Re: Social Contract
%% Mike McCarty [EMAIL PROTECTED] writes: mm Isn't that one of the claims of most people who support the use of mm the GPL? That, since everyone just labors on it for love, or mm whatever, and that the source is available, then the quality will mm be better? I don't know about what most people claim, but that's not part of the manifesto of the FSF or reason behind the GPL. I haven't read this entire thread, but let me say that the primary goal of the FSF, and the reason they created the GPL, and the reason they USE the GPL, is in no way, shape, or form about making better quality software. That may, or may not, be a side-effect. Also, the FSF and GPL also aren't concerned with maximizing the freedoms of any particular individual; if they were then obviously public domain is the right way to go. What the GPL is all about is maximizing the amount of available free software (where free is defined by the traditional freedoms to examine, modify, and redistribute, as discussed on the FSF's web site). That goal means that some individual freedoms are not available, but this isn't uncommon: there is often a trade-off between individual freedoms, and freedom of the group in general. mm So, the GPL tells me what I can do with something I purchased. It mm tells me how I can dispose of it, and under what conditions. And mm discourages me from making modifications to it, because then I've mm got to release that, too. If the fact that you don't want to distribute your modifications when you distribute the result of your modifications is a discouragement to you, then definitely the GPL is not for you. The people who choose the GPL are doing so for a very specific reason: They are not totally altruistic. They want something in return for their work that they provide you. It's not necessarily money: instead it's that you contribute any enhancements you make to their work back into the commons. If that quid pro quo is not to your liking then you should definitely stay away from the GPL. mm Oh, and I've got to assign the rights to the Free Software mm Foundation. That's a primary point in the GPL. Because otherwise mm the FSF and you and whoever cannot get standing. You might mm investigate that part of it. That's totally, absolutely untrue. Not even close to being true. IF you modify a program where the FSF is the copyright holder (and there is far more software under the GPL where the FSF is not the copyright holder than otherwise--the Linux kernel for example), and you want to contribute your changes back to the FSF, then yes, the FSF will ask you to assign your copyrights before they accept the changes. This is so there is one unambiguous copyright holder for the entire software package. When you do this, the FSF will send you an agreement giving you complete freedom to use all the code you contributed in any other way you like, under any other license you like (not anything derived from the GPL, but if you extracted out your code and only your code). If you don't want to do this you STILL have an alternative: you can fork the project and provide that version (still under the GPL of course). The copyright assignment has exactly nothing to do with the GPL. It's solely a bookkeeping/legal protection procedure that the FSF has instituted before it will accept changes into _ITS_ software. mm If the fit is good, then fine. For me, the fit is not good, so I mm don't use it. For people who try to make a living writing mm software, who are not members of the idle rich, and who cannot mm afford to donate a significant portion of their lives to giving mm away software it generally is not a good fit. One part which makes mm this a bad fit is that anything which the GPL touches it invades. You are looking at this incorrectly. The FSF isn't against anyone making money. There are many ways to make money on software that does NOT involve using a proprietary license. The GPL can actually _HELP_ you make money. Why do you think the MySQL folks, the Qt folks, etc., release their stuff under the GPL? -- --- Paul D. Smith [EMAIL PROTECTED] Find some GNU make tips at: http://www.gnu.org http://make.paulandlesley.org Please remain calm...I may be mad, but I am a professional. --Mad Scientist -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Social Contract
%% Mike McCarty [EMAIL PROTECTED] writes: What the GPL is all about is maximizing the amount of available free software (where free is defined by the traditional freedoms to examine, modify, and redistribute, as discussed on the FSF's web site). That goal means that some individual freedoms are not available, but this isn't uncommon: there is often a trade-off between individual freedoms, and freedom of the group in general. mm We certainly agree there, except I would omit the phrase traditional mm freedoms and substitute privilege and add restricted privilege mm before redistribute. I do not agree at all with those changes. I suspect the fact that we disagree on this wording is an excellent indicator of our respective positions. The people who choose the GPL are doing so for a very specific reason: They are not totally altruistic. They want something in return for mm I'm leery of imputing motives to people I don't know. And I'm leery of imputing stupidity and/or laziness to people _I_ don't know. mm Oh, and I've got to assign the rights to the Free Software mm Foundation. That's a primary point in the GPL. Because otherwise mm the FSF and you and whoever cannot get standing. You might mm investigate that part of it. That's totally, absolutely untrue. Not even close to being true. mm No, no, it is. It is not. Please quote one sentence of the GPL, or even the rationale, that supports your position. If it is indeed a primary point in the GPL, it should not be difficult to find. IF you modify a program where the FSF is the copyright holder (and there is far more software under the GPL where the FSF is not the copyright holder than otherwise--the Linux kernel for example), and you want to contribute your changes back to the FSF, then yes, the FSF will ask you to assign your copyrights before they accept the changes. This is so there is one unambiguous copyright holder for the entire software package. mm And that's what I meant. The only person who can know what you meant is you. What you actually WROTE, however, was quite inaccurate, to the point where I can only consider it FUD. Unintentional perhaps, but FUD nonetheless. You are looking at this incorrectly. The FSF isn't against anyone making money. There are many ways to make money on software that does NOT involve using a proprietary license. mm Umm, do you presume to speak for the FSF? In private e-mail back in mm 1986 or so I discussed Richard Stallman's goals with him, and his mm goal, AIUI, is that people should *not* make money off of writing mm software. I can say with certainty that your understanding of FSF's goals is incorrect. I can't say what RMS's goals may or may not have been back in 1986, but I'm personally quite confident that he never intended to keep everyone from make money writing software. I don't speak for the FSF, obviously. However, I have read the many statements of their goals posted on the fsf.org website and their position on this subject is quite clear. Perhaps you could point to a statement which supports your claim? mm If I understand him properly, he disbelieves in any form of mm intellectual property. But, since he lives in a world which is not mm to his liking, he uses the intellectual property laws to try to mm reshape it as closely as he can to a world where people cannot mm make money merely by writing and selling software. No. Again, you assume that selling software under proprietary license is the only way to make money writing software. This is a false assumption. The GPL can actually _HELP_ you make money. Why do you think the MySQL folks, the Qt folks, etc., release their stuff under the GPL? mm Huh. You like to speak for others, I guess. I don't have any idea mm why they do that. Have you had conversations with them? How would mm you know? Because I've read their mailing lists and their web sites, where they explain it. They release the fully-featured version of their code under the GPL. This allows any other software that is released under the GPL or a GPL-compatible license to use it. They also say, if you want to develop a proprietary program using our software, come to us and we'll sell you a license to use it in ways that the GPL does not allow. If they released their code as public domain, or using a license such as BSD or even the LGPL, they obviously would not be able to do that: those companies could use their code in their proprietary products and would not need to pay for it at all. -- --- Paul D. Smith [EMAIL PROTECTED] Find some GNU make tips at: http://www.gnu.org http://make.paulandlesley.org Please remain calm...I may be mad, but I am a professional. --Mad Scientist -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject
gprof with -lc_p: anyone able to do it?
I wanted to do some profiling on GNU make, so I installed the libc6-prof package (I'm using testing) and built make like this: $ make CFLAGS='-g -pg' LDFLAGS='-g -pg' LIBS=-lc_p But this is very unhappy: $ ./make Segmentation fault :( :( GDB says: Program received signal SIGSEGV, Segmentation fault. 0x08099133 in _dl_lookup_versioned_symbol () (gdb) bt #0 0x08099133 in _dl_lookup_versioned_symbol () #1 0x40008385 in ?? () #2 0x400029c9 in ?? () #3 0x4000bc29 in ?? () #4 0x4000175a in ?? () #5 0x400016cb in ?? () Not very helpful. Note that if I leave of the C profiling library (no -lc_p) then it works fine and I can generate profiling output... but when I do that it only accounts for 18 seconds of a 3+ minute run! Am I doing something wrong here? Thanks! -- --- Paul D. Smith [EMAIL PROTECTED] Find some GNU make tips at: http://www.gnu.org http://make.paulandlesley.org Please remain calm...I may be mad, but I am a professional. --Mad Scientist -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Show postscript files in Landscape
I found a great answer on the ALE (Atlanta Linux Enthusiasts) mailing
list, posted in May by Joe Steele. Gotta love Google!
After reading his post, I simplified his suggestions slightly into this
handy-dandy shell function:
# Takes a PS file and generates a PDF, rotated 90 (e.g., in Landscape)
#
# Usage: ps2pdfr [input.ps [output.pdf]]
#
# If no input file is given, reads PostScript from stdin.
# If no output file is given, writes PDF to stdout.
ps2pdfr() {
pstops -w0 -h0 '1:0R(0in,8.5in)' $1 | ps2pdf -g7920x6120 - $2;
}
Hope this helps someone else!
--
---
Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools
Please remain calm...I may be mad, but I am a professional. --Mad Scientist
---
These are my opinions---Nortel Networks takes no responsibility for them.
Re: Show postscript files in Landscape
%% the Edward Blevins [EMAIL PROTECTED] writes: teb On Wed, Oct 10, 2001 at 11:54:01PM -0400, Paul D. Smith wrote: I have some PostScript files that show (in gv) in Portrait. However, they are slides and they are formatted in Landscape, so to read them on my screen I have to crane my neck over 90 degrees to the left :(. teb You might want to check out this page: teb http://rocky.wellesley.edu/downey/orientation/ Thanks, that was an interesting read (all my Google searches with pdf, postscript, landscape, convert, etc. didn't come up with this one). Unfortunately, none of the suggestions there helped me... I'm generating a PostScript file from OpenOffice Impress. I tried to add the new kind of slide paper to the ps2pdf config so I can use it to convert these PostScript files into PDF in Landscape (er, 11x8.5 :) but it absolutely didn't work. Both gv and acroread continue to display the resulting PDF file exactly as before: sideways. In fact, except for the creation date the PDF file generated by ps2pdf -sPAGESIZE=slide was exactly, byte-for-byte identical to the one generated without that flag. I know I changed the right file because if I use -sPAGESIZE=foo I get an error, but -sPAGESIZE=slide works fine. -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Show postscript files in Landscape
Hi all; this is a big off-topic but... I have some PostScript files that show (in gv) in Portrait. However, they are slides and they are formatted in Landscape, so to read them on my screen I have to crane my neck over 90 degrees to the left :(. Surely there must be some way to change the PostScript definition or something to rotate these files, so that when gv (etc.) displays them it chooses to automatically display them in Landscape mode instead of Postscript. I'm _sure_ I've seen Postscript files that worked in this manner. Really what I want to be able to do is then convert such a Postscript file to PDF and have acroread (etc.) _also_ display the results in Landscape mode. So far as I can tell the _only_ way to get this to happen is to boot Windows and use Acrobat Distiller or similar. Speaking in complete ignorance, this seems like a straightforward thing; is it really true that there is no way to perform this operation on Linux? -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
%% der.hans [EMAIL PROTECTED] writes: dh Am 13. Sep, 2001 Paul D. Smith so: I just asked on debian-devel. dh Did you get an answer? Yep. Apparently apt isn't being pulled in because it would break other packages, like aptitude or gnome-apt, which don't have suitable updates for some architectures. Until those tools that depend on APT have updates available for all architectures that allows them to go into testing, APT itself won't go into testing. I pulled it by hand and it only updated apt, apt-utils, and aptitude (once I got rid of gnome-apt), so that was OK. -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
%% Bob Nielsen [EMAIL PROTECTED] writes: Unfortunately, it still hasn't migrated into testing, and when I tried to grab it from unstable it wanted to pull too many other unstable packages for me to be comfortable with. bn What often works for me in a situation like that is to use: bn apt-get build-dep apt bn apt-get -b source apt bn and install the resulting package with dpkg -i. This will bn probably install some -dev packages you might not otherwise need, bn however, but you can often avoid massive upgrading. I thought of that. But then I took a closer look at the packages, and I apt-get remove'd the gnome-apt package (I don't much like it anyway--it _could_ be so great, but... it's not--yet, maybe) and after that the list of packages it wanted to update was much more reasonable, so I went ahead and did it. Thanks... -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
%% Jorge Santos [EMAIL PROTECTED] writes: Looking through the APT bug reports, I think this bug has been fixed in apt 0.5.4, uploaded to unstable around Aug 20. Unfortunately, it still hasn't migrated into testing, and when I tried to grab it from unstable it wanted to pull too many other unstable packages for me to be comfortable with. js http://ftp-master.debian.org/testing/update_excuses.html I'm stumped. According to the excuses, there's no reason for APT to not be moved to testing, as long as it doesn't depend on broken packages. Well, apt-search shows that the unstable version of APT (0.5.4) has these depends: Depends: libc6 (= 2.2.3-7), libstdc++2.10-glibc2.2 Now, on my testing system right now I already have libstdc++2.10-glibc2.2 installed, so that's no problem. And, I have libc6 2.2.4-1, which is definitely = 2.2.3-7. So, I have no idea whatsoever why APT 0.5.4 hasn't migrated into testing yet... -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
%% der.hans [EMAIL PROTECTED] writes: So, I have no idea whatsoever why APT 0.5.4 hasn't migrated into testing yet... dh Does it depend on apt or something that uses apt to get the dh correct results? No idea... dh BTW, for the workaround, just taking unstable out of sources.list dh fixed the probs. Didn't even have to do an update again to get rid dh of them. Yep, that's exactly what I did too: On a whim I just commented out the unstable lines in sources.list and re-ran apt-get upgrade (no update first), and it worked! -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
I just asked on debian-devel. -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
%% Jorge Santos [EMAIL PROTECTED] writes: js I think that, if it is a bug, it should be submited against apt, js since dpkg didn't seem to get to execute. Looking through the APT bug reports, I think this bug has been fixed in apt 0.5.4, uploaded to unstable around Aug 20. Unfortunately, it still hasn't migrated into testing, and when I tried to grab it from unstable it wanted to pull too many other unstable packages for me to be comfortable with. I can live with it the way it is for now. I can't remember the URL of the excuses page that tells you why something is not migrated into testing... -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: what command in linux such as mem in dos
%% thomas anderson [EMAIL PROTECTED] writes: ta is there a command in linux to show a more detailed information on ta memory usage and alternatively also cpu usage? currently I use 'ps ta aux' but I need more information... Tried top? There are also other, graphical programs. -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
apt-get problem workaround... is this a bug?
... -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: apt-get problem workaround... is this a bug?
%% Jorge Santos [EMAIL PROTECTED] writes: js Similar problem with my setup, thanks for the work around. Do you have the same setup WRT the apt.conf and preferences files, too? Maybe I will submit a bug. I'm not sure whether to file it against apt or dpkg, though... I guess dpkg. -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Quoting styles, cont (Was Re: Fonts in GTK)
%% Regarding Re: Quoting styles, cont (Was Re: Fonts in GTK); %% Eric G. Miller egm2@jps.net writes: but this practice is strongly deprecated. egm^^^ egm Hell does that mean? egm Webster's Ninth Collegiate Dictionary has this to say... egm dep-re-cate 1. to express mild or regretful disapproval of 2. egm DEPRECIATE egm I strongly mildly dissapprove of that quoting convention! Huh? Merriam-Webster's Collegiate Dictionary http://www.m-w.com: 1 a archaic : to pray against (as an evil) b : to seek to avert deprecate the wrath ... of the Roman people -- Tobias Smollett 2 : to express disapproval of 3 a : PLAY DOWN : make little of speaks five languages ... but deprecates this facility -- Time b : BELITTLE, DISPARAGE the most reluctantly admired and least easily deprecated of ... novelists -- New Yorker Webster's Dictionary (New Lexicon / Deluxe Encyclopedic Edition, 1988): v.t. To express disapproval of Strongly deprecated makes perfect sense. -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Fonts in GTK
%% J. Roger Jones [EMAIL PROTECTED] writes: I had a working GNOME install from Woody. I just updated to Sid, and everything seemed to go quite nicely, but now in GNOME, or rather anything that uses GTK, I see only a bunch of little squares where I should be seeing text. Can anyone tell me what I hosed, and what I need to do to correct it? j I had the same problem a few weeks ago. I used KDE2 for a while, j but didn't care for it. I tried researching the problem, but came j up with few references. I then came across the Debian TrueType j Mini-HOWTO, and installed xfs and xfstt, and then fiddled with the j FontPath setting if the XF86Config file (following instructions in j the mini-HOWTO). Needless to say, whatever I did works. Gnome, GTK, j and X apps are now readable. You don't need to use either xfs or xfstt when you're using XFree86 4.x, which is the version in both Woody and Sid. XFree86 4.x contains loadable module capability, and there are loadable server modules for TrueType fonts. Thus you do not need a separate font server. You might check out my TrueType info page: http://www.paulandlesley.org/linux/ -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Woody: problems upgrading lately?
%% Bill Wohler [EMAIL PROTECTED] writes: bw No problems here. If you already have a woody system, you don't bw need apt-get dist-upgrade. Just apt-get upgrade. Neither them work. Sorry, I should have said that. Since I've updated last, now if I try apt-get upgrade I get: # apt-get -s --ignore-hold upgrade Reading Package Lists... Done Building Dependency Tree... Done The following packages have been kept back cygnus-stylesheets debian-policy expect5.24 gsfonts gsfonts-x11 libxml-generator-perl 213 packages upgraded, 0 newly installed, 0 to remove and 6 not upgraded. Sorry, but the following packages have unmet dependencies: am-utils: dpkg-perl: Depends: perl5 Depends: libnet-perl but 1.0703-4.1 is to be installed libgtk-perl: Depends: libgtk1.2 (= 1.2.0) but 1.2.10-1 is to be installed Conflicts: libgtk-imlib-perl ( 0.7000) but it is not going to be installed mesag3+ggi: Depends: libggi2 ( 1:1.99.2.0b2.1) but 1:1.99.2.0b3.1-2.1 is to be installed Depends: xlib6g (= 3.3.6-4) but 4.0.3-4 is to be installed Conflicts: libgl1 pspell-ispell: Depends: libstdc++2.10 but 1:2.95.2-14 is to be installed task-c-dev: Depends: task-devel-common but 0.7 is to be installed E: Internal Error, InstallPackages was called with broken packages! I don't really understand some of these messages; for example libgtk-perl says it depends on libgtk1.2 = 1.2.0 but 1.2.10-1 is to be installed... well, 1.2.10-1 _is_ = 1.2.0, isn't it? Is that last line normal in this situation, or does it mean that something more sinister is going on? If I try apt-get dist-upgrade now it doesn't fail, but: # apt-get --ignore-hold dist-upgrade Reading Package Lists... Done Building Dependency Tree... Done Calculating Upgrade... Done The following packages will be REMOVED: am-utils dpkg-perl glutg3 gnome-control-center gs-aladdin gv kdebase-crypto kdebase-libs kdelibs3 kdelibs3-crypto konqueror libgtk-perl libkonq3 libqt2 libxml-dom-perl mesag3+ggi mysql-navigator pdl pspell-ispell pstoedit ssystem task-c-dev task-gnome-apps task-gnome-desktop task-tex task-x-window-system-core xbase-clients xf86setup xscreensaver-gl The following NEW packages will be installed: defoma dialog expect libxml-enno-perl The following packages have been kept back cygnus-stylesheets debian-policy 214 packages upgraded, 4 newly installed, 29 to remove and 2 not upgraded. Well, some of those I don't care about, like konqueror (I just wanted to try it out). But I can't lose am-utils, that's critical to my system! Likewise gv, pspell-ispell, and a few others. What's going on here? -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Woody: problems upgrading lately?
I've been doing apt-get update's for about a week now, on three different Debian systems which are all more-or-less recent testing systems, and none of them will let me do an apt-get dist-upgrade. I keep getting some packages listed, then this error: Error, pkgProblemResolver::Resolve generated breaks, this may be caused by held packages. I have no held packages on my system, though (I used --ignore-hold and it made no difference). I've gone through and tried to fix these problems by removing unused and older packages; often this changes the packages that are named in the error, but then another package gives an error. I have to believe that there's something a bit more fundamentally wrong here, but I don't know how to figure it out :( For example, on one of my systems: # apt-get -s --ignore-hold dist-upgrade Reading Package Lists... Done Building Dependency Tree... Done Calculating Upgrade... Failed Sorry, but the following packages have unmet dependencies: libgal4:E: Error, pkgProblemResolver::Resolve generated breaks, this may be caused by held packages. # dlocate -l libgal4 ii libgal40.5-7 G App Libs (run time library) # apt-get -s install libgal4 Reading Package Lists... Done Building Dependency Tree... Done Sorry, libgal4 is already the newest version. 0 packages upgraded, 0 newly installed, 0 to remove and 220 not upgraded. Before this it complained about libncurses4, blt4.0 vs. blt8.0, etc. Anyone have any ideas on what's going on here, or ways to investigate further? Thanks... -- --- Paul D. Smith [EMAIL PROTECTED] HASMAT--HA Software Mthds Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: TrueType Fonts
Look here for full instructions: http://www.paulandlesley.org/linux/ -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Building an SRPM on Debian?
%% Joey Hess [EMAIL PROTECTED] writes: jh Quelle strange. rpm wants a rpm database just to apply patches to a jh build tree? It must be on drugs. Maybe it wants to do source dependancy jh checking though. Does --nodeps help? Ah. That did it. Thanks. jh Could you send me the srpm in question? I'll see if I can fix jh this. It was the kernel-2.4.2-2.src.rpm, straight from the Red Hat download (I got it from rpmfind.net). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Building an SRPM on Debian?
I have an SRPM that I need to build on my Debian box. Really, I don't need to build it I just need to get to the patch phase (-bp) so that it constructs the proper source hierarchy and applies the patches (there are a lot of patches and they have to be applied in a certain order). I'd installed the Debian RPM and I used rpm -i to install the SRPM, which put the source in the SOURCES directory and the spec file in the SPECS directory just fine. But, when I try to run the rpm -bp I get this error: error: cannot open Packages index using db3 - No such file or directory (2) error: cannot open rpm database in /var/lib/rpm I created that directory and put an empty Packages file there, but I get: error: cannot open Packages index using db3 - Invalid argument (22) error: cannot open rpm database in /var/lib/rpm I don't want to do anything else with RPM, I just want to get this source set up correctly. It's annoying that I appear to need a Packages DB just for that, but I apparently do. Does anyone know how to construct a legal Packages file (I guess it can be empty--not contain any packages) on my Debian (woody/testing) box? Thx. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Man page output width
Back a long time ago when I used to know roff in general, and groff in particular, really well (at Data General I wrote a lot of docs using groff, and talked to James a good bit about it), I took the tmac.an macros and modified them to decrease the margin on the man pages displayed on the TTY. It is _SO_ annoying to lose at least 10 characters of perfectly good space on the right of every single line, to a quite useless right hand margin. Unfortunately, that knowledge has passed out of my ever-shrinking set of useful brain cells. I see that there's an /etc/groff/man.local file which is loaded after the normal tmac.an (or tmac.an-old or however that works--looks confusing). That seems promising. However, I've completely forgotten what magic incantation I need to put there to change the right-hand margin from 1 inch or whatever it currently is to something more reasonable... Hints, anyone? -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: NO! chmod strikes!
%% Dave Sherohman [EMAIL PROTECTED] writes: ds Changing it to: ds find . -name .[^.] -print0 | xargs -0 chmod r-owx Not to be anal retentive, but ITYM: ... -name '.[^.]*' ... Your glob won't match anything except .x where x is exactly one character long but != .. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: MUAs that compare with Outlook (your chance to show how much better Linux is than MS!!)
%% Kurt Lieber [EMAIL PROTECTED] writes: kl OK, I've read with great amusement all the chest-thumping going on kl about MUAs, MTAs and how Microsoft email products are things that kl you scrape off the bottom of your shoe. ... kl If there isn't, then I hope the person who stated anyone who uses kl MS email products is ignorant will reconsider their statement. I'm using Emacs+Gnus+Procmail and get all the power I want. I don't know if it can do everything you want or not; I don't see anything obviously missing. I more wanted to respond to these particular statements. The reason people here don't like MS Email products typically has little to do with how easy they are to use. It's also not true that we have to have actually used these products before abusing them. It may well be (I don't know, myself) that Outlook has the best, simplest to use user interface ever invented. That's not the point. The point is, Outlook etc. cares more about how easy it is for you to send the message. We care more about how easy it is for the _receiver(s)_ of the message to read and understand it. So, when we see Outlook send messages with broken MIME, with bogus default settings like Rich Text that more than doubles the bandwidth and storage capacity required for no reason, with quoting capabilities that defy every real and defacto standard developed over the years, etc., then we say that it is a crappy product and should be avoided and shunned, and we know we're right. And we never once had to run it ourselves. UNIX tools typically start with the basic premise that the underlying behavior must be correct, and user friendly bells and whistles can be added later (and this is starting to happen). It often seems that Microsoft starts from _exactly_ the opposite position. I think (and I think experience has shown this to be correct) that it's much easier to add a nice interface to a fundamentally strong base than it is to go in and fix up a broken base underneath a nice interface. To paraphrase the famous performance enhancement mantra, it doesn't matter how simple the product is to use, if it generates the wrong result. This probably doesn't make much difference in your search, but I thought I'd throw in my $0.02 anyway. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Who's got a cheap scanner working?
%% Robert Voigt [EMAIL PROTECTED] writes: rv On Monday 09 July 2001 22:40, Andrew Perrin wrote: I'd like to add a cheap scanner (probably USB) to my machine, running debian potato. I don't need to do anything fancy, but would like basic scanning to work. It looks like I can get more than enough power from something like the Visioneer Photoport 7700 or the UMAX Astra 3400. Can anyone share successful experiences with a USB scanner? rv I have a cheap parallel Plustek scanner working here. Most Plustek rv scanners are supported, but not in Sane yet. I had to compile the rv stuff myself. My Plustek is in SANE (heh :); when was the last time you checked to see if yours was supported? -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Kernel 2.4.5 at last! Problems
%% User zos [EMAIL PROTECTED] writes: uz IIRC in all the kernels I've compiled from 2.2.x and on you could uz always select ext2fs support or not. This makes sense as linux is uz just a kernel and should not be dependent on any particular file uz system, especially now with the advent of more robust filesystems uz (eg: ReiserFS, ext3). I think you miss the point Andon was trying to make: You _can_ select whichever filesystem type you want for your root partition (AFAIK). The rule is that whatever filesystem type you select for your root partition _must_ be built into the kernel. It can't be a loadable module. Consider: how can the kernel load the module that contains the code needed to access the filesystem, when it can't access the filesystem until the module is loaded? uz On Tue, 10 Jul 2001, Andon M. Coleman wrote: I've had this problem too... It appears EXT2 support is no longer built-in to the kernel... I've switched to ReiserFS since then, but you may want to look into re-compiling the Kernel with EXT2 support built-in (or whatever FS you're using). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: CUPS and LJ (was Re: LaserJet Plus and Samba)
%% D-Man [EMAIL PROTECTED] writes: d Will unzip work on .exe files? Yes. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Kernel 2.4.5 at last! Problems
%% User zos [EMAIL PROTECTED] writes: uz Hah. That's really funny. I didn't even think of that. I dunno, if uz I need a file system supported I just compile it straight into the uz kernel. While I realize the benefits of having loadable modules uz for various things, it doesn't make much sense to me to have the uz kernel load a VFAT module just so I can mount my windows uz partition, especially considering that it automounts it at boot. In that case, true. But, I only very occasionally mount my Windows partitions. Further, I rarely use my CDROM drive, or my floppy drive, or my sound card, or my parallel port. Since I don't use my CDROM much, I don't need the ISO9660/Joliet filesystems very often. So, I leave those things as loadable modules. On the other hand, my network card I always build into the kernel :). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: NO! chmod strikes!
%% Dave Sherohman [EMAIL PROTECTED] writes: ds On Mon, Jul 09, 2001 at 12:29:40PM -0400, Paul D. Smith wrote: BTW, the best way to do what you wanted to do is this: $ chmod -R o-owx .[!.]* dc Or even better, ignore the -R in the chmod command and use find: dc find . -print0 | xargs -0 chmod r-owx Not to be argumentative, but what's better about it? ds While I won't presume to judge which is better, there is a difference: Certainly there's a difference; my second sentence was: First, your example doesn't do what the OP wanted to do, or what my example does do. ds The chmod -R version will affect .foo/bar and ignore foo/.bar (it ds looks only at names in the directory where the command is issued) Yes, that was exactly the point of my second paragraph. ds while the find version will leave .foo/bar alone and change ds foo/.bar (it looks at each filename independently). Not true at all; the find version will change _both_ .foo/bar _and_ foo/.bar. find . matches everything in the current directory and below, _including_ hidden files. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: debian: mozilla-0.9.1 not as good as mozilla-0.8.1
%% Brendan J Simon [EMAIL PROTECTED] writes: bjs I have manually installed versions of mozilla on my PowerPC bjs laptop. Version 0.8.1 worked fine with my internet banking and bjs other java web sites. I was so pleased to see Mozilla-0.9.1 in bjs the testing distribution and promptly did an upgrade. I now can bjs not access any of my banking. I'm not sure if Mozilla-0.9.2 will bjs fix this. I hope so. It sounds to me like you haven't installed the Mozilla PSM (Pesonal Security Manager) module. Try apt-get install mozilla-psm -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: NO! chmod strikes!
%% Dave Carrigan [EMAIL PROTECTED] writes: dc Paul D. Smith [EMAIL PROTECTED] writes: BTW, the best way to do what you wanted to do is this: $ chmod -R o-owx .[!.]* dc Or even better, ignore the -R in the chmod command and use find: dc find . -print0 | xargs -0 chmod r-owx Not to be argumentative, but what's better about it? First, your example doesn't do what the OP wanted to do, or what my example does do. In order to do that you _still_ have to use my globbing expression (or else use some truly bizarre contortion in find--see other posters to this thread as they try to come up with one): find .[!.]* -print0 | xargs -0 chmod r-owx Second, your example has a lot more typing in it, and although anyone doing UNIX sysadmin needs to be familiar with this pattern, it's definitely more complex than -R. You _could_ have argued that the -R flag is not standard, and so the -R version won't work everywhere... but the -R is a lot more portable and works on a lot more versions of UNIX than the find -print0/xargs -0 flags, which _only_ work with GNU find and xargs. So... I'm not seeing it :). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: msttcorefonts package and X fonts?
%% Geoffrey Romer [EMAIL PROTECTED] writes: gr Does anyone know of a workaround to incorporate those fonts into X gr in the absence of defoma support in-package? I have already done gr the following: gr -ensured that the line gr dir /usr/share/fonts/truetype This isn't the path you want to add. Assuming you're using XFree 4.x, take a look at this doc: http://www.paulandlesley.org/linux/xfree4_tt.html and all will be made clear... If you have XFree 3.x, then there's a sibling doc there, somewhat more out of date, that will help with that. Have fun! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: NO! chmod strikes!
%% [EMAIL PROTECTED] writes: j 1. ( ) text/plain (*) text/html No need for HTML on the mailing list. j I tried, in a subdir of /root, the command j chmod -R o-rwx .* j It changed the permissions on the parent directory, j the parent's parent directory, all the way up. j Now only root can use my computer. j Was chmod supposed to understand .* so differently j than /bin/ls does? Both chmod -R and ls -R behave the same way. However, your description above is not actually what happens, and I haven't seen anyone else here correct it in so many words. So, just to be clear: Remember that the shell expands all wildcards, not the application. So, supposing you had these files/directories in your current working directory: . .. .foo .bar .biz baz boz Then, the command chmod -R o-rwx .* is identical (from the point of view of chmod, which only gets the postprocessed commandline) to having typed this: chmod -R o-rwx . .. .foo .bar .biz _Neither_ ls nor chmod (nor any other tool that recurses that I've ever heard of) will follow .. in a directory as it's recursing. However, if you give .. on the command line, of course it will operate on that directory just like it would any other directory on the command line. The upshot is, chmod -R o-rwx .. will change the permissions in the parent directory and all of its subdirectories, recursively; it will start from the directory above the current directory and walk down. It will _NOT_ walk back _up_ the tree any further up from the parent. Ditto for find, ls, diff -r, etc. etc. So, the what happened here depends on your working directory. If you ran the command in /root (or any other directory which is an immediate subdirectory of /), then the chmod started at / and your entire system got chmod'd and your system is relatively screwed. If you ran it from a deeper directory, say /root/foo, then only the files under /root were changed; that may be less terminal. BTW, the best way to do what you wanted to do is this: $ chmod -R o-owx .[!.]* That will change everything, recursively, beginning with a ., except . and ... Modern shells also accept the more standard RE format: $ chmod -R o-owx .[^.]* -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: TTF fonts
You don't have a fonts.scale or fonts.dir file in your fonts directory; I'm pretty sure you need to. If you are using XFree86 4.x and a testing distribution of Debian, you can use my doc to very easily install TrueType fonts, and you don't need a font server to do it (XFree86 4 has built-in support for TrueType fonts via its loadable modules feature). http://www.paulandlesley.org/linux/xfree4_tt.html There is also a version I wrote for XFree86 3.x, here: http://www.paulandlesley.org/linux/debian_tt.html This requires more steps but it _does_ work. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: testing packages not consistent?
%% [EMAIL PROTECTED] (Colin Watson) writes: cw testing ignores the Recommends: (it's installable without it), and cw resolves the Conflicts: by assuming that you can remove debhelper cw in order to install the package. It's not optimal in this case, cw but is necessary in others. OK. No problem; I just used apt-get -t unstable install debhelper and then alsa-source installed fine. I love the new apt features! :) -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: TTF fonts
%% Martin F. Krafft [EMAIL PROTECTED] writes: mfk also sprach Paul D. Smith (on Sat, 07 Jul 2001 06:07:40PM -0400): You don't have a fonts.scale or fonts.dir file in your fonts directory; I'm pretty sure you need to. mfk first, i translate to Type-1, got to stay native, you know: There's nothing non-native about TrueType as a format... ? At least, not moreso than other formats. All the TrueType capabilities in both the xfs and builtin solutions are free software, developed natively on Linux (most of them). And I don't see how turning Microsoft TrueType fonts into Type-1 makes them any more native than leaving them TrueType. It just makes them look worse :(. But, whatever, it's your monitor :). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
testing packages not consistent?
I just upgraded my system to all the latest woody stuff as of about an hour ago. Now, I want to install the alsa-source package so I can rebuild alsa for my system (I have an older version, 0.5.10b-6, installed now) But whenever I try to install it, it wants to remove some critical files from my system first! # apt-get install alsa-source Reading Package Lists... Done Building Dependency Tree... Done The following packages will be REMOVED: alien debhelper dh-make 1 packages upgraded, 0 newly installed, 3 to remove and 2 not upgraded. Need to get 1333kB of archives. After unpacking 634kB will be freed. Do you want to continue? [Y/n] n Abort. Uhm, excuse me, but I'm _using_ those! :) Turns out that alsa-source wants version 3.0.35 or better of debhelper: # apt-cache show alsa-source | less Package: alsa-source ... Version: 0.9+0beta4-5 Depends: debconf ( 0.2.26), make, gcc | c-compiler Recommends: dpkg-dev, kernel-package, debhelper (= 3.0.35), debconf-utils Suggests: devscripts Conflicts: debhelper ( 3.0.35) ... but the latest version in testing is 3.0.15: # dlocate -s debhelper Package: debhelper Status: install ok installed Priority: optional Section: devel Installed-Size: 357 Maintainer: Joey Hess [EMAIL PROTECTED] Version: 3.0.15 ... # apt-get install debhelper Reading Package Lists... Done Building Dependency Tree... Done Sorry, debhelper is already the newest version. 0 packages upgraded, 0 newly installed, 0 to remove and 3 not upgraded. If I want to install a better debhelper, I have to go to unstable. I thought the idea behind testing was that packages would only go into testing once all the packages that they needed were also in testing... How am I confused? -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: New TrueType fonts doc version 1.1
%% Walt Mankowski [EMAIL PROTECTED] writes: wm I haven't read your document -- you didn't provide a link :-) -- Doh! http://www.paulandlesley.org/linux wm but you might want to point people to wm http://www.microsoft.com/typography/fontpack/default.htm if you're wm not already doing so. wm That webpage allows you to download what Microsoft calls their wm TrueType core fonts for the Web. The downloads are free, and while wm IANAL the EULA seems to me to allow personal use on Linux. That font set is already packaged for Linux (testing/unstable), so you can install it with apt or whatever... and I do mention it in my doc. In fact, my doc deals quite explicitly with how to install it. Thx! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: blocky fonts in woody
Not sure what you used to have, so I can't really say why things got worse moving to woody (BTW, almost certainly the important thing here is that potato (2.2) used XFree86 3.3.x while woody has XFree86 4.x). But, you may find the following article interesting and/or helpful: http://www.paulandlesley.org/linux/xfree4_tt.html -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Why is setting up X so arcane?
%% Nikki Locke [EMAIL PROTECTED] writes: nl I ran XF86Setup, and chose the SVGA driver. I left everything else nl alone. It hung my machine solid (again). Are you _sure_ your machine was hung? That rarely happens. More likely you just can't use your monitor. Try using ALT-F1 or maybe CTRL-ALT-F1. That should switch you to the console you were using before you started X. Alternatively, you could use F2 instead of F1 to get to a different console and log in there. You can then reboot your system properly if you like, although you almost certainly don't need to reboot (this ain't Windows! :) Instead, use ps to find the X server process, then kill it: # ps -aef | grep /X ...make sure you get the right process: the PID is the second field... # kill 1234 # or whatever the PID was If that doesn't work (ps shows the process is still running), you can try kill -9 1234. Then try again with some of the other suggestions here. As others have pointed out, the version of X in Debian 2.2 is not as good at this as later versions are. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Why is setting up X so arcane?
%% Nikki Locke [EMAIL PROTECTED] writes: nl Good. Is it worth me upgrading to another version? The machine nl is there mostly to run samba, apache, apache-ssl and php4. I need nl a solid, stable ftp client which will work through a gateway, and nl the ability to ssh in from other machines on the network. If that's all you need, why install X at all? I don't have it on my firewall system, which does basically just exactly what yours is doing. Of course, this doesn't address your question directly. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Why is setting up X so arcane?
%% Nikki Locke [EMAIL PROTECTED] writes: Are you _sure_ your machine was hung? That rarely happens. nl Pretty sure. Ctrl-Alt-F1 didn't work. Ctrl-Alt-Del didn't work. A nl telnet session into the machine from elsewhere on the network nl stopped responding. nl It may be rare, but it happened to me. Twice :-( Sounds like you have one of those old pieces of hardware that are problematic. I've screwed around with X a good bit in the 8 years I've been using Linux, but it's never locked up my entire system. I even installed it on a laptop--_that_ was an adventure :) Sorry I can't be of more help, but you might try asking a wider audience, such as the comp.windows.x.i386unix or comp.os.linux.x newsgroups, to see if anyone else with this same card and/or monitor will send you an XFree86 3 config file for it. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
New TrueType fonts doc version 1.1
I uploaded version 1.1 of my TrueType fonts for XFree 4.x on Debian document. This version discusses more explicitly how to install TrueType fonts that are not packaged for Debian (for example, ones you copy from your Microsoft Windows partition). Note I have no idea or opinion on whether using those fonts in this way violates your EULA or not: you need to decide that for yourself. There are a few other touch-ups that people suggested. Thanks to my readers :) -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: TrueType Font Guide feedback
It's a doc that describes how to get TrueType fonts set up correctly on your Debian system. There is an older (and slightly outdated) version for Debian 2.2/stable with XFree86 3.3.6, and the new version I just posted for Debian testing/unstable with XFree86 4. There are some Debian packages for the actual fonts; unfortunately they don't do everything you need to get them working. This doc is that bit of glue. I'm hoping to have time to add more info to it on how to use non-packaged TrueType fonts; as of now you have to kind of infer a lot of stuff in that situation :-/. But, note that the Microsoft core fonts (like Arial, etc.) do have a Debian package (in testing/unstable). http://www.paulandlesley.org/linux/ -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: TrueType Font Guide feedback
). Any clarification of that would be helpful. My doc says: In XFree86 4, loadable modules functionality was added to the X server and among the loadable modules provided were two providing TrueType font support directly in the X server, without requiring an external font server (obviously if you need a font server for some other reason, you are still free to use one). Maybe this should be expanded. rb The whole X setup seems to have a lot of redundancy (e.g., *.dir rb and *.scale), which is asking for trouble. I don't know if there's rb anything you can say about it that would help, though. Well, the *.dir vs. *.scale is how the X setup is done. The *.dir files are used for regular, non-scaled fonts. The *.scale file is used only for scalable (TrueType, etc.) fonts. I'm not sure why there have to be two or why they have to have the same contents, though. rb Anyway, it might be worth letting people know they can fool with rb such things in, e.g., /etc/X11/app-defaults/Xman. Personally, I wouldn't change anything there. Again, that's a system file controlled by the package manager. I know it's a config file, so it won't be updated without asking first, but even so it's a PITA to keep updating that file with that kind of modification. I really avoid modifying any package-controlled files unless there's no other alternative. With X resources there are more alternatives than you can shake a stick at :). There's really no need to modify these config files. Personally, I create a $HOME/app-defaults directory, then in my shell setup I set: export XUSERFILESEARCHPATH=$HOME/app-defaults/%N Obviously if you use XDM or GDM or whatever you need this set in your session setup file as well. Now I can create a $HOME/app-defaults/Xman (for example) file, and every time I start xman it'll be loaded just like the system Xman resource file, overriding those settings. The nice thing about this method, as opposed to using .Xdefaults and xrdb -load or whatever, is that they're dynamic: you change the file and restart the app and you don't have to reload them into the X server. rb When I configured NS (Mozilla, actually) as described in part 7 the rb choices actually appeared as Monotype-arial black. Arial Black is a different font. You should have _both_ Arial _and_ Arial Black in the Monotype foundry. If you don't, you're missing some fonts or some aliases or some contents of your .scale file or something. rb Do you want to make a recommendation for the serif fonts, e.g., rb monotype-times new roman? Personally I don't have any real opinion. I rarely use serif fonts. Use what you think looks good :). rb A couple lines on exactly how and when to restart the font server rb and the X server would be good. I was concerned doing so would rb kill my session, so I just logged out and then back in. I'm also rb not sure whether, e.g., /etc/init.d/xfs reload is adequate, or if rb restart is necessary. I thought I mentioned this, but perhaps it was too brief. You can either restart the server (by logging out and in), _OR_ you can use xset to change your font path dynamically. They're equally good, so use whichever you're more comfortable with. My document doesn't deal with font servers at all, it's completely geared towards using the builtin font management not an external server. Therefore, /etc/init.d/xfs is irrelevant; I actually uninstalled the font servers on my system altogether. Thanks for the notes...! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: TrueType Font Guide feedback
%% Anthony Campbell [EMAIL PROTECTED] writes: ac In section 5.1 an extra has crept into the command: ac grep 'iso8859-1' fonts.alias msttcorefonts.alias Actually, that's not a typo. I meant to do that! :) If you leave out the then you get far too many fonts, because the grep matches iso8859-10, iso8859-11, etc. as well as iso8859-1. The anchors the end of the match. ac And in section 6.2 the font path should be: ac /usr/X11R6/lib/X11/fonts/TrueType No, this is the right path. On any properly installed X system (this is not just Debian, and not even just XFree86), /usr/lib/X11 is a symlink pointing to the current installed version of X's lib directory (in this case, /usr/X11R6/lib/X11 as you point out). But the correct, and documented, way to install fonts is via the /usr/lib/X11 link and not /usr/X11R6/lib/X11. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
TrueType fonts with XFree86 4 in Debian? Here's how...
In case anyone's interested, I installed another installment of instructions for getting TrueType fonts working in Debian. This version describes how to do it with XFree86 4. It uses the Debian packages I discovered (or which were added) since my last installment, which did the same for XFree86 3.x. You don't have to build anything by hand in this one; much cleaner than the old one. I also wrote it using the NewbieDoc stylesheet; not too bad. http://www.paulandlesley.org/linux/ Let me know if you have comments/suggestions/corrections/etc. Have fun... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Any Clearcase users on Debian?
%% Regarding Any Clearcase users on Debian?; %% Nico De Ranter [EMAIL PROTECTED] writes: ndr I'm trying to get Clearcase (huge, expensive code revision ndr system) to work on Debian but I can't get the mvfs module to ndr load. Has anybody tried this before? and succeeded? Please don't ndr tell me to revert to Redhat 6.2 :-( I've been using ClearCase 4.1 on Debian since it was released last November. Works fine. There are a number of issues to consider: 0) Be sure you have all the latest patches: most particularly patches 11 and 12. Before that, ClearCase on Linux (even Red Hat) was not so stable. 1) ClearCase only supports Linux kernels 2.2.14-2.2.16. To be precise, they only officially support the Red Hat versions of those kernels. I have used ClearCase with the Linux 2.2.17 kernel, but others have reported problems. ClearCase MVFS will _NOT_ work with any kernel prior to 2.2.14 (why would you want to? :) or later than 2.2.17 (the NFS changes introduced in 2.2.18 change the filesystem module interface sufficiently that MVFS won't work). I have used the vanilla kernels from www.kernel.org with ClearCase and they worked fine, but what I did to be extra-safe was go to the Red Hat site and get their kernel sources package for the Red Hat 7.0 kernel (2.2.16-22, in Red Hat-speak). This is 2.2.16 plus a lot of patches Red Hat applied. Remember that, as with Debian's kernel source packages, this is a regular RPM not an SRPM. Then I used rpm2cpio to unpack it, and built it and installed it using Debian make-kpkg etc. 2) You have to edit /etc/issue to fool the ClearCase installer into thinking that you are running Red Hat; add something like Red Hat version 7.0 to /etc/issue when installing--but be sure to take it out again after so people don't think you're really running Red Hat! :) 3) You must also apply Rational's small kernel patch as described in the ClearCase release notes, and install the MVFS code before you build the kernel. Also be sure the various kernel module settings are correct, as described in the release notes, and that you have picked the MVFS module to be built when you're in the kernel filesystems config menu. And of course you need NFS enabled. 4) You don't say what version of Debian you're using, but note that ClearCase wants GLIBC 2.1.3 (as shipped on Red Hat 6.2/7.0). I was using testing, and upgraded my GLIBC to 2.2.1 with no problems. However, last month when I upgraded to GLIBC 2.2.3, the ClearCase albd_server stopped working :(. It comes up fine and goes into a select() wait for connections, but no clients can connect to it. Very weird. What I did was to unpack a copy of the Debian potato libc6 2.1.3 DEB in a different directory, then play games with ld.so to have the ClearCase programs use that libc instead of my standard system libc. It took a bit of hacking but it works fine now. Let me know if you need more details. If you stay with Debian potato you won't have to mess with this. Finally, if you have an extra system lying around I strongly suggest you install Red Hat 6.2 or 7.0 on it, then install ClearCase there. If you run into any issues you should verify that they exist on the Red Hat system as well before reporting them as bugs to Rational: Rational does not officially support any version of Linux other than Red Hat. This system doesn't have to be anything special; just use that P100 doorstop you have lying around :). CheapBytes will sell you a CD with RH 6.2 or 7.0 on it for $5 or less. HTH. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Customizing the console key map?
%% Wayne Topa [EMAIL PROTECTED] writes: I already know how to do this; what I want to know is whether there is any way to do it in Debian, so that when I upgrade my Debian packages console-tools (or kbd, I don't care much which one) my customizations are kept automatically without my having to go back in and fix them up. wt I made a custom.map file, for slackware, from that article 5-6 years wt ago. I have used it, without modification, on Debian since bo. I am wt currently using it on potato, slink and woody boxes. I have never wt had to change it. There must be some reason why upgrading these packages keeps installing new versions of the keymap files. Look, for example, at how the X fonts are done: there is a directory /etc/X11/fonts/dir for each font type, and in there you can add your own files. For example, you could add a file /etc/X11/fonts/100dpi/my-fonts.alias and put your own font aliases in it; the system will automatically include and install those aliases for you when you run update-fonts-alias. This way, you don't have to edit a system config file to add your own aliases, which would entail re-editing those config files (or merging them or whatever) every time you updated to a new Debian X fonts package. wt My custom.map file is only for the console. I setup keys for X in my wt .xmodmap file. Well, of course. Your comment is a non sequitur. I'm talking about X fonts above solely as an illustration of a good method of handling user customizations. Since you bring it up, though, I'll point out that modmap is another example of good separation between system files and user customizations: you can write your own .xmodmap file as a separate file containing _only_ the changes to the standard key bindings. This is what I want to do with the console key map: have my own file with _only_ the changes, not the entire keymap. wt Maybe RedHat has a tool for that. They seem to be pretty good at wt making tools for helping new users. I don't use redhat as I wt prefer to understand what I am doing. Another non sequitur. What tools Red Hat may or may not have is not helpful to me since I use Debian. I was hoping there was some similar way to defined my own console key map customizations file as a separate file, and have the console-tools (or kbd) automatically install/append/whatever that file into its default keymap when it's created. wt Well I must not be understanding you correctly then. I made my wt console key map customizations file myself and don't know or wt want any file so general that it would modify my custom key maps. ?? I don't want anything to modify my custom key maps, of course--that's what it does now and what I'm trying to avoid. I want a way to customize the standard key map, without completely replacing it. This latter is what you seem to have done: if I'm understanding you correctly you don't have a console key map customizations file, you have a customized console key map file. The distinction is very important. Again, X modmap is a good example: you don't put the entire X keymap in your .xmodmap file, you only put the _changes_ there. That's what I want for the console: a file containing my changes, which is automatically loaded and overrides settings in the standard key map, rather than keeping an copy of the entire keymap myself with my changes embedded in it. This can be done at the time the keymap file is created, of course, unlike X modmap which is done dynamically. It has nothing to do with knowing the system or any other such stuff: I've been using Linux since 1993 (I used dd on my SPARC SunOS 4.1 workstation to create my first Linux boot floppy :) and UNIX a lot longer than that--I'm not _incapable_ of making these changes myself. Now that I'm an old fogy I just don't have time to fool with it; there are much more important things I want to do. That's why I moved to Debian from Slackware: I just don't have any urge to build everything myself anymore. I've done that for many years (and this was back before autoconf, when building Emacs and GCC and stuff was actually an adventure :)), I know how to do it, and it doesn't interest me any longer. I build locally and hack the packages I'm actively involved in, and let other people worry about the rest. But, I am willing to spend some time explaining what I want in the hopes that it will save me, and/or someone else, time in the future. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: ~/.Xmodmap (was: Customizing the console key map?)
%% [EMAIL PROTECTED] (John S. J. Anderson) writes: On 30 May 2001 22:08:33 -0400, Paul D. Smith [EMAIL PROTECTED] said: Paul Second, my point is, with so many truly _useful_ and Paul _interesting_ things to learn, why waste brain cells on Paul something as basically useless and uninteresting (and baroque) Paul as modmap syntax? john Personally, I find the xkeycaps interface to be baroque, john relative to the not-really-that-complicated Xmodmap syntax. OK... I'll just mention, though, that it took me a total of about 15 seconds to modify my Windows key: started xkeycaps, left-click on the Windows key on the displayed keyboard, and picked Edit Keysyms of Key. Of course, I still don't know much about X modmap syntax... but I never had to look up whether Menu was a valid keysym or not. john Not everybody thinks better in GUI; some of us prefer text. Some of us prefer practicality, and use the best tool for the job... even if it does happen to have a graphical interface. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: abbreviations for non-native english speakers
There are many collections of definitions. I asked Google, and here's one it gave me back: http://www.harley.com/abbreviations/ There are undoubtedly many more. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: ~/.Xmodmap (was: Customizing the console key map?)
%% Mike Pfleger [EMAIL PROTECTED] writes: mp What keyboard are you using? I have a PC-104 kbd, and I have been mp trying to bind the windoze and menu keys to be Mod3 and Mod4 mp modifiers, respectively. mp What am I doing wrong? You're trying to use xmodmap directly instead of using JWZ's most excellent xkeycaps program. # apt-get install xkeycaps and try that... :) -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: ~/.Xmodmap (was: Customizing the console key map?)
I'll repeat my previous advice: use xkeycaps. Life is too short to futz with modmap. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: ~/.Xmodmap (was: Customizing the console key map?)
%% Mike Pfleger [EMAIL PROTECTED] writes: mp On Wed, May 30, 2001 at 04:01:13PM -0400, Paul D. Smith wrote: I'll repeat my previous advice: use xkeycaps. Life is too short to futz with modmap. mp Yes, well; using xkeycaps doesn't teach me anything, does it? First, as someone else pointed out, it certainly _can_ teach you something, if you want... Second, my point is, with so many truly _useful_ and _interesting_ things to learn, why waste brain cells on something as basically useless and uninteresting (and baroque) as modmap syntax? Anyway that's my opinion but, of course, YMMV :). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Customizing the console key map?
Is there any approved way to create customizations for the console key maps? There appears to be some kind of include functionality, so it seems to me that the package should be able to support customization files which are separate from the main files, and construct a local customized map which includes both the standard files and local additions. See, for example, how the Debian X packages handle font aliases, and any number of other similar customizations. What I want is simple, small, and common: I want to always, always, always have my CAPS LOCK key rebound to CTRL. I can do this by hand but every time I upgrade my system and a new console-tools, etc. is installed it overwrites my custom keymap (and/or I have to merge in changes by hand) and I have to do it all over. It seems like this area is ripe for some kind of local modification ability... any hints? I'm using Debian woody (console-tools 0.2.3-23). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Customizing the console key map?
%% Wayne Topa [EMAIL PROTECTED] writes: wt Quoting Paul D. Smith([EMAIL PROTECTED]): Is there any approved way to create customizations for the console key maps? wt Way back when John Fist was writing 'The Linux Gazette he had a wt nice column on how to make a cudtom key map file. I believe it wt was somewhere in the first 15 issues. As that was over 5 years wt ago I don't recall the exact issue. Thanks; I found it (http://www.linuxgazette.com/issue01to08/keys-n-consoles.html), and it is a good article, but it talks about how to rebind console keys in general. I already know how to do this; what I want to know is whether there is any way to do it in Debian, so that when I upgrade my Debian packages console-tools (or kbd, I don't care much which one) my customizations are kept automatically without my having to go back in and fix them up. Look, for example, at how the X fonts are done: there is a directory /etc/X11/fonts/dir for each font type, and in there you can add your own files. For example, you could add a file /etc/X11/fonts/100dpi/my-fonts.alias and put your own font aliases in it; the system will automatically include and install those aliases for you when you run update-fonts-alias. This way, you don't have to edit a system config file to add your own aliases, which would entail re-editing those config files (or merging them or whatever) every time you updated to a new Debian X fonts package. There are many examples of this separation of user customizations into separate files in Debian, which makes system customization and upgrades much simpler. Look, even, at /etc/network/interfaces: same principle is at work here. All the package details are taken out of that file and only user customizations are left. This makes upgrading these network packages much simpler for the user. I was hoping there was some similar way to defined my own console key map customizations file as a separate file, and have the console-tools (or kbd) automatically install/append/whatever that file into its default keymap when it's created. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Debian2.2 and XFree86 version 4
%% Derya PALANCI [EMAIL PROTECTED] writes: dp I had the same problem so i did what you told someone else... but dp it removes some packages now i cannot make xf86config it says dp command not found... do i have to do something? dp I have a i810 board, debian potato 2.2 , newly installed xfree86 4.0.3 If the potato 4.0.3 is anything like the woody 4.0.3, you don't want to configure it with xf86config, etc. Instead, configure it through the Debian package setup. As root, you can run: dpkg-reconfigure xserver-xfree86 -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: [users] Unkillable process
%% MaD dUCK [EMAIL PROTECTED] writes: md also sprach Andrei Ivanov (on Tue, 22 May 2001 10:31:26PM -0500): scorpio 7314 0.0 3.8 2 4876 tty1 DMay10 0:00 /usr/local/mozilla/mozilla-bin md this is a straight-forward failure of the linux kernel. it's a dead md process, it doesn't listen to anything anymore. It's possible, under severe error conditions, to get processes which won't respond to kill -9 on any kernel. The KILL signal may not be blocked by the process in user space, but that doesn't mean that it can't be blocked by the kernel in kernel space, and it often is. Simply whacking a process within the kernel at the instant you kill -9 would leave all sorts of resources unreleased, etc. etc. Remember that when you kill a user process the kernel cleans up all its memory, open file descriptors, etc. after it. If you kill a process within the kernel, who cleans up after that? Thus, the kernel doesn't allow processes to just disappear no matter what state they may be in within the kernel itself. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Christian Meder?
Has anyone seen or heard from Christian Meder? The last message I can find from him anywhere, including deja/Google, was 31 Aug 2000. I'm curious because the lclint package, which he maintains, in Debian is _WAY_ old and I really wish someone would install a newer version... Thanks... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Using old libc for some programs?
I've just upgraded my testing distribution which installed libc6-2.3.1. Now, one of my apps (a daemon that uses sockets (I guess)) isn't working anymore. That is, the daemon starts and appears to be running normally with no errors in the log, but none of the clients can attach to it: they all give can't find ... errors. My suspicion is that its the new libc which needs to be recompiled. However, the app is proprietary so I don't have the source to recompile it. What I'd like to do is run it with LD_LIBRARY_PATH (or whatever) set to use a different, older libc. I've downloaded the stable libc6 (2.1.3) and use dpkg-deb -x to extract it... but this doesn't work because of ld-linux.so.2 or something: $ LD_LIBRARY_PATH=/opt/src/old-libc/lib:/lib app app: error while loading shared libraries: /opt/src/old-libc/lib/libc.so.6: symbol _dl_debug_impcalls, version GLIBC_2.0 not defined in file ld-linux.so.2 with link time reference Anyone have any hints or help for me? Thanks -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: SSL : RE: water, water everywhere, but not a drop to drink.
%% 'Dave Sherohman' [EMAIL PROTECTED] writes: ds I removed telnetd-ssl when I discovered this because I couldn't ds find a way to turn the fallback behaviour off. Probably someone already said this (I can't figure out why _some_ of my email is taking 3 days to reach me :-/), but you must not have looked very hard--the man page for both telnet and telnetd tell you how to keep it from switching back to insecure mode if SSL is not available. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: water, water everywhere, but not a drop to drink.
%% Ethan Benson [EMAIL PROTECTED] writes: eb On Wed, Apr 11, 2001 at 03:39:55AM -0400, Paul D. Smith wrote: # groupadd kevin # useradd -g kevin -m kevin # passwd kevin eb um you only have to do it that way if you use roothat. Um, Red Hat did not invent these. They are the standard tools for creating users on most SysV-like flavors of UNIX, such as Solaris, HP-UX, AIX, etc. While Debian may have some better, more unified way, I'll stick with the portable ones, thanks (although I hardly ever use these anyway, I just edit /etc/group and /etc/passwd directly :)--I only use useradd to get the shadow password stuff set up right initially. I _certainly_ never use usermod :). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Should arrogant, self-important people be encouraged to use Linux?
%% [EMAIL PROTECTED] writes: t Ok, so let me get this straight. Calling people dumb lazy idiots t for wanting decent documentation is OK No one is calling anyone a dumb lazy idiot for wanting decent documentation. Hell, _everyone_ wants decent documentation. We'd all _love_ decent documentation. We don't enjoy hearing from frustrated newbies any more than you (the general you, not you you :) enjoy _being_ a frustrated newbie. People are called dumb lazy idiots when their sole and total contribution to Debian, an all-volunteer organization which provides incredible value to many people, is to tell everyone that they shouldn't be doing what they want to do, but should drop that and instead immediately start doing what _you_ think they should. Maybe you can see how this advice is, after all, not very useful. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: water, water everywhere, but not a drop to drink.
%% Known Human Nick Rusnov [EMAIL PROTECTED] writes: khnr As for running telnet, telnet is fine as long as you know your khnr network is private and secure (eg a private subnet lan of which khnr you are the only user) .. otherwise your passwords are exposed khnr in cleartext to anyone who cares to listen. On my local network I have installed and run telnet-ssl and telnetd-ssl. This is normal telnet authentication, but your password, etc. is sent encrypted instead of in the clear. That's enough paranoia for me, since I have a very strict firewall guarding it. Check it out (apt-get install telnet-ssl telnetd-ssl). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: water, water everywhere, but not a drop to drink.
Tera ks Term Pro with the SSH extenstion to log onto my Linux machine from ks a Windows machine on the local network. ks Right now if I type: ks ssh -v -l root rocky Here's the thing. You can't login remotely as root, by default, over ssh: the ssh setup disallows this (as with everything in UNIX, this is configurable if you really want to do it--it's a bad idea so it's disabled initially). You don't want to work as root, at all, ever, anytime, anywhere, anyplace. Even for testing. _Especially_ for testing. Use root only when you must do root operations, then run screaming into the bushes again immediately after you've done that operation. Go now to your system and create yourself a user account all of your own. Debian install will have highly recommended that you do so. If you did not take its advice, then do this (say you want user kevin): Be root: # groupadd kevin # useradd -g kevin -m kevin # passwd kevin and give yourself a password. Now, log out from root and log in as kevin. Here's the second thing: there are _two_ ways to use SSH. One is to authenticate using normal password methods; in this method you have to type a password and SSH just encrypts the transport. I actually have never once used this, but I think you don't need to do anything special (like make keys with ssh-keygen, etc.) to use it. ks I get this error message about authenticity not being established. ks I made a 'indentification' and an authorization file in the ~/.ssh ks directory along with the keys created by ssh-keygen, but I really ks don't know what I'm doing. To use RSA (public/private) key authentication, do this: 1) When logged in as yourself, run ssh-keygen. This will give you a new public/private key pair, and put them in the right spot (~/.ssh). 2) The _server_, or remote, system must have your _PUBLIC_ key in the ~/.ssh/authorized_keys file, one key per line. Anyone trying to log into the system who has a private key that matches any of these public keys will be granted access. The private part of the key you just created is ~/.ssh/identity. The public part of the key is ~/.ssh/identity.pub. Since this is your first key, you can just copy the public one to authorized_keys: $ cp ~/.ssh/identity.pub ~/.ssh/authorized_keys If you have more than one you have to append them, of course. 3) To test it locally, you can now ssh to your own system: $ ssh -l kevin localhost If that doesn't work, report back as to exactly what happens (what errors you see, etc.) Also you can turn on verbose mode (add the -v option) to maybe see more information. If that does work, now you need to give your private key to the Windows SSH client. I'm not familiar with it, so I can't really help, but you need to get the contents of the ~/.ssh/identity file installed in your Windows client somewhere. Be careful! That private key is like your password; anyone who gets a copy can get into your system. It's a good idea to sign the key with a passphrase when ssh-keygen asks for one: then people not only need the private key but they also need your passphrase. This is more secure because the passphrase is used only to unlock the key locally; neither the passphrase _NOR_ the key itself are ever transmitted over the network. Public/private key cryptography is not the most straightforward thing in the world, unfortunately. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: water, water everywhere, but not a drop to drink.
%% Noah L. Meyerhans [EMAIL PROTECTED] writes: nlm On Wed, Apr 11, 2001 at 01:22:42AM -0400, Paul D. Smith wrote: On my local network I have installed and run telnet-ssl and telnetd-ssl. This is normal telnet authentication, but your password, etc. is sent encrypted instead of in the clear. That's enough paranoia for me, since I have a very strict firewall guarding it. nlm Are there any decent SSL telnet clients for non Unix platforms? Ah! Yes. The $64,000 question. Well, maybe these days closer to $6,400 :). I don't know. But it would be nice if there were. However, I think SSH can act like SSL telnet, in that it can authenticate using normal passwords rather than RSA or DSA keys. And of course there are numerous SSH clients. Equally obviously, it's not telnet, so if you have firewall issues or something that might be a consideration (although I sure don't know what admin in his right mind is going let telnet through a firewall while blocking ssh! :) -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: C-A-Fx not working in XFree86 4.0.2?
%% Erik Steffl [EMAIL PROTECTED] writes: Thanks for the reply. es does chvt work? I don't have chvt anywhere on my system (locate says no), nor can I find it with apt-cache search or the Debian web-based package search. Can someone who has it do a dpkg -S chvt and tell me where it comes from? Oh, wait, my Linux box at work has it--it's in console-tools? Why isn't that installed on my home system? OK, I installed it, and using chvt 1 does switch me to the console... but C-A-F1 still doesn't work. es are any other key-combinations ignored? there is this application es (can't remember the name) that hsows which keys were pressed - try it to es see how your keyboard works... xev. It doesn't show anything: it just gives me 3 keypress events, one for CTRL, one for ALT, and one for F1 (the appropriate keyrelease events). Nothing else. I also can't find anything of interest in any of my log files. Damn weird! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
C-A-Fx not working in XFree86 4.0.2?
It seems to me that I saw something about this a few months ago, but 30
mins+ with Google and the Debian mailing list archive search engine
netted me zippo :(.
I upgraded to the latest testing version a few days ago, and ever since
I did I can't use C-A-F{1,2,3...} to switch back to my text console.
The key sequence is simply ignored as if I'd done nothing at all.
Prior to this I was using a testing version from about 4-5 weeks ago,
with X 4.0.1 (IIRC).
I switched window managers (from fvwm2 beta to sawfish), but the problem
persists, so I it's not a WM taking those keys (is that even possible,
or does X intercept them too early?)
I haven't changed my XF86Config-4 file (I have nothing in my ServerFlags
section). I haven't changed anything in any other setup that I can
think of. I _do_ have getty's running on the first 6 consoles; once I
quit X I can A-F2, etc. to switch to them.
I've tried this both with a DM (GDM) started, and without one (using
startx from the console).
Darn weird: anyone know what's going on here?
--
---
Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools
Please remain calm...I may be mad, but I am a professional. --Mad Scientist
---
These are my opinions---Nortel Networks takes no responsibility for them.
Re: XFree86 4.0.2 in Debian
%% Bill White [EMAIL PROTECTED] writes: bw I have heard that people have gotten the sources and built them bw for potato, but it's not clear to my untutored eye which source bw packages I need to get. Look on DebianPlanet (www.debianplanet.org). There's an article there somewhere about building X 4.* source debs for potato. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
webmin?
I could have sworn that I saw a Debian package of webmin before, but now it doesn't show up in my database with apt-cache search, nor does it appear in a search of the Debian packages on www.debian.org. Did I imagine this? If it went away, is there any interest by anyone in reviving it as a package? Thx. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Log rotation in Debian /var/log
%% Ethan Benson [EMAIL PROTECTED] writes: eb On Wed, Mar 21, 2001 at 09:08:00PM -0500, Paul D. Smith wrote: This seems needlessly complex: how many different log rotating tools do we need to have on a system in order to do the job? Debian should pick one and try to move all the standard system services to use that one tool, IMO. eb debian packages have been rotating logs for a lot longer then eb logrotate has existed. Sure; I meant going forward it would be good to converge on one method. eb current debian policy *suggests* using logrotate, so packages should eb and some do, but its not required. Unfortunately after playing with logrotate, it has some significant drawbacks that would need to be addressed before it could be considered sufficiently generic to be used by everyone. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Log rotation in Debian /var/log
%% Ilya Martynov [EMAIL PROTECTED] writes: im Just curious: What drawbacks? I found two, but only one is really critical: 1) You can use globbing to have a single stanza match multiple log files, BUT the entire stanza is run each time for each log file, _including_ the postrotate scripts! IMO, this makes it useless for any daemon that generates multiple log files, like Apache or even INN (the example they use in the man page!) Who wants to have their Apache or INN servers bounced 15 times in quick succession every night just because it generates 15 log files that need to be rotated? The postrotate script should be run _one_ time for each stanza, after all the logs are moved and before they are compressed. 2) If #1 is fixed, then there's a lesser problem that only one filename phrase is allowed per stanza. This is probably OK for a system log utility since it's not a good idea to put your log files in different places, but for a tool like Apache where you might have virtual hosts for different users and want to put the log files for each virtual host in that user's directory, it's going to be pretty hard to come up with only one globbing expression that matches all the log files! Of course, Apache has more serious requirements since you don't even know the names of all the log files; the current Apache setup parses them out of the Apache config file. It would be nice if logrotate handled that; maybe a special include syntax that invoked the file via popen() or something, and read its output. Or, something. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Log rotation in Debian /var/log
So, how does this work? I mean, I know about logrotate and I ass-u-me that the logs in /var/log are rotated using logrotate... no? When I look in /etc/cron.daily/logrotate I see an invocation of logrotate with the config file /etc/logratate.conf. All well and good. But, there is no entry in /etc/logrotate.conf that pertains to the general contents of /var/log! It does /var/log/wtmp and /var/log/btmp and that's it. How are the rest of the log files getting rotated? Is this built into logrotate somehow so it doesn't need to be configured? Or what? Also, in the logrotate.conf file it has a comment that it includes the /etc/logrotate.d directory because RPM packages drop log rotation information into this directory. The only file there on my system is for junkbuster (which is a Debian package, not an RPM). If this is for RPM logrotate configs, is there somewhere else for Debian ones? Hmm, looking at the policy manual I see /etc/logrotate.d explicitly mentioned as the right place to put logrotate configs. Maybe this comment in logrotate.conf needs to be removed/reworded. Thanks... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Log rotation in Debian /var/log
OK, I see sysklogd. This seems needlessly complex: how many different log rotating tools do we need to have on a system in order to do the job? Debian should pick one and try to move all the standard system services to use that one tool, IMO. Next question: how in the heck are the Apache logs rotated? As far as I can tell, sysklogd doesn't do so. Anyway, none of the invocations of syslog-listfiles I came up with listed any of them to be rotated. Also, the apache config doesn't appear to use the method suggested in the Apache web site (pipe to rotatelogs or whatever). Nevertheless, my /var/log/apache/* log files are without a doubt being rotated. Who is doing it, and from where?!?! Enquiring minds want to know... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Linux wannabe
%% Matthew Dalton [EMAIL PROTECTED] writes: I think that Linux boots up faster than Windows, so it's maybe an advantage. md 2.4 boots quite a bit quicker than 2.2 did, but I have a feeling that md Windows is quicker still. Of course, the golden rule of optimisation is md that you optimise the most common task... This must depend on your system setup. There is absolutely no question that Linux (2.2.18) boots _much_ faster than Windows98 on my system (homegrown PII 450, 128M RAM); I timed it once and Linux was over twice as fast as Windows. Note by boot for Windows I mean the time it takes to get to a stable screen with all the silly little applets running in the tray, etc. (I don't have any login required on my Windows partition). For Linux I counted to the time it takes to get the XDM (really GDM) login screen up. Since I use vanilla FVWM 2.x as my WM and no desktop apps beyond FvwmButtons it takes only 2-3 seconds or so to login to Linux anyway. YM, of course, MV. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: New user missing his true-type fonts
I did this a few months ago and wrote it down; try this and see if it helps: http://www.paulandlesley.org/linux/debian_tt.html -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
XF 4.0.2/testing: won't use new server!
Hmm. This is baffling me. So, last week I installed XF 4.0.2, but it was having problems with my video card (Nvidia). So, I switched back to the XF 3.3.6 SVGA server, although the rest of my system is still 4.0.2. To the best of my recollection, I switched back simply by modifying the /etc/X11/Xserver file and changing the first line from: /usr/bin/X11/XFree86 to: /usr/bin/X11/XF86_SVGA However, now that I have some time and want to investigate further, I can't make my system run the new server anymore! I've switched that file back to the first version but it's still invoking the XF86_SVGA (as both the debug output and xdpyinfo attest). In fact, I tried putting a completely bogus file in there as the first line and it doesn't care; it ignores it and uses XF86_SVGA. No errors or messages regarding this are generated anywhere that I can find. This is true running both startx from the console _and_ using xdm (gdm really). How in the heck do I tell the X system I want to use the new server again? -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: True Type fonts
You might also find this link helpful; it's a little more up-to-date than the TT-Debian mini-HOWTO. http://www.paulandlesley.org/linux/debian_tt.html -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: True Type fonts
%% Hall Stevenson [EMAIL PROTECTED] writes: hs That's the page I used in order to get my TT fonts set up, along hs with the http://home.c2i.net page that it references. Since I'm hs using XFree86 v4.x.x, I skipped the step regarding xfs-xtt and it hs still works fine. As I recall, the font alias portion was what hs really helped everything work ideally. Since I wrote that I've upgraded to 4.0.2 in testing, and everything is still working, and still using xfs-xtt. I've heard some people say that server is better at rendering TT than the default 4.x server, but I have no idea if that's true. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Host lookup for different apps...
I'm seeing a very weird thing on my Linux box (Debian testing, but not the latest). I'm wondering if anyone has any insight. So, I have a local hosts file with a few hosts in it. I also have NIS, with a whole _boatload_ of hosts in it (8782, to be precise). And, I have DNS nameservers. My /etc/host.conf file says: order hosts,bind multi off My /etc/nsswitch.conf file says: hosts: files dns and my /etc/resolv.conf file says: search one.domain.com two.domain.com three.domain.com four.domain.com nameserver xx.yy.zz.qq nameserver ss.tt.uu.vv Suppose I have a host which is in my /etc/hosts, and in NIS and DNS. If I use nslookup on this host, it is found on my first nameserver, as myhost.one.domain.com. I cal also use nslookup on the IP address and get back the right value. If I write a tiny program which invokes gethostbyname() on my hostname, and I use strace on it, I see that it obeys the settings in host.conf, gets the value from /etc/hosts, and never tries to contact my DNS servers. Ditto for programs like rsh and rlogin. _However_, if I try to use telnet or ssh it _doesn't_ stop at /etc/hosts, but rather continues and tries to look up the hostname in DNS. Strace shows this definitively. I see an open of /etc/hosts, then a read of the contents, then a close. Then it loads libresolv.so.2, then I see a connect to the first nameserver entry, and a send() whose argument includes the first domain on the resolv.conf search path. Then I see a recvfrom(), that returns an OK value (0). Then I see another connect to the same server, and a send of the host plus the second domain on the search path. This also succeeds. Then it appears to try to do it couple of times more, but doesn't ask about the other two domains in the search path. Now that is weird enough, but it gets more strange: there are _some_ hosts where it doesn't stop at the second domain, but instead also tries the third and fourth. Now, these domains don't know about this host, so they fail. The poll waiting for a result from the send() request takes 4 seconds to return (strace -r shows an elapsed time of 3.921295s). Then it tries the fourth, with similar results. Then it looks like it just tries the hostname by itself. All fail, but not before waiting 4+ seconds. This latter behavior is what started me investigating: in order to telnet or ssh to these particular hosts I have to wait 15-20 seconds, just sitting there, before I get a prompt. With other hosts its essentially instantaneous. And, I can't come up with any differences in the way these hosts are configured; adding them or not to /etc/hosts makes no difference, and they're equally available in my DNS server. I'm stumped! BTW, I should point out that I have some Solaris boxes whose telnet and OpenSSH don't have this delay to any hosts... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: openssh and non-free ssh coexisting
%% Robert Waldner [EMAIL PROTECTED] writes: rw I´ve got network equipment which only supports RSA-ciphers rw sigh, so obviously I cannot use OpenSSH (which supports only rw blowfish and 3des) with it. Not obvious at all. The RSA patent expired last September, and there is no restriction to using RSA in OpenSSH (or anywhere else, for that matter). OpenSSH does support RSA (or rather, it will if your OpenSSL library has support for RSA built in, since OpenSSH uses OpenSSL for this). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Boston area: which ISP would you recommend
FWIW, I've had RCN in Lexington (MA) for 8 months now and it's great; no bandwidth or latency issues at all, and I've not noticed any service outages. But, they had to basically lay all new infrastructure here so maybe it's different in Boston proper. RCN is also DHCP only (well, maybe you can get static for more $$, I'm not sure), but so far I've only had my address change once in 8 months, and I think it was some kind of infrastructure change since the new address was _completely_ different than the old one (new class A even, IIRC). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: CPU Usage
%% Sven Gaerner [EMAIL PROTECTED] writes: sg I'm running Debian woody with a 2.4.0 kernel. I configure my sg kernel not to use ACPI but APM. When my system runs one process sg (kapm-idled) is using about 60% of my cpu time. I'm a little bit sg confused if this is right. It is right. sg Can anybody tell me how to decrease this cpu usage or if this is sg 'normal' for this process. It's normal. That represents the idle CPU (that's why it's called idled). If any process on your system wants that CPU, it'll get it. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: internet connection sharing
%% Omar Shuja Siddiqui [EMAIL PROTECTED] writes: oss i have recently installed a network of to Linux boxes. oss one of them is a working dial up machine. i want to oss share the internet connection with the other Linux box oss also. please tell me what is the whole procedure for oss doing this. The most common way to do this in Linux is called IP Masquerade, or IPMasq. See the HOWTO for how to do it: http://www.linuxdoc.org/HOWTO/IP-Masquerade-HOWTO.html I had this working with a dialup 56k modem over ppp, where my wife's laptop was sharing the link (we have wireless enet cards for the home lan). Worked fine (well, as fine as two people sharing a 56k modem could be expected to work :). Now we have a cable modem connection, and that also works great--with much better bandwidth :). If you're going to be connected more continuously than dialup you should seriously consider hardening that system to avoid someone hacking through it into your home network. Check the IP Chains HOWTO, at the above site. There are other Security-related docs floating about as well. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Want TrueType fonts for Debian?
I just spent a few hours wrestling with getting my Windows TrueType fonts working in my Debian 2.2 system (XFree86 3.3.6). I used the xfs-xtt server to do this, not the xfstt server. Not sure which is better, but the results I got were quite nice. Since it involved a good bit more than the typical trivial apt-get install that most Debian packages require, I wrote up some instructions that might help the next person get there more quickly. If you don't want to move to XFree86 4.x yet but would like to have TrueType fonts, you might find these notes handy. Check it out: http://www.paulandlesley.org/linux/debian_tt.html -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Setting console keybindings
%% MH [EMAIL PROTECTED] writes: m What's about using loadkeys and other keymap-utilities... This sounds promising, but I can't find a package that contains these things, after apt-cache search'ing on loadkeys, keymap, dumpkeys, etc. etc. Any hints? Thanks! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Setting console keybindings
What's the good/right/proper way to set console keysyms? I want to change the CapsLock key to a CTRL key in my Linux console. I see that there are two packages which manage this, either kbd or console-tools. The latter is installed by default; is this the preferred tool going forward? With either of these I don't really see any way to customize the keymap that it installs; I could edit the default keymap before or after installation, but I would much prefer to come up with some out-of-band way of adding an override for this. I'm worried that if I modify these files my changes will just get overwritten when I upgrade these packages. Thanks for any hints! PS. Do people realize the FAQ-O-Matic is down? -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: Setting console keybindings
%% [EMAIL PROTECTED] writes: m Once you have built the keymap file that you want, store it in m /etc/console-tools/default.kmap.gz. It will then get loaded at boot m time and will not be disturbed by package updates. OK. But, it still seems to me that if I rerun kbdconfig or whatever it will copy a new keymap and overwrite my current default.kmap.gz. Right? Sure, probably it'll ask me first, then I can go back and do that same modification to the new keysym file. It's sure a shame that there's no way to include extra, user-specifiable settings that don't require modification of standard files; say you create a /etc/console-tools/user.kmap file and the toolset automatically appends the contents of that file to the default keymap when it builds it, or something. Thanks. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
ttmkfdir vs mkttfdir
So, I'm trying to follow the X Font Deuglification howto. I have a bunch of Windows fonts on my other partition (the old dusty one that came with the system but is almost never used anymore). I want to grab those fonts, just like the above doc says. I copy them all over. I've installed xfs-xtt (I'm still using XFree86 3.3.6 and am not interested in moving to 4.x yet, thanks anyway...) Now it says use ttmkfdir; I've searched everywhere and I can't find this in any Debian package. The mailing lists suggest using mkttfdir instead (why? Why isn't the original included with fttools or something?) I tried that (a perl script apparently), and I just get Segmentation fault: # mkttfdir Segmentation fault. I checked and I _do_ have a fonts.dir file after this bombs, but I don't have a fonts.scale, which is what the above doc says I need, then I'm supposed to run mkfontdir to get fonts.dir. How can I get a fonts.scale file? Why is mkttfdir dumping core? And why do we have a different implementation of this tool in Debian than everyone else uses; this seems counterproductive? I'm downloading the tarball and building ttmkdir myself now... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: glibc devel info pages
%% David Purton [EMAIL PROTECTED] writes: dp On Tue, 2 Jan 2001, Mike wrote: hal9000:~$ man vasprintf dp useful dp what package provides these man pages? My system sez: $ dpkg -S vasprintf manpages-dev: /usr/share/man/man3/vasprintf.3.gz So, use: # apt-get install manpages-dev -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Package pool information?
I'm really interested in the package pools implementation that's going on now. Which list(s) are the best to subscribe to/and/or read archives of if I want to keep up-to-date with this new feature (issues, decisions, announcements, etc.)? I found a debian-pool list, but it seems dead (just a few posts back in August). Thx! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: tt fonts listing but unavailable?
%% Alson van der Meulen [EMAIL PROTECTED] writes: I used to use this font: -*-barbedor t heavy-medium-r-normal-*-100-*-iso8859-1 but now when I ask for it, FVWM says it's not found. When I use xlsfonts and grep for barbedor I can find this font: -2rebels-barbedor t heavy-medium-r-normal--0-0-0-0-p-0-iso8859-1 When I use xfontsel -scaled, I can choose the exact font I want, but my WM doesn't grok it. This worked fine with the old X server and xfstt; is there something I have to do to allow clients to use the scaleable fonts in X 4.x? avdm i had the same problem with X3.3.6 + xfstt when freetype1 was avdm not installed yet, xfontsel/gimp showed it in the list but avdm displaying the font failed. I definitely have freetype2 installed; is that good enough? Thanks for the reply... -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
tt fonts listing but unavailable?
I just installed X 4.0.1. After some tweaking it seems to be mostly working. I followed the directions listed here previously to get my TT fonts (copied from my Windows partition, plus other places) visible. Now I can see them all with xlfonts. However, my previous window manager (fvwm 2.3.x) title bar fonts are not working any more; they worked before the upgrade (I was using XFree86 3.3.6 plus the xfstt server--I've now removed the xfstt server package and I'm not using any font server at all, just what comes with X). I used to use this font: -*-barbedor t heavy-medium-r-normal-*-100-*-iso8859-1 but now when I ask for it, FVWM says it's not found. When I use xlsfonts and grep for barbedor I can find this font: -2rebels-barbedor t heavy-medium-r-normal--0-0-0-0-p-0-iso8859-1 When I use xfontsel -scaled, I can choose the exact font I want, but my WM doesn't grok it. This worked fine with the old X server and xfstt; is there something I have to do to allow clients to use the scaleable fonts in X 4.x? Thanks! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: installing packages: can't locate file/glob.pm
%% Kelly Corbin [EMAIL PROTECTED] writes: kc I had the same problem. For some reason, perl-5.6-5.6 was not kc installing before the apps that needed it. I downloaded the .deb kc manually, installed it, and the errors when away. My understanding (perhaps incorrect) is that debs are supposed to only need perl-5.6-base, not the complete perl-5.6. Unfortunately, perl-5.6-base doesn't include the File::Glob.pm. In previous releases of Perl this wasn't needed on most platforms: perl invoked csh (ugh!) to handle all the globbing the glob function needed. In Perl 5.6.0 they changed that to use internal globbing functions, in File::Glob. I submitted a bug against perl-5.6-base for this yesterday, saying that it should include File::Glob.pm. I suppose an alternative would be to disallow the use of glob in deb config scripts. Not sure what the final outcome of that will be. At any rate, the answer here is the right one: find a way to install the real perl-5.6, not just perl-5.6-base, and you'll be all set (for this problem, at any rate). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: emacs without backup option
%% [EMAIL PROTECTED] (robert_wilhelm_land) writes: rwl Jay Ford wrote: Does someone know which option is regarded to avoid emacs from imediately creating a backup file by opening a new or any other file? Just a clarification: Emacs doesn't make a backup when you open a file, new or otherwise. It only (potentially) makes a backup when you _save_ the modified file. Put the following in ~/.emacs: (setq make-backup-files nil) ; Never create backup files It works for me. rwl Just tested it - exactly what I wanted. Thank you very much! rwl Where did you get the synthax from and why does this setting rwl require brackets? Parentheses. This is Lisp. The vast majority of Emacs programmed in Lisp; Emacs is really a Lisp interpreter with a bunch of primitive functions appropriate to an editor, written in C with Lisp bindings. This is why Emacs is called extensible; with a complete programming language you can add huge amounts of functionality on top of the basic editor. Consider, for example, the Gnus newsreader, VM mailer, etc. These are programmed entirely in Lisp; _no_ actual C code in Emacs was changed or added for these packages to work. Pretty impressive, IMO. And, Lisp is cool. However, if you're not into Lisp and just want to customize Emacs a little, you probably should be using Emacs' customize feature to make these changes to how Emacs behaves. It's much simpler. To use it, select the Customize submenu off of Help. If you just want to look through all the thousands of things you can customize, start at the top of an options menu with Top-Level Customization Group. Browse Customization Groups will give you a tree-like list of all the groups you can customize. If you know, as in this case, something about what you want to customize, you can use Apropos... or, since this is an option, Apropos Options Enter some keyword you want information apropos of, such as backup, and you'll get a list of all the options associated with that. In this case, I get 8 possibilities. The second from the last is Make Backup Files, set to on. In Customize, you use mouse-2 to select things, so click mouse-2 on the [Toggle] button to change it to off. Then, at the top of the buffer, click mouse-2 on the [Save for Future Sessions] button. Done. Note that you can do _significant_ customizations; really, complete overhauls, of Emacs functionality by writing Lisp, and it's not that hard. But, that's for later once you're a true Emacs junkie :). Have fun! -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: VI probs
%% Stefan Janecek [EMAIL PROTECTED] writes: sj Whats called 'vi' in suse is actually vim. do 'apt-get install vim sj vim-rt' and use vim to edit your files - it should behave like vi sj on suse... Just a note: you don't need to use the command vim to edit your files. When you install VIM, Debian will use the alternatives functionality to set the vi alternative to point to vim. So, you can just continue to use the vi command. If that doesn't happen, use update-alternatives --config vi to pick the alternative you like--I can't remember when the alternatives are set for you and when you have to do it by hand. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: wm
%% Erik Steffl [EMAIL PROTECTED] writes: es it also has better pager (IMO) and overall virtual screens behaviour es then (most) others. you can drag windows from pager to current screen, es from screen to screen, switch between screen and move view many ways es (mouse, keyboard, explicit action on pager)... Yes; the coolest thing about FVWM IMO (aside from the modules concept, which I like) is that it always pays close attention to the keyboard: _any_ FVWM operation can be bound to a key, if you like. I know many people who use a rich graphical environment, including multiple screens and desktops, and who never use the mouse for _any_ WM operations. es however, as of now, the configuration is done by editing files, some es other WMs have GUI based config (which I don't like too much, it's not es as flexible, but some people like it more). Note the dotfiles collection has a graphical configurator module for FVWM. I haven't tried it in a long time so I don't know how many of the fancier features are available. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: system.map
%% Gnanasekaran Thoppae [EMAIL PROTECTED] writes: gt How do i generate new System.map file? Do I need this file gt whenever I generate new kernels and modules? The simplest thing to do is to use the kernel-package tool to build yourself a new kernel package, then install it using dpkg -i. Not only will this put all the right files in all the right places (including System.map), ask you about Lilo, etc. etc., but it makes visible the version of the kernel you're using in your Debian database. It's really very easy to use, too. Highly recommended. # apt-get install kernel-package # cd /usr/share/doc/kernel-package # zcat README.gz | less (or view it directly if you have LESSOPEN set up properly). To actually build a kernel, you just: # cd /usr/src/linux # make config (or xconfig or menuconfig or oldconfig or whatever) # make-kpkg clean # make-kpkg --revision=3:custom.1.0 kernel_image (See the README for info on the epoch value (here, I used 3:)) Now you can install the result like this: # cd /usr/src # dpkg -i kernel-image-2.2.17_custom.1.0_i386.deb (if you have module sources to be installed, like ALSA or something, use make-kpkg again with the modules_image target instead of kernel_image, then use dpkg -i to install those as well). -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
Re: how to get the latest STABLE releases?
%% Daniel Borgmann [EMAIL PROTECTED] writes: db why aren't stable packages for the unstable tree moved to the db potato tree? Because that would be very difficult. First, packages all depend on other packages. The package in woody may well depend on other packages with newer versions in woody, so you can't always just stuff one package by itself from woody into potato. Second, people aren't testing the package on potato, they're testing it on woody (usually), so just because it appears stable on woody doesn't mean it will work as well on potato. db am i dammned to use years old unstable software packages until db woody is released next year? There is help coming for you. In the next few weeks, hopefully, the entire unstable package archive will be rearranged completely (although, it won't be obvious to us peons). The goal is to create a concept of package pools, so instead of just three package areas (stable, frozen, unstable) we'll have as many as we like. Packages can/will appear in more than one. The idea is you can declare, for example, that you want to upgrade this pool of packages but not that one, etc. Also, there's the test package pool, where packages will automatically be moved into that pool if that package (and all the packages it depends on) has been available for 2 weeks or so with no critical bugs logged against it; this should provide a pretty solid distribution. The whole package dependency thing will allow this to all work without simply generating a huge mess. Once package pools are available, I predict many wonderful things will start happening. Stay tuned. db can i use single woody packets (e.g. licq) on a potato base db without problems? and what should i do to do so? i mean, if i db only want to update licq but remain in the potato tree for every db other package. You can update individual packages _if_ their prerequisites are all met. Add the unstable tree to your apt.source list, then use apt-get install to install just package . _Don't_ use apt-get upgrade or apt-get dist-upgrade or you'll get all of woody. If APT needs to install other woody packages in order to meet the requirements for the package you want to install, it'll list them then you have to decide whether you want to go ahead or not. Your other alternative is to download the source to the woody package, then build it into a package using your potato libraries, etc. using the Debian package tools, then you can install that. -- --- Paul D. Smith [EMAIL PROTECTED]HASMAT--HA Software Methods Tools Please remain calm...I may be mad, but I am a professional. --Mad Scientist --- These are my opinions---Nortel Networks takes no responsibility for them.
