $ORIGIN directive explained in there, for example, but this may be BSD related and not part of Linux. This is an ls of my /etc/namedb directory: --------------------------------- bash# cd /etc/namedb bash# ls -la total 45918 drwxr-xr-x 2 root wheel 512 May 21 19:56 . drwxr-xr-x 8 root wheel 1536 May 23 10:14 .. -rw-r--r-- 1 root wheel 394 Nov 16 1995 PROTO.localhost.rev -rwxr-xr-x 1 root wheel 44042321 May 19 16:27 com.zone -rwxr-xr-x 1 root wheel 524234 May 19 16:22 edu.zone -rw-rw-r-- 1 root wheel 472 May 22 00:50 localhost.rev -rw-rw-r-- 1 root wheel 562 May 21 13:26 localhost.rev.BAK -rw-r--r-- 1 root wheel 695 Nov 16 1995 make-localhost -rw-r--r-- 1 root wheel 766 May 22 00:58 named.boot -rwxr-xr-x 1 root wheel 1658 May 21 11:59 named.ca -rw-r--r-- 1 root wheel 510 May 22 01:00 named.hosts -rw-rw-r-- 1 root wheel 530 May 22 00:51 named.rev -rw-r--r-- 1 root wheel 2121 Nov 16 1995 named.root -rw-r--r-- 1 root wheel 2121 Nov 16 1995 named1.root -rw-rw-r-- 1 root wheel 564 May 21 11:53 named2.boot -rwxr-xr-x 1 root wheel 2119 May 21 14:55 named2.root -rwxr-xr-x 1 root wheel 2354211 May 19 16:23 net.zone -rw-r--r-- 1 root wheel 394 May 20 00:07 proto.bak -rwxr-xr-x 1 root wheel 1658 May 19 16:25 root-ser.ca bash# --------------------------------- You'll notice that I got a copy of the com.zone and other files from internic a few days back and the big one is 44 Mbytes! Some files like named1.root are backup files or examples from elsewhere. Any idea if these *.zone files will ever be used by my system as part of the caching scheme?--or should I dump 'em? >> 2. Does one edit each config file or is there a config utility >> program to manipulate the file? >vi..... note that section 5 of the manpages is on file formats. I was using vi! Maybe I shouldn't have been messing with the configuration files though?-- but I had to in order to get anything working half-way! My point is that if a program were to manage some of this configuration manipulation, it would be able to locate the config file in whatever directory, and then manipulated it (to a certain degree of competence a la MS-windows) so that I could begin my programming tasks rather than wrestle with the configuration issues of one flavour of an operating system. >> 3. For 2 or more example installation complexity levels, what are the >> key elements to understand about the each type of config file >> on the system? Give at least a beginner, and an intermediate example. >I think that this is too broad of a question Why? There are probably only a finite number (20 or fewer?) key configurable items that do not require the user to re-compile the kernel that could be considered like anything else on a PC running whatever operating system. It's just like using a word processor or using a desktop publishing package--once you've become reasonably advanced on one flavour, all you need to know is what these other guys call the same feature on their competing word processor or DTP package. One can look for the common elements. Eg, serial ports, modems, terminal emulators, ethernet cards, hard disk partitioning, disk caching, cd-rom drivers, scsi drivers, display drivers, video capture devices, scanners, laser measurment devices, pcmcia slots, tcp/ip host drivers, tcp/ip client applications, other widgets and their drivers. If a person has already dealt with it somewhere else, then they're ready for whatever info is required for their setup in the new environment. The setup, if it defaults to something too beginner'ish, then it will not help them when they get in there editing those config files--they've got to look lots of stuff up and add it using the correct syntax, whereas, if another, intermedate level config template was also provided, they only might have to delete a few things or set some YESs to NOs. Config template A is a setup for simple 5 node LAN with tcp/ip. Config template B is for an internet connection on ethernet where the machine being configured is the name server for a small domain. These two configurations must be very common. The adding or subtracting of a couple of client host names with the vi editor then becomes less grief! >> 4. What are the differences between X-windows and MS-windows? >X is a protocol for communicating information about windowing systems >between two machines. MS-Windows in a window manager. The layer boundaries of the client / server for X-windows are distinct from what you are saying. But when the user is poking around using X on a machine that is running both the client and the server for his X-windows activities, then he's now interested in its functionality for launching applications and inter application window object data transfer--can I get my contact manager database to provide data to my report writer application using some kind of common data passing mechanism that does not involve RFC first principles? Can the same data passing be done manually via some kinda clip-board concept? [trying to be amusing:..] Whaddya press ta get this thing to go?! [snip] >> ... What steps can be made to bridge any deficiencies >> that X-windows might not have over MS-windows like file >> type associations, >those are there. They are the first couple of bytes in the file. That >way you do not need reserved names (although they still get used for .c >.o, .s, .f, etc. Where is the database that links the file type header codes with the installed applications that use these data/document files? Where is the mechanism for dragging a file icon into a folder to suddenly make it a launchable application? [trying to be amusing again:..] How d'ya make this thing go?! >> OLE2, TTFonts, DDE, NetDDE, filesharing, icon >> files in a program manager or file manager user interface >> shell program--I dunno, I'm no expert--other stuff? >those can all be done in external programs (like the window manager can >be done in an external program. actually _has_ t be done in a an >external program. I recommend ctwm. It would be nice if the How-to section of the documentation for Dos users was renamed to ...For Dos and MS-Windows Users and expanded to answer a bunch of my questions... >> Microsoft has steadily improved its installation mechanisms so that >> applications are easy to install. I forgot to also say that apps / packages are easy to uninstall too! >but still impossible to use and broken :) Microsoft's code is about as buggy as anyone else's. It depends on testing and quality control and the political leanings about whether something is a bug or a disabled feature (denyability sp?). >> INI files and system registration >> database file(s) are in the Windows directory and aren't scattered >> all over the place. >in most unices, they are in /etc. and that is _all_ that is in /etc. >(well, siome scripts too). Some of the startup scripts or files pointed to config files like /etc/namedb/named.boot and I found the one smb.conf file in the /stand directory. That samba is another thing... =^S that I'll have to pour over the man pages about. Seems that there are files somewhere? that tell it the list of workgroups and users that are permitted to connect with it. Then there is kerberosIV... The directory is there but no files therein? I understand about the export restrictions!--and why they can't be circulated on the distribution CDs! I got some of that kerberos stuff from MIT athens ftp site, but the documentation files compressed seemed to be about 820 kbytes so I'm not sure what's next. The source code compressed with gzip seems to be about 1000 kbytes though, so it might be managable. More later on... >> I hear that there is a 5 volume set (one with CD Rom(s) and one >> without). The cost is US$150 so I was quoted CDN$220 by a >> bookstore here in Ottawa. Anybody got any comments about the >> set--do they go out of date too quickly to be worth the >> money? > errrr... they _cannot_ go out of date. They define BSD 4.4. there >will never be another release of CSRG's OS. they are worth the money >simply because they are complete documentation and papers (academic >papers) on the OS. >Soren Hmmm... maybe the manual set is not exactly what I'm looking for... I wanna know what to type into the computer to make it go! and want some small understanding of what each thing I'm typing in does so that I might be able to guage if it will have the right effect that I'm going for. When I do a xstart, where do I go from there? A login and a couple of xterm windows pop up along with an analog clock. I'll have to read about the manager you talked about. Does it come in the distribution?--or do I get it from an FTP site? (I'm a gumby... I'll look around for it, and if its there, I'll find it!) But until you pointed it out, I wouldn't know of its name or what it does! Can ya recommend any good X-window beginner to "expert- developer" kinda books? I speak C++! I downloaded Netscape 2.02 for FreeBSD/BSDI, and ran it. It goes, but only after spitting out 3 pages of warning messages that it can't understand some keymapping. I compared its XKeysym--whatever file to the existing one and they were almost identical except for the release date in the file, so I used theirs and backed-up the old one, and still the warnings happen. Due to all of this key mapping missinterpretation, the backspace/delete keys and others do not seem to work. Is it because I've got to pour through the man pages some more to see what other configuration files I've got to put in the right directories and tweak with? Getting back to samba... What example config files exist for it? Where are they now? Where should they be? Should they be renamed to some standard name? What associated configuration files do I have to set up to make it go? Are there examples-- where are they now? Where should they go? Do they need to be renamed to something standard? Isn't there some way that this user database info could be automatically gathered by samba doing queries from time to time on the LAN to ascertain the users, workgroups, domains and so on. Have each user set as disabled until further action, and then a simple editing session is all that's needed to flag one or more of the users as authorized? Thanks for reading everybody. Hope that I get pointed in the right directions soon! Regards, Chris K. Skinner.