Re: DRM failure with ATI radeon 9200 and Kernel 2.6
Daniel Johnson [EMAIL PROTECTED] wrote: I have a multihead setup, and you can't avoid the bus ID's, and I have yet to see a multihead config program. You might not be able to avoid reading the docs. They are big, and complicated. Sometimes the setup debian gives you isn't enough. I love reading docs*, but *which* docs do you mean? I've already looked at /usr/share/doc/xserver-xfree86/README* and http://www.xfree86.org and various man pages (maybe not the right ones?). Yes, I am woefully ignorant about X. Most of the time it just works, so I ignore it. I'm installing my Mom's first Linux box, though, and it just won't be right until Tux Racer works. ;-) As a practical matter, what she'll care most about is streaming video, which is probably 2D acceleration, though. That's part of the reason I was looking at the Gatos site --- but it does have a page on debugging DRI problems, so I thought it might be relevant. Thanks, Terry *As opposed to not having any docs anyway. ;-) -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: DRM failure with ATI radeon 9200 and Kernel 2.6
On Monday 27 June 2005 02:22 am, Terry Hancock wrote: Daniel Johnson [EMAIL PROTECTED] wrote: I have a multihead setup, and you can't avoid the bus ID's, and I have yet to see a multihead config program. You might not be able to avoid reading the docs. They are big, and complicated. Sometimes the setup debian gives you isn't enough. I love reading docs*, but *which* docs do you mean? Maybe you meant man XF86Config-4? That does seem to cover the options for the Device block. Anyway, that's what I'm reading now. -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
DRM failure with ATI radeon 9200 and Kernel 2.6
Hello! Hardware acceleration (DRI) is not working with my ATI Radeon 9200 card on a fresh install of Debian Sarge (3.1), using the AMD 64 build of Kernel 2.6. From /var/log/XFree86.0.log I'm seeing this message: (II) RADEON(0): [drm] drmSetBusid failed (5, PCI:1:0:0), Invalid argument (EE) RADEON(0): [dri] DRIScreenInit failed. Disabling DRI. Googling, I found some references to a similar problem with the same driver when gcc 3 was used to compile the module (or when it was mixed with a kernel built with 2.95?). However, that was a Permission denied error instead of Invalid argument. That was with Debian unstable in 2003, which probably was Sarge, IIRC, although given that there were such problems, it seems unlikely that it would've gone unchanged. This line has been reported with Invalid argument for some other drivers, but so far, I haven't found a similar report for the radeon. Unfortunately, the only sites I found are in German and Russian, which I don't speak, so I can't really follow the discussion, so I'm just guessing based on the log messages they posted. The XF86Config-4 file has a very simple entry for DRI: Section DRI Mode0666 EndSection so I don't *think* it's a problem with the configuration. That leaves me with the possibility of a (probably different) compiler issue. How can I tell which compiler built the kernel and radeon modules that I'm using? What exactly is the Invalid argument message I'm seeing? I'm guessing this might be a C API mismatch? Or maybe there is something being passed from a config file that I can change? Any recommendations for what to try next? Any ideas appreciated, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com
Re: DRM failure with ATI radeon 9200 and Kernel 2.6
Daniel Johnson [EMAIL PROTECTED] wrote: On 6/26/05, Terry Hancock [EMAIL PROTECTED] wrote: Hardware acceleration (DRI) is not working with my ATI Radeon 9200 card on a fresh install of Debian Sarge (3.1), using the AMD 64 build of Kernel 2.6. I have a radeon 8500, and it works with a fresh install of sarge with a 2.6.8 kernel. That is a similar card I think. I don't remember if I had to manually uncomment the dri, and glx lines in the config file. From /var/log/XFree86.0.log I'm seeing this message: (II) RADEON(0): [drm] drmSetBusid failed (5, PCI:1:0:0), Invalid argument (EE) RADEON(0): [dri] DRIScreenInit failed. Disabling DRI. Could you have the busid wrong? Well, I suppose, but this is all automatic. To my knowledge, I haven't specified it anywhere. One thing bothers me a little --- this is an AGP card, why is the bus id a PCI specification? Is that part normal? There is a series of lines in the XFree86.0.log file which appears to show the autodetection process: drmOpenDevice: minor is 0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is -1, (Unknown error 999) drmOpenDevice: open result is -1, (Unknown error 999) drmOpenDevice: Open failed drmOpenDevice: minor is 0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is -1, (Unknown error 999) drmOpenDevice: open result is -1, (Unknown error 999) drmOpenDevice: Open failed drmOpenDevice: minor is 0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is 5, (OK) drmOpenDevice: minor is 1 drmOpenDevice: node name is /dev/dri/card1 drmOpenDevice: open result is -1, (No such device) drmOpenDevice: open result is -1, (No such device) drmOpenDevice: Open failed (I've included some failures for context -- there's a dozen or so failed combinations, and only the one device at /dev/dri/card0 found. One odd thing -- what is the difference between the first 3 attempts? They all look like they are using the same parameters to reach /dev/dri/card0, but only the last one fails. Is that normal? Why Unknown error instead of No such device?) Did you check it against lspci? I have done so now. ;-) Here's the relevant lines: :01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01) :01:00.1 Display controller: ATI Technologies Inc RV280 [Radeon 9200] (Secondary) (rev 01) I'm a little uncertain about how to read the bus ids here, but they look plausible. Maybe it should read PCI:1:0:1 or PCI:0:1:0? Even if I want to change this, though, where do I put the information? Is there a line in XF86Config-4 for this? I suppose I can try to set these with dpkg-reconfigure xserver-xfree86 -- I'm pretty sure there is an option to specify the bus. But what should I tell it, if this isn't right? Or maybe just the syntax of that line wrong? Um, which line? Again, I don't know where this would be specified. A syntax error sounds promising, but I don't know where to look for it. ISTM, whatever syntax is being used is being created by the driver software? I'm willing to go to a new driver or whatever if that's required, but this looks like a very minor problem --- it seems like there *ought* to be something I can specify somewhere to make this behave. I'm also looking at the Gatos project and the Debugging DRI page: http://gatos.sourceforge.net/dri-debug.php but it appears to me that this would involve compiling and installing new drivers, which means getting kernel-headers and so on. I'd like to avoid that if there's a simple configuration solution (which I'd likely have to solve anyway, even for new drivers). -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PDF spec (Was: Re: ooh! debian jewelry)
On Sunday 14 December 2003 07:31 pm, Monique Y. Herman wrote: On Sun, 14 Dec 2003 at 15:54 GMT, J.H.M. Dassen (Ray) penned: On Sat, Dec 13, 2003 at 14:08:21 -0700, Monique Y. Herman wrote: Hrm, I could have sworn that PDF was a spec published by Adobe and freely usable, but google seems to disagree. http://partners.adobe.com/asn/tech/pdf/specifications.jsp has e.g. the PDF Reference, Fourth Edition, Version 1.5 (1172 pages). xpdf seems to handle the Acrobat 5 version of it just fine. Thank you! I don't think I remembered to say Thank you, myself, Ray. So, thank you for the link! It appears that I stand corrected on the PDF spec issue. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: RFC: Linux compatibility test framework (was: Re: OT: Letter to TigerDirect)
On Sunday 14 December 2003 02:16 am, Karsten M. Self wrote: on Sat, Dec 13, 2003 at 11:15:07AM -0600, Kent West ([EMAIL PROTECTED]) wrote: John Hasler wrote: Terry writes: Now, if you are just doing this in good faith or making a best effort, you're asking for a lawsuit the first time somebody buys your stuff and can't get the backalley joe Linux distro version 0.0.2 to work with it. Either that, or you're just going to be giving people their money-back an awful lot. So you guarantee that it works with a specific version of a specific distribution. And ship it with that version installed and running. Or to make it even simpler: Designed for Knoppix 3.3 11-02-2003 and include that Knoppix CD without an OS installed. That'd be good enough guarantee for me. I've recommended this (strongly) in the past. What does this mean, exactly? There's more than one idea here. It would be nice to see HW vendors pick up on this. It would also be helpful if a HW regression test fram for various peripherals could be created, automated to the extent possible. There's some work toward this (lshw, hdparm, my own system-info script, etc.), anyone care to mention subsystems and possible tests? You know, I'm not really sure I know what regression test framework means. I've heard it before, of course, but never really heard a definition. But if you mean that it would be good to make testing much easier for the hardware guys by providing automated testing procedures, I think that's a great idea. Regarding supporting Knoppix 3.3 11-02-2003 or some such exact distribution, I think it's missing the point. Not enough people in the Linux users community is going to settle for one distribution. And in this case, the user is optimistically asking for a distribution ONLY ONE MONTH OLD! Think about what you're demanding! It's hard enough to stay in the business of selling Debian CDs, when every few months your customer base decides your inventory is all stale and isn't interested anymore. And CDs are *cheap* compared to computer systems. How many good systems must you sell to justify the cost of all the hardware you had to test and send back because it didn't work? How many Linux users will *actually* buy these hypothetical computers, instead of just agitating for them? The actual turn-out tends to be rather underwhelming, from what little I've been able to see of it. In order to make a useful size run, the vendor's going to want more than 1 month lead-time! You need time to decide on a platform design, test the components, replace any that need replacing, and then fill bulk orders for enough units to make the sale profitable. I'm not sure Linux users as a class are that patient. I'm reminded of the college human geography teacher who described a survey on the usage of public buses in Austin, TX. He said the surveys overwhelmingly showed that people [said they] would use the buses. But this survery was devastatingly wrong -- hardly anyone used them when they were provided, and they lost a lot of money. His explanation was that what people really wanted was for OTHER PEOPLE to ride the buses, so they could drive their cars with less traffic on the roads. Are you sure this isn't what's going on here? Otherwise, why isn't e-Linux (for example) a household name? Do you shop there? I mean -- I don't shop there: I can too easily find what I want at PC Club or the local swap meet, and figure out the compatibility issues on my own. And I think that's probably closer to the average Linux users' attitude. I think that a hardware vendor system that really meshes with the FL/OSS community is going to have to take this kind of thing into account. It will have to be sufficiently flexible and robust that it can actually cope both with lots of different distributions, and with the small core market that it will have to depend on. And it will have to (somehow) engage in the same responsibility and AT YOUR OWN RISK risk sharing that the open-source software movement uses. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel-HOWTO has been removed for review
Roberto Sanchez [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]: Arnt Karlsen wrote: .. ;-) And I need to get time to learn the Debian way. ;-) The Debian way is really a piece of cake. ..patching too? How about Netfilter Patch-O-Matic? From `man make-kpkg': [...] Thanks for the tip on the patching options, I didn't know it included that. Cool! (Because actually, I do have a few patches I want to try out hmmm). The most difficult aspect is getting past the initial psychological fear. My first 6 months using Debian I was petrified of trying to recompile a new kernel. After I finally took the plunge, I wondered why I had been so scared. I started compiling the kernel with make-kpkg when I needed to get ALSA working (there's basically no other way to get it right, as the modules have to be compiled in any case). I found that compiling the kernel with make-kpkg was actually easier than finding the right pre-compiled kernel and module packages. In fact, it was almost as easy as using dpkg (not quite). I think it's so easy, in fact, that I'm questioning whether -- with just a little bit of improvement -- if it could become the PRIMARY way to install a Debian kernel. I wonder if it'd be possible to make the make-menuconfig script look like (or be preceded by) a debconf configuration process. It's already pretty similar. My feeling is that with the Debian kernel-package available, the main obstacle to compiling the kernel is psychological. It just feels like you're doing something drastic, because there's all this lore out there that dates from the old way, and makes it seem like a major hacking experience. But the reality just isn't like that. It's no worse than a basic Gnu Autoconf source install, now. And given Debian's positioning as a convenience distribution for experts, rather than a real newbie distribution, it seems like a very compatible idea. Of course, I am not a Debian developer. Just a very satisfied user. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Copy all desktop settings for a new user
On Sunday 14 December 2003 08:49 am, Philipp Schulte wrote: Nunya wrote: I do it for these files: desk:/mnt/apt/inst/dotfiles# ls dot.fetchmailrc dot.gnome dot.gtkrc-1.2-gnome2 dot.procmailrcdot.xsession dot.fluxbox dot.gnome2 dot.gtkrc-2.0 dot.profile dot.gbuffyrc dot.gnome2_private dot.kde dot.qt dot.gconfdot.grun_historydot.kderc dot.xmms dot.gconfd dot.gtkrc dot.muttrc dot.xscreensaver I don't get it. You are talking about copying the files. This of course is not a problem, but what do you do if you have to change the content of hundreds of files for each user? Something like this (this is tcsh syntax, so sue me): set newuser = set olduser = foreach file ( dotfiles/* ) cat $file | sed s/$olduser/$newuser/g /home/$newuser/$file end Untested, of course, but that ought to give you the idea. man sed man tcsh Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PDF spec (Was: Re: ooh! debian jewelry)
On Sunday 14 December 2003 09:54 am, J.H.M. Dassen (Ray) wrote: http://partners.adobe.com/asn/tech/pdf/specifications.jsp has e.g. the PDF Reference, Fourth Edition, Version 1.5 (1172 pages). xpdf seems to handle the Acrobat 5 version of it just fine. Hmm. Yes, that's very interesting. I saved a copy, although Adobe apparently thinks that was illegal of me: No part of this publication (whether in hardcopy or electronic form) may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of Adobe Systems Incorporated. Please note that the content in this guide is protected under copyright law even if it is not distributed with software that includes an end user license agreement. That's cute wording, isn't it? Doesn't that pretty much debar saving the file on my hard drive? Surely that's a retrieval system? Which says they're trying to deny certain fair use rights, as I read it. In any case, this may be one reason it's so hard to find, since no one is legally allowed to mirror it. I accessed this without registering because you provided a deep-link, but normally, Adobe makes you go through a forms process to get this far, AFAICT. That's probably why Google can't find it, too. So, yeah, it's there. And that's better than I thought. But it still doesn't quite give me warm-fuzzies. ;-) Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Letter to TigerDirect
On Saturday 13 December 2003 03:57 am, Wayne Sitton wrote: I mean holly fril, If Gateway had a sticker that guaranteed that every piece of hardware on a specific Box was Linux compatible, no matter what os was shipped, they'd probably become better than they are currently. But, why do they not get this? HP recently announced that they will start supporting Debian, but where on their site can I gat a laptop pre-installed with Debian? I just want these companies to recognize that we buy too, and that we want to know if a particular product is Linux compatible. This is a good idea, which I've had before. In fact, I liked it so much, I thought maybe I should open a small shop that did precisely this. Then you find out the problem: When you slap a sticker on hardware like that, you aren't just suggesting that it *probably* works with Linux, you are *making a legal guarantee* that it works with Linux and *promising to support it* if it doesn't work on Linux. That can be an incredible burden! Windows (and Sun and Macs) come in relatively few flavors. Furthermore, one flavor is always recommended and all others deprecated, because these companies want to force you to keep upgrading to their latest and greatest OS. So, as a hardware vendor, you're okay if you just support that variety. What's more, you've got a lot of backup -- you have suppliers who *also* guarantee the hardware with that OS, and you've got an OS tech-support number that you can dump people onto if you can't figure it out, and a legal agreement that they're going to do something for you (that's what those designed for Microsoft Foo stickers are all about). Now, if you are just doing this in good faith or making a best effort, you're asking for a lawsuit the first time somebody buys your stuff and can't get the backalley joe Linux distro version 0.0.2 to work with it. Either that, or you're just going to be giving people their money-back an awful lot. Now there *are* a *few* hardware vendors who do officially support Linux (e.g. Netgear NICs -- which is why I buy from them. They not only provide Linux drivers, they provide *source* for their Linux drivers in the consumer-packaged box. This is good practice!). But if you had to build a retail computer business around only those suppliers, you'd never get anywhere. It already sucks badly enough being a hardware vendor -- hardly anything is more perishable than the value of computer hardware. You can *not* afford to hang on to inventory. So either you sell in volume, with extremely efficient JIT supply, or you don't stay in business. No time to hang around waiting for a community driver to be written or fixed. A couple of companies have actually tried to support Linux online. You may be interested in http://www.elinux.com/ for example. But, even they don't really *guarantee* that their hardware works with Linux (they are also quite biased towards the server rather than the desktop and have limited selection -- which just means they've followed their market). And although, I'm pleased to see that e-Linux is still in business, I'm not sure that all Linux users are exactly flocking there. So, as an entrepreneur, it's hard to feel too motivated by this type of business plan. It's *not* obvious that the market is sitting there waiting for you. This kind of dilemma is a direct consequence of Linux's development model: Distributed open-source development and an indemnity from responsibility (as in the Gnu disclaimer we put at the top of our source files), makes it awkward for a company to underwrite our successes in this way (whereas corporate-written proprietary software comes with that guarantee, so the vendor can pass the buck if things don't work as advertised). My point? If we're going to hang onto our software model (and we should), then we're probably going to need a new distribution model for the hardware, too. I'm still not sure what that model is, though I have a few ideas about pieces of it. But it's a problem to be solved. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ooh! debian jewelry
On Saturday 13 December 2003 12:46 pm, Monique Y. Herman wrote: On Fri, 12 Dec 2003 at 03:00 GMT, Greg Madden penned: http://debian.org/logos/ Ooh. I was wondering where to find logos. Why eps as opposed to pdf or whatever? Well, for one reason, postscript is an older and more widespread standard in the printing field (whereas PDF is ostensibly only for electronic distribution). Any printshop running whatever OS on their computers will have an easier time reading and handling encapsulated postscript. That would be the printshop's reason. Furthermore, PDF isn't really an open data format, just a closed one that turned out to be easier to crack than .doc files. Adobe isn't any nicer about sharing their standards than Microsoft is. The fact that we have good Linux readers for PDF has more to do with slow releases of new versions of the standard, and more widespread use (especially in the academic community), than with any intrinsic quality. But the latest versions (5.x?) still don't seem to be accessible with open source tools. OTOH, postscript has been around for so long that it is definitely a well-understood standard (there are published manuals explaining the language in detail). I'm not sure whether it was intended to be open or not, but it is in effect, at least. And that would probably be Debian's reason. ;-) Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PDF spec (Was: Re: ooh! debian jewelry)
On Saturday 13 December 2003 03:08 pm, Monique Y. Herman wrote: Oops. Hrm, I could have sworn that PDF was a spec published by Adobe and freely usable, but google seems to disagree. It references some old links from the adobe site, but they seem to have been removed. PDF 5.x is supposed to contain the same document encryption technology that everyone gets so steamed about MS doing with the DOC format. There are good reasons for distrusting that. It would be conceivable to call PDF 4 an open standard, since Ghostscript can already handle it. But we really ought to make a distinction, since the newer versions are incompatible. This is equally true of DOC format, too, though. We *could* adopt some prior version of it as a standard, seeing as several open word processors can handle them already. I'd rather see something like the OASIS standard http://lwn.net/Articles/16043/ work, though. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PDF spec (Was: Re: ooh! debian jewelry)
On Saturday 13 December 2003 10:11 pm, Nunya wrote: On Sat, Dec 13, 2003 at 10:08:07PM -0600, Terry Hancock wrote: It would be conceivable to call PDF 4 an open standard, since Ghostscript can already handle it. But we really ought to make a distinction, since the newer versions are incompatible. Many PDFs I get don't display correctly in gv. Yes. I have the same problem. I'm not sure whether this means they 1) meet the later standard, 2) are non-compliant with the standard as published (but are compatible with the de-facto it looks okay in acroread standard), or if the problem is actually 3) bugs in GV making it not meet the standard it was written to. I'm inclined to think it's mostly the #1. PDF has evolved, and acroread 5.x supports a lot of bells and whistles that I know GV is not prepared for. I attended media production classes for staff at Caltech in which making maximum use of these PDF 5 features was *really* pushed hard (sometime last year). No doubt they had also received some serious direct marketing, and were just passing it on. (On the other hand, they also have a program pushing a TeX + XML pipeline for publishing theses and dissertations, which I think is a move in the right direction). Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Firebird losing toolbar configuration after restart
I installed Firebird 0.7. It worked fine for a good time, but now it's very very slow closing down tabs that have had pages in them, it's become progressively worse. Just FYI, I've had similar problems with Gecko-based browsers (mostly Galeon, because that's what I use), and here's what it usually turns out to be: If something interferes with the program writing its preferences files, they will not necessarily be saved in a safe format. E.g., if the disk drive fills up, the prefs.js file may well end in the middle of a statement. Even after you resolve the original problem, the file will remain messed up like this. I think the occasional segfault can do this too. Given that browsers are exposed to a rather hostile environment and may be your heaviest used application, this sort of thing tends to happen a lot, and the damage can accumulate. This causes syntax errors on read-back, and the program can't cope with it. The result is a series of weird failures -- password fill-in broke once, and I think you can also run into slow downs as the code gets lost in trying to recover from errors, etc. You may (but not always) get an error message if you run the program from a terminal command line instead of launching from the menu as you probably do normally. Anyway, the fix is to look for such corrupted files. Every line of Javascript should end in a ; -- if the last line in the file doesn't, just delete the whole line, and try to restart. Javascript files are .js in the program's preferences directory. You don't have to understand the actual code to fix this, usually. The program can apparently recover from that point. This usually works for me. It probably isn't the only thing that can go wrong, and it may have nothing to do with your problem, but it's the most common problem I've had with them. The trick is, the symptoms may not always be the same (probably depends on exactly where the problem is in the preferences files and which files are affected). HTH, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Firebird losing toolbar configuration after restart
Some relevant bug reports -- these appear to coincide with the problem I described: http://bugzilla.mozilla.org/show_bug.cgi?id=123929 http://bugzilla.mozilla.org/show_bug.cgi?id=203343 http://bugzilla.mozilla.org/show_bug.cgi?id=223162 http://bugzilla.mozilla.org/show_bug.cgi?id=192425 Since it seemed relevant that this sort of problem needs to be reported. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Linux is not for consumers!
On Thursday 11 December 2003 04:58 pm, s. keeling wrote: Incoming from Richard Kimber: Aren't you missing the point that you need to understand it before you can document it, and that in many cases understanding does not come without documentation. If you need to understand it to use it, you've got the source. What more could you need? That's not good enough? Don't use it. You think it would be better with good documentation? Great. Go write some. Oh, you want me to write some? Why? I don't think it needs it since I have no trouble using it. This defensive attitude is common, but not universal amongst developers. I actually like writing documentation for stuff I write. It often starts before the code does (but that usually means it needs to be revised a lot). Saying that the source code *is* the documentation is not unlike saying the Human Genome is an Anatomy Physiology textbook. :-P I feel that good documentation (not to mention accurate spelling and grammar, etc) is an important craftsmanship issue, and I don't like to short-change it. I don't know how many times I've gone through my code just to fix the bad spelling in comment lines. The most embarrassing thing was when I had to change the API to my Zope product because I misspelled a word in a function name! That was really dumb. The LDP is always looking for volunteers. Hint, hint. The common sense answer to this whole thread is, Don't look a gift horse in the mouth. At one level, you're right -- no one has a right to force you to write documentation if you don't want to. Nor to space and indent your code so other programmers can read it. Nor to use variable names that are remotely related to the things they represent. That doesn't make obfuscation a good practice. And it's good practice that's being discussed here -- which is entirely appropriate for discussion, IMHO. I for one, like to keep improving at what I do. The service-bureau management model is highly flawed for many applications. Given a complete failure of the embedded approach, we turn to it as a last resort. Don't confuse that with success. That's like saying that having a fire department justifies sub-standard wiring. The people who know a program best are the ones who work on its internals. No one else can write documentation like the guy who built the thing in the first place. Failing that, you can have someone step in and write it, yes. But it'll never be as accurate as you'd like, nor as up-to-date. The ideal, for me, is to include documentation within the program in a literate-programming style, and generate the documentation from the code. This solves the other bad problem with Free Software documentation -- being WRONG. Wrong documentation can be worse than none! (With specific reference to Zope and Python code, this seems to be the biggest problem from my POV). IMHO, good documentation means more people using, testing, and (perhaps ultimately) contributing to my project. For those of us for whom those are principle motivators for releasing the code under a free-license in the first place, that's more than enough reason to make documentation part of the task. I also like to write documentation because I have a crummy memory. I write documentation so *I* can use it, too. :-) In fact I just spent all day yesterday reading one of my own manuals. OTOH, when I first released the program, I thought it was so simple that it didn't need documentation. But some of the people on the mailing list I announced to suggested that it really did. In retrospect, I can certainly see that they were right. I like Python for that, because there's so many great introspection tools built into it (I was never able to get anywhere with the original cweb literate programming system). Of course, I come from an academic background, so writing has always been part of my job. Anyway, I just wanted to point out that we're not all such curmudgeons about writing. :-D To the O.P. though -- bear in mind when you make this kind of criticism that you are criticizing somebody's baby. They do tend to get quite defensive. It takes a lot of confidence in yourself to accept criticism of your work without feeling personally attacked. ;-) One thing you get with free-licensed open source software that you don't get with the proprietary beast *is* actual contact with the author. That's unheard of for proprietary stuff! Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Why should non-root users have a password?
On Sunday 07 December 2003 01:28 pm, Tom wrote: If I have a firewall, and I'm the only person who uses my computer, do I really have to have a password on my non-root account? I know the answer is yes but -- why? They can't do anything to my machine anyway, except use it. And due to the firewall that never happens anyway. If you really know that, then the answer is no, you don't need it. But for those of us who are paranoid enough to think that our firewall might not be perfect or that someone might try to access our computer from the console, there are reasons. The first step in most root exploits is to get normal user access, and so it's helpful if that's not too easy. *That* is why you don't want just anybody to use your system. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Moving programs into KDE menu
On Sunday 07 December 2003 08:20 pm, Hoyt Bailey wrote: Can someone advise me on how to move a working program into the KDE menu. I have gimp installed, configured and working but cant figure out how to put it into the menu system. If you mean the system-wide way to do it, I don't know. But if you just mean for yourself, right click on a blank spot on the panel and select menu editor. It should be self-explanatory from there. Oddly, most packages seem to automatically appear in the menu when I install them under KDE (I had to run update-menus for that to work with fvwm). I'm not sure what happens when you run update-menus with KDE (maybe nothing?). Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT - archiving mail
I have never used it, but I understand that procmail is the thing you want: a tool which can process mbox contents and do arbitrary things with them according to headers, etc. Python also has some modules for handling email formats (and I'm sure Perl does too). So there's probably a variety of tools that can do it. But they're all going to involve some kind of scripting, not just flipping through menus in your mail client. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian Server Compromise -- A Fire Drill ??
On Thursday 04 December 2003 12:17 pm, Dave wrote: On Thu, 04 Dec 2003 18:00:18 +0100, Tom [EMAIL PROTECTED] wrote: On Thu, Dec 04, 2003 at 10:15:12AM -0600, John Hasler wrote: ... That's why the kernel developers thought it was just an ordinary bug: they could see no way to exploit it. That statement is somewhat disconcerting. The hypothesis is that many eyes detect secure bugs, and here is clear case evidence contradicting that hypothesis. There is no contradiction. Many eyes detect most security problems, but not all. This is certainly better than just a few eyes with access to proprietary code. There is also the point that *somebody* found this bug. Just not the folks we were hoping would. ;-) Letting real crackers hammer your system is another way to find bugs, although we hope it's a last resort. -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Importing Corel Draw
On Wednesday 26 November 2003 01:34 am, Dr.-Ing. C. Hurschler wrote: I've exported drawings to wmf and then opened them in OpenOffice Draw with no apparant information loss. Right, because WMF is a vector format. -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Importing Corel Draw
On Wednesday 26 November 2003 03:59 pm, csj wrote: I'm actually having problems importing SVG with Sketch (I'm missing a Suggests depends which I'm too lazy to install). But I'm using Sketch as an EPS (PostScript) to SVG converter. Last I checked, the SVG import in Sketch didn't work with text (apparently SVG text is pretty complicated, and the converter isn't fully implemented), but it works okay with curves. As for the dependency -- yes, you need libxml or whatever it asks for for SVG to work. And, since CorelDraw supports exporting to EPS , I would change your step (1) to: 1) Use Corel Draw to convert CDR - PS I don't really recommend this. Postscript is not a general purpose vector format -- it's very specialized for printing, and most programs that generate it, don't make very good code. The resulting files are hard to edit, and often highly lossy methods like converting circles to lots of line segments. It's better if you can use a vector format that's design for editing. My understanding is that CMX is a more recent Corel Draw format, and Sketch supports it directly. That's why I recommended it. Someone else recommended WMF, and that should work too -- WMF is supported by Sketch and apparently Open Office. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Importing Corel Draw
On Tuesday 25 November 2003 01:07 am, Thanasis Kinias wrote: scripsit Luis Fernando Llana Díaz: I have some pictures made with Corel Draw. Is there any possibility of use them under Linux? I haven't been able to find any. I have many of these, too, from an earlier career -- as well as many WordPerfect and QuattroPro files. No one has ever been able to point me to a converter for any of these, and neither the Gnome apps nor OpenOffice can read the files last time I chedked. :( If anyone knows differently, let us both know, please! Sketch (http://sketch.sourceforge.net = sketch package) can import Corel Draw CMX format files. The author and some of the more involved users are former Corel Draw users. HTH, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Importing Corel Draw
On Tuesday 25 November 2003 03:36 pm, Andreas von Heydwolff wrote: There once was a free Linux port of PhotoPaint 9 from Corel running on Corel Wine. It was buggy but I had it running quite nicely on my potato system two years ago. I think it had import filters for CDR. [...] Why not grab a spare machine, throw an older distro at it and install PP from some leftover archive on a forgotten ftp server. Seeing as PhotoPaint is a *bitmap* program, this would not truly import the CDR data, but *render* or *rastorize* it. Which is probably not what the original poster wanted. It would be far better to use a tool which can actually convert *vector-to-vector*. Sketch does this with CMX files, but I don't really understand the relationship between CMX and CDR (I stopped using Corel Draw around version 3.0, and there was no CMX format then). I would *guess*, though, that there is some intermediate version of Corel Draw which can read and write both formats. So I'm suggesting: 1) Use Corel Draw to convert CDR - CMX 2) Use Sketch to convert CMX - sk (I think you can probably also get SVG out with Sketch, though I'd have to check that -- it can certainly import SVG, but I'm less sure about exporting). Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Workspace/desktop switching
On Tuesday 18 November 2003 01:19 am, Alex Malinovich wrote: I've been seeing a lot of discussions about various WM's lately, and everyone seems to be extremely concerned about easy workspace switching. I'm just wondering what exactly everyone uses workspaces for? Use #1: Efficient parallelizing Obviously it's going to depend on what you do with your machine, but I do a lot of time intensive hand-off tasks. This often has to do with download/upload times, but also sometimes with disk-access, or even CPU time on a remote machine (won't help if it's local, unless we're talking about multiple processors which I don't have). Why wait for a long download when you can be browsing the web or working on your development project at the same time? And switching desktops is usually *much* faster (2-3X) than opening/closing individual app windows in actual computer time, and even faster when you consider that it takes fewer mouse clicks. Use#2: More space to spread out in [Doing this now] Just as an example (and this incorporates some of #1) I develop Zope products (web apps). I keep a master copy on the file server, and periodically mirror it into the Products directory and refresh as part of the development cycle. During this time, I usually have one desktop with two gVim windows (I prefer separate windows over the internal panes, BTW), one on the main source module I'm editing and one on to browse files it depends on -- usually to read, but I sometimes need to make edits there. Then I have an xterm logged in as the Zope user which I use to mirror the results. I use a 2nd desktop to have a browser window pointed at the local Zope server, usually with several tabs addressing the Product refresh page, an object-tree browser page, and the page where I actually see the output. I use a 3rd desktop with a tabbed browser pointing at the local Python manual, Zope.org, Python.org, and google for researching questions as they come up. For testing: 1) save source files, 2) up arrow and enter to run the cp command that copies the sources to the Products directory, 3) swap desktops, swap tabs, click refresh wait for response, 4) swap to output screen, click reload and check the results. It's not quite as simple as compile, link, run, but web apps are like that. There's also a unit-testing mode which hopefully I've mostly completed before getting to this level. I use a 4th desktop to hold my email client. I may have one or two emails open at a time. Sometimes I'll start writing a question to a list and then realize I can figure out the answer for myself and stop. Of course, I also use it to take a break and answer a question myself. Like now. Sometimes I need to swap over to my reference desktop to check something about my answer or verify a URL. So that's four. Right now I have two more in use, because I also have a separate development project that I'm just collecting information for. That's three xterms logged onto remote machines at my clients site: one each on two machines (different architectures, as I have to install software for both), and one with w3m running in the window. I'm using it to download software package files. That's on desktop #5. The 6th desktop just has XMMS in it, because I'm listening to music. Occasionally, I minimize apps. But as I said, it's usually more faster to switch desktops than to go to and from the taskbar. I used to be limited by the CPU speed and RAM (2 or 3 big apps would strap the computer), but I've upgraded, so now I'm mostly limited by how much I can think about at once (which is how it ought to be ;-) ). Note this is KDE 2.2.2 that comes with Woody. My biggest complaint is that I haven't figured out how to sweep virtual-desktop style from desktop to desktop (like I could with FVWM), but must manually click which one I want. I figure there's probably a setting somewhere that controls that (or will be in KDE 3?), but it's not really such a pain -- I've already gotten used to it. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: classic deficiancy in both windows and linux ?
On Thursday 23 October 2003 05:08 pm, James D. Freels wrote: Can you figure out a way to get a listing of a directory (folder in Windows) and print it, without resorting to command prompt ? I was about to pooh-pooh this a bit, but in fact, he's right: both Konqueror and Mozilla (at least) lack the ability to print a directory listing, though they will show it on the screen. Mozilla even responds with a cute We are unable to print this page message. Hmmph. This is with Konqueror 2.2.2 and Mozilla 1.0.0, things may have improved since then. Surely some file browsers do have a print option? Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: classic deficiancy in both windows and linux ?
On Friday 24 October 2003 10:43 am, Mental Patient wrote: it. The one thing perl has on python is that there's an imagemagic module for perl. My python scripts are doing lots of os.system() calls :) Look for PythonMagick (caps may be different). There was one, then it died and lost compatibility with Image Magick changes, then somebody else picked it up and/or re-implemented it (I think this is the Boost implementation, IIRC). I'm not sure what the status is now. I opted to use Python Imaging Library from http://www.pythonware.com/products/pil/ to handle this kind of stuff, largely because of the availability/compatibility problems with PythonMagick. I'm probably going to reconsider that though, if the PythonMagick situation becomes better, because libmagick does some things much more nicely than PIL does (the reverse is true too, though). Just wanted to pass along some hints for when you decide to get rid of the os.system() calls. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: classic deficiancy in both windows and linux ?
On Friday 24 October 2003 09:57 am, Craig Dickson wrote: Presumably because he normally uses a GUI file manager (Konqueror or some equivalent) and it would be convenient, if one already had a window open on some directory, to simply select Print from the menu to get a hardcopy of the directory listing, in the same format as it appears on the screen. Essentially, he's suggesting that a view of a directory in a GUI file manager should be considered a document that can be printed, just as one can print a web page from a browser or a text file from an editor. This seems perfectly sensible to me; I've never understood why file managers don't provide this functionality. Indeed, and Netscape does this. Why doesn't Mozilla, I wonder? (as of 1.0.0 anyway -- maybe it does now?). I agree that this is an oddly missing feature. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How Do You Know If It Works In Linux?
computer in the reviewer's opinion (actually 3 different ones for 3 different price/ performance breaks). Unfortunately, I can't remember where, and I'm not finding it with Google. This information is obviously ephemeral -- such a site would have to be maintained, and it may not have been. (?) Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
OT: Re: patents, Re: Multi-user Debian
On Sunday 28 September 2003 12:56 pm, [EMAIL PROTECTED] wrote: It's completely out of control. Be afraid that someone will patent the act of typing on a keyboard, or of breathing in and out, and try to charge you a royalty. Or *hope* it happens and makes the revolution happen sooner. Maybe enough absurd cases will finally motivate eliminating or reforming the patent system. Hey, I can dream, can't I? Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dselect, was Re: OT: RH and Debian brothers now?
On Friday 26 September 2003 10:29 am, Pigeon wrote: On Fri, Sep 26, 2003 at 02:52:12AM -0700, Paul Johnson wrote: On Sat, Sep 27, 2003 at 08:26:32AM +1200, cr wrote: I appreciate that dselect is only part of the install process, albeit the largest part timewise if one uses it. You don't have to, though. Yeah, but what are the choices? Run tasksel (y/n)? (frequently too coarse) Run dselect (y/n)? (we all know about this one :-) ) Do it by hand afterwards - somewhat inconvenient and daunting for a new user Yeah, I agree that dselect is awful (to be fair, it probably wasn't so bad when there were fewer packages to wade through). I have stopped using it, myself. I recommend using tasksel to rough out the system, then using apt-get to finish the job. The tasksel will give you a working system with the basics you need that you may not know by name. Then the apt-get will allow you to ask for everything you specifically know you want. The downside is that it may be a little hard to find the correct package names by what they do or what commands they include. I have only just recently learned that there are ways to query this on the command line (but they may only work for installed packages?). Anyway, myself, I make used of the packages search on the Debian website to make those determinations. Usually, you can just use apt-get to load stuff as you need it after that. And of course, once you get a collection of packages you like, you can use the dpkg --get-selections and dpkg --set-selections to save and retrieve your choices or replicate onto multiple computers. HTH, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: RH and Debian brothers now?
On Tuesday 23 September 2003 06:03 pm, Paul E Condon wrote: RedHat's business model is moving toward support services for enterprises and away from sale of boxed sets of CDs. I don't think it makes much sense for them to continue work on the RedHat Linux distribution, but I can see why they might want to pretend to do so. Their corporate customers probably wouldn't notice if RedHat started loading Debian onto the corporate computers, so long as Red Hat, the company, continued to provide support. I think they would save themselves a lot of head aches if they did move to Debian. This collective support of the RedHat distribution, without selling CDs looks to me like Debian done badly. It will wither away, and the people will drift into the Debian community. It occurs to me that that might very well be a near-ideal and possibly planned outcome. If I were running Red Hat, I would probably think so. No need to scent conspiracy theories -- it's not a bad approach to transition: gives plenty of time for the two to grow towards a standard (e.g. the LSB stuff), and then make the transition smooth for the end user. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Slide show software
On Monday 22 September 2003 01:20 pm, William Bradley wrote: Does Debian in their deb files have anything that will allow me to run my photo's as a slide show on my computer? Sure, depending on what you want exactly: imagemagick (probably closest to what you want), e.g.: % display --delay 300 --geometry 800x600 mypictures/*.jpg kpresenter (can create a presentation) kde (can rotate the desktop wallpaper) And I'm sure there are other choices too. I used to have a nice hack to set my FVWM root window to a random image on login (I used signify and xloadimage to do this). Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get problem (bug in apt-get?)
On Monday 22 September 2003 08:32 pm, Darryl Barlow wrote: Are you running unstable? If so, as of the last time I checked, the kdemultimedia package is broken. A bug report was filed some time ago but it has not yet been fixed. This is one of the joys of running unstable and the price to be paid for running a more cutting edge distribution. Personally I find it remarkably stable but these things do happen from time to time. Well, I'm actually running a mixed system with stable preferred (but libc6 and other development packages are from unstable). So it's not a flat single distribution system -- and that may indeed be why I'm running into this kind of grief. The version of kdebase-audiolibs I have is from testing. But the point is, I'm not having ANY technical problems with kdebase-audiolibs (my KDE environment has sound effects which I can hear as I use it, and games etc all have sound as expected). But apt-get is refusing to *recognize* that it is installed at some times (when it's a dependency), but insisting that it is at other times (when you try to install it or when it conflicts). I'd say your problem is with the version of the package rather than the package itself not being installed. I think you will find all will install fine when the bug is resolved. If you are really desperate you can force the install of whatever package has the unresolved dependency and hope it works or that you can make it work by symlinking libraries etc,. or perhaps install from source. Like I said -- it IS installed. Dpkg thinks so, and even apt-get thinks so when you ask it to install it again. And it works -- I use it everyday. But apt-get still complains that that it's not installed whenever I try to install something that depends on it. And it's NOT a version issue, because non of the packages in question ask for a specific version of the package, they just depend on kdebase-audiolibs OR kdebase3-audiolibs. I can work around it by using --reinstall and including kdebase-audiolibs on *every*single*usage* of apt-get. But that's just really stupid. Especially since that actually does cause it to reinstall needlessly. So you see, apt-get is making logically contradictory claims. So what we're talking about is a package database that contains a logical contradiction. And furthermore, the claim that the package is not installed is refuted by the simple fact that the program it installs is working just fine. So I'm still inclined to suspect that there is something wrong in the database, and therefore in the programs that manage it, rather than a problem in the package (of course, maybe the package has some irregularity that triggers this mismanagement, but that still should be under apt-get's control). Hopefully apt-get is not designed to assume that packages will cooperatively manage the database?! (Remember how successful cooperative multi-tasking was?). Unfortunately, even if this is all true, I don't have any way to know when or on what stimulus the problem occured, so it'd be nearly impossible to identify a specific bug to report. I was more wondering if anybody knew how to verify the contents of the Debian package database. I pretty much treat it as a black box, and never touch it. :-) The better option is probably to wait. Ah, but for what? ;-) Well, it's not crippling at the moment -- it's only blocking installation of a few games my kids want, and there are other ways around that. It's just a matter of maintenance. Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
apt-get problem (bug in apt-get?)
Hi, I suspect this is an actual bug in apt-get, but I'm not sure where the problem would be, or when it occured. Anyway, it looks like it must be some kind of database inconsistency, and I'm really more interested in fixing my package database. Lately, I've started to get this message when trying to install a package: kde: Depends: kdebase-audiolibs but it is not going to be installed or kdebase3-audiolibs but it is not installable Well, for starters, my sound works fine in KDE, so it seems quite improbable that kdebase-audiolibs isn't installed. And indeed, dpkg says I'm right: % dpkg --status kdebase-audiolibs [...] Status: install ok installed [...] So what if I try to install kdebase-audiolibs? % apt-get install kdebase-audiolibs [...] Sorry, kdebase-audiolibs is already the newest version. [...] So, here, even apt-get agrees that kdebase-audiolibs is installed. So what if I force it to install? % apt-get --reinstall install kdebase-audiolibs [...] Unpacking replacement kdebase-audiolibs ... Setting up kdebase-audiolibs (2.2.2-14.4) ... Well, that *seems* to work (it downloads the package and installs it). But unfortunately, the behavior is exactly the same as before -- when trying to install packages which depend on kde, apt-get continues to claim: kde: Depends: kdebase-audiolibs but it is not going to be installed or kdebase3-audiolibs but it is not installable So I'm right back where I started! How can I fix this, then? And how can apt-get simultaneously think that a package is installed so that it can't re-install it, but then think that *isn't* installed when I need it? Clearly the two opinions contradict, so apt-get's database (the package database) must now be inconsistent in some way. That's why I think it has to be a bug in apt-get and not something wrong with kdebase-audiolibs that gets us to this point. Right now, though, it's probably a corruption of my package database. Anybody know how to check it or correct it? TIA, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Mozilla/Konqueror + Kaffe JVM?
Hi, I just upgraded my system to Debian 3.0 and am happily playing with Konqueror and Mozilla which are now real replacements for Netscape which was starting to show its age. NS was actually my last proprietary software package, so I'm feeling really good about that. :-D There is one problem though -- even though I opted to install Kaffe and Jikes as Free Java replacements, the browsers are not apparently configured to use them. I kind of figured dpkg would work that out for itself (I've gotten spoiled, I guess ;-)). Going to the Mozilla webpages I found instructions for installing a Java 1.4 plugin, which I could obviously do, but I really want to run a Free version. The main reason for this is that I'm going to be developing some Java-based web applications, and I want them to work in this Free-software-only configuration. So I'd like to have at least one of these two browsers configured to run a Kaffe JVM/JRE so I can test in what is basically going to be a stressed configuration. I'm hoping it's actually possible to get a Java 1 / AWT / Jython based applet to run this way. But the first step is obviously getting the browser to use the JVM in the first place. I opted against using Java 2 / Swing / Jython because of the SCSL licensing issues described in http://www.debian.org/doc/manuals/debian-java-faq/ as well as some performance issues I was warned about, BTW. I'm assuming that's still true. For reference, starting Jython on the command line gives me this message, which leads me to believe I actually do have a working system: Jython 2.1 on java1.0.5 (JIT: kaffe.jit) Although I do see some messages like this: *sys-package-mgr*: skipping bad jar, '/usr/share/kaffe/tools.jar' Hmm, while I'm at it, I notice that: import awt import java.awt give me errors like: import java.awt Traceback (innermost last): File console, line 1, in ? ImportError: no module named java even though kaffe.org claims to have AWT. Hmm, that doesn't sound so promising. MInd you, I'd be happy with any Free JVM -- Kaffe just seemed like the natural choice when I was looking at the Debian packages. Any advice greatly appreciated! Cheers, Terry -- Terry Hancock ( hancock at anansispaceworks.com ) Anansi Spaceworks http://www.anansispaceworks.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
jigdo checksum error
Hi, I'm trying out jigdo to retrieve a Debian 3.0r0 CD image set, and after quite a long download, I get a checksum error. Is there any way to fix this short of starting the whole process from scratch? (I really hope so). I used http.us.debian.org to get the .jigdo file, and mirrors.kernel.org as the Debian Mirror site. Both are listed as Push-Primary mirrors on the Debian site ( http://www.debian.org/mirror/list-full ). I originally tried using http.us.debian.org for both, but this seemed to be very slow (perhaps this mirror restricts bandwidth or is simply very busy?). I also tried using mirrors.kernel.org for both, but it doesn't seem to have the jigdo files (or not in the same place anyway). I have downloaded other CD ISO images through my connection before, and they don't take this long, so the claim that jigdo is faster doesn't seem to hold true (though I can see that it might be a big help to the servers, which is sufficient justification for it). Basically, I'd like to ask if I'm doing this right -- is it okay (preferred? discouraged?) to use two different push-primary mirrors like this? Also, is the slow download simply to be expected, or the result of some error I'm making. Why would I get a bad checksum? Simply corrupted files, or is there a chance that I'm somehow getting the *wrong* files? Thanks for any advice you might be able to offer, Terry -- -- Terry Hancock Anansi Spaceworks http://www.anansispaceworks.com -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
dpkg database repair
Hi, Apt-get and dpkg work so well I hardly ever have to think about how they work. Unfortunately, this means I'm pretty clueless when something does go wrong. I have two computers which apparently have damaged dpkg systems -- the status or package databases seem to be corrupted or destroyed. For example, attempting to run apt-get update results in this error message (after checking for packages data from the Debian site): Reading Package Lists... Error! E: Could not open file /var/lib/dpkg/status - open (2 No such file or directory) E: The package lists or status file could not be parsed or opened. (looks like I've made a blunder somewhere along the line!) The other machine constantly sends email to the administrative email address complaining that various packages aren't being found in a database of some kind (I forget what the exact message is). It seems like there ought to be a way to have dpkg do a consistency check/repair of some kind (I though of reinstalling dpkg, but then do I need dpkg to install dpkg?). I'm sure there must be some facility for doing this, but I haven't found anything by just doing searches. Is there a standard way to fix a corrupt dpkg system? The only thing I can think of at this point is to manually backup all the user data areas and reinstall the whole system, but that's pretty painful, so I'd really like to try something less destructive. Thanks, Terry -- -- Terry Hancock [EMAIL PROTECTED] Anansi Spaceworks http://www.anansispaceworks.com P.O. Box 60583 Pasadena, CA 91116-6583 -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
package size stats
Often, I wish I could extract certain information out of the package database, such as the size, age, and distribution the package came from. This would make it a lot easier when deciding what packages can be removed to make room for others (i.e. very old, too large, etc). Also which ones I ought to upgrade and so on. Right now I have packages from three separate Debian distributions, and probably a lot of stuff that's just taking up disk space. It'd also be neat to be able to tell if a package is depended on by another, though that sounds much harder to do. It's clear that the data is in there, since apt-get can compute the consumed disk space of a potential installation before doing it. Does anyone know how to do this? Please CC. Thanks, Terry Hancock -- -- Terry Hancock [EMAIL PROTECTED] Anansi Spaceworks http://www.anansispaceworks.com P.O. Box 60583 Pasadena, CA 91116-6583 --
Linux ALSA sound ThinkPad 380ED
[Cross-posted to debian-user and alsa-user, not sure which is more appropriate] Okay, I have a ThinkPad 380ED (Actually two of them -- so I'll refer to them as #1 and #2, I'm installing #1 while using #2, which has a pre-installed Windows 95 OS and working sound). I'm trying to install ALSA (=Advanced Linux Sound Architecture http://www.alsa-project.org ) sound support onto it. The problem seems to be that I don't know whether this is actually a plug-and-play (PnP) device or not. If not, I do not see how to select parameters for it. If so, I am not able to detect it with the isapnptools package in Linux, which reports no boards found. DETAILED DESCRIPTION OF PROBLEM: Supposedly this has a CS4236B chipset sound system (and I have no reason to doubt this, though I'm just quoting the specifications from other sources). Sound was working fine with pre-installed Windows. It still does on #2. Laptop #1 however, has now got Debian Potato 2.2 Linux installed. Kernel 2.2.17. I download ALSA support for 0.4.1i from the testing version of Debian. From the Windows INI file on Laptop #2, I see the following information: [PNP] WssIO=534 WssInt=5 WssDmaPlay=1 WssDmaCapture=0 SbIO=220 OplIO=388 OplInt=Disabled GameIO=200 4232IO=538 4232Int=Disabled MPU401IO=330 MPU401Int=9 CDIO=Disabled CDInt=Disabled CDDma=Disabled A peek at the Device Manager in windows, gives: Crystal PnP Audio System CODEC/Joystick I/O range 0201-0201 I/O range 0534-0537 I/O range 0388-038B I/O range 0220-022F IRQ 05 DMA 01 DMA 00 Crystal PnP Audio System Control Registers I/O range 0538-053F (I'm not too familiar with Windows, so I'm just making educated guesses about the meanings of these lines in both the INI and the Device Manager). It appears that Device Manager is only reporting the resources consumed, and not what their function is, though the INI is a bit more informative. I know the 0x388 port is the old Adlib compatibility FM synth port, which was preserved in the Soundblasters. I also know 0x220 is a default Soundblaster port. Manually configured sound cards usually use IRQ 5 or 7. Not really familiar with the rest of it. I have tried various permutations of these numbers in the alsa configuration (/etc/modutils/alsa). Here is the present incarnation, plus some notes about variations I tried (on Laptop #1): # --- BEGIN: Generated by ALSACONF, do not edit. --- # --- ALSACONF verion 0.4.2 --- alias char-major-116 snd alias snd-card-0 snd-card-cs4236 alias char-major-14 soundcore alias sound-slot-0 snd-card-0 alias sound-service-0-0 snd-mixer-oss alias sound-service-0-1 snd-seq-oss alias sound-service-0-3 snd-pcm1-oss alias sound-service-0-12 snd-pcm1-oss options snd snd_major=116 snd_cards_limit=1 snd_device_mode=0660 snd_device_gid=29 snd_device_uid=0 options snd-card-cs4236 snd_index=1 snd_id=CARD_1 snd_port=0x538 snd_cport=0x538 snd_mpu_port=0x330 snd_fm_port=0x388 snd_irq=5 snd_mpu_irq=9 snd_dma1=0 snd_dma1_size=64 snd_dma2=1 snd_dma2_size=64 # --- END: Generated by ALSACONF, do not edit. --- Most of the relevant parts ar in the options part. I have tried: snd_port= 0x538, 0x534, 0x53f, 0x53e 0x53d 0x535,0x220 (with 0x534, 0x538 being tested the most) snd_irq= 5,6,7,8,9,10,11,12,15 (actually some of these aren't supposed to be legal) snd_mpu_irq=9,11,12,15 With any of these I get the same behavior: Starting ALSA sound driver (version 0.4.1i):isapnp: No Plug Play device found snd: isapnp detection failed and probing for CS4236+ is not supported snd: CS4236+ soundcard #1 not found at 0x534 or device busy snd: CS4236+ soundcard #2 not found or device busy (cs4236) I thought that the 0x534 part changed according to what I had snd_port set to, but I tried changing this to 0x220 while working on this e-mail, and it didn't change. I DID use update-modules and modules.conf shows my selection of snd_port. I don't know what that means, but it suggests that my settings aren't being used somehow. Okay I investigated this a little just now -- using modprobe with the snd-card-cs4236 and the above options shows the failure at the address I selected, but if I edit the config file, run update-modules, and then either reboot or stop and start alsa, I get the message referring to 0x534. Hmm. I've been struggling with this for four days now, so I'm getting pretty frustrated with it! If you can offer any suggestions, I'd really appreciate it! -- Terry Hancock [EMAIL PROTECTED]
xbanner w/ X4.0 upgrade
Hi, I recently had to upgrade many packages on my system to the unstable Woody distribution, which included upgrading to X 4.0, which -- not surprisingly -- broke scads of things, most of which are now fixed. The one continuing annoyance is that xbanner is no longer doing anything! I suspect this has to do with relocating the config files or possibly with the xdm configuration files themselves. I know there were changes in the organization of the config files (for X). I note that xbanner was not upgraded to a new upstream version going to Woody (both Potato and Woody use v1.31), although the Woody version is a later package release. I assume this means that xbanner is expected to continue working -- but it appears that it doesn't. At least not without modification or re-configuration. Has anyone else noticed this? And what exactly do I have to do to recover my xbanner setup? Any help would be appreciated -- I really liked my old login screen! And it's so much harder to intimidate Windows users this way. ;) -- Terry Hancock [EMAIL PROTECTED]
X 4.0 Upgrade /etc/X11/X ?
Well, I needed a bunch of library packages from the unstable (Woody) release, so I tried using apt-get to do it, and it decided I needed to upgrade the X packages as well. So okay, I wanted 4.0 anyway, but here's the problem (after eliminating some intervening steps): $ X /tmp/x.out 21 X: Cannot stat /etc/X11/X (No such file or directory), aborting (This test was recommended by the Xfree86 HOWTO -- the same thing happens with startx, and xdm, while it doesn't visibly complain, also doesn't work -- no errors, but no visible effects either, after a few minutes it shuts down again). Indeed, there is no /etc/X11/X, but I don't know what should be there. Is it a symlink? If so, to what? I believe xf86config was run during the install process, as I had to answer some questions about my video card, mouse, and monitor. Of course, I was previously using XF86Setup, so the interface has changed visibly from what I was used to -- but it appeared to be pretty much the same information. Now, of course, I did not force all of packages to upgrade to Woody, just the ones following the dependencies I needed, but I think I have all the new X packages despite the changes (but something could've fallen through the cracks). Please CC me -- I'm using a Yahoo mailer because my usually mail system is down too. Fortunately Lynx still works! :D Thanks! Terry Hancock reply to -- [EMAIL PROTECTED] __ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/
Re: X 4.0 Upgrade /etc/X11/X ?
Thanks! This was the problem. Now I have X windows again. Some other stuff is not working completely (the fonts look _bad_ all of a sudden), but I'm running again. Thanks for the help (and in 15 minutes, too!), Aaron and the two others who replied. Terry Hancock Indeed, there is no /etc/X11/X, but I don't know what should be there. Is it a symlink? If so, to what? This file should be a symlink to your xserver. Under xfree86 4, it should look like this: X - /usr/X11R6/bin/XFree86 If you're running xfree86 3, it will point to XF86_driver such as XF86_SVGA for a generic video card. __ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/
Unstable packages on Stable distr.
Hi, I have only one computer (working anyway), which I use for fairly important information. So, I use the stable Debian distribution on it. However, I also do development on this machine, so I often need later versions of libraries and so on that I'm using in my projects. If I could just install the unstable versions of those packages it would save me a lot of trouble, because otherwise I wind up having to install from source. The reason is that the unstable packages seem to have the assumption built into them that they will never be used on a stable distr system -- that is they have dependencies on later versions of basic packages. In many cases these seem to be frivolous assumptions. It seems very implausible to me (for example) that compiling libsdl1.1 _really_ requires libc6 = 2.1.97. I'm pretty sure that, installing from source, libc6 version 2.1.3 (in Debian 2.2) will work. So why this dependency? This appears to exist only to force using the unstable distribution throughout -- not because the software would actually require it. (?) And this happens all the time. But if libc6 2.1.97 is defined as an _unstable_ package, is it going to break my stable distribution? Is it buggy? It should just differ by a patchlevel, right? So, how have people done this before? Is it possible to run a mixed box like this, or am I just out of luck? I suppose I could just acknowledge that I'm going to have to install development packages from source -- but if so, why have the development packages at all? Thanks for any suggestions. Please reply direct or CC me. -- Terry Hancock [EMAIL PROTECTED]
A Solution: LILO stops at LI
I noticed a few different posts from people having similar problems to mine (a couple of examples follow), in which LILO would stop after just typing LI on the screen and hang up. Various different reasons for the problem were proposed, centering on hardware compatibility problems. After examining the changelogs, I got the impression that LILO went through a major clean-out of obsolete features between the versions used in Debian 2.1 and Debian 2.2, so I suspected that the Slink LILO might work better. And, indeed, when I downgraded the lilo package to the one on my Slink (Debian 2.1) CDs, this fixed my problem. As you might expect, there's no major problems associated with using the older LILO with the newer distribution (nothing really seems to depend on it, since all it does is load the kernel, which then takes over). The computer I'm using is pretty old -- most likely I need one or more of the obsolete features in order for it to boot. Anyway, for the inexperienced, downgrading a package is simple, with the Debian 2.1 CD Binary CD #1 in the drive (mounted at /cdrom): dpkg -i /cdrom/dists/slink/main/binary-i386/base/lilo*.deb ought to do the trick (check the directory first with ls to make sure this is right). You don't have to do anything special to downgrade a package, it's allowed by default. After installing, you will be prompted for whether you want to run LILO to create a new lilo.conf and install the new boot block right away. The instructions above this warn you not to do it, but it seems to be the right thing to do. I made a backup of /etc/lilo.conf (copied it to lilo.conf.old), before doing the downgrade. Oh yeah, of course, in order to get this far, I had to have booted without a working LILO. Use the CD (or boot-disk) to boot in rescue mode, i.e., at the boot prompt, type: rescue root=/dev/hda1 (or whatever the correct boot device is on your system). I thought this might be worth posting in case other people run into this problem. This is perhaps not the ideal fix, but it's easy. Jeff Davis wrote: I have had this problem before as well. I do not know what causes it, although it usually happens after some kind of hardware issue. Reinstall LILO and it should work fine. I am not extremely familiar with this process, but there are good man pages (man lilo and man lilo.conf). Here are some ideas I have: I have not had this problem in a while. I assumed that the newer versions of lilo are rid of this. What version are you running? Andrew McRobert wrote: I tried adding a ULTRA SCSI hard disk to the server [] If I disconnect it, ie. take my system back to exactly the way it was before, LILO stops at LI???!?!?!? -- Terry Hancock [EMAIL PROTECTED]
How to set up ISA SCSI card.
Briefly, does anyone know how to set up a jumpered ISA SCSI card properly? I have the I/O and IRQ settings on the card, but don't know where to put the information. The card itself (Adaptec 1502) is listed in the Linux Hardware HOWTO as supported, but I don't know what modules to load, what arguments to give them, how to get them configured at install time, or where the documentation is. I posted before, but maybe at a bad time. Sorry for the repost, anyway. Thanks! -- Terry Hancock [EMAIL PROTECTED]
Re: I can burn CDs, but can't mount any CDs.
John Travis wrote: [In response to a previous poster who said:] I have no problems burning CDs, its just that now I can't mount any CDs - data or audio. /dev/cdrom is the correct symlink to my CD-ROM device (not CD-RW), and this was all working before I changed the kernel to be able to burn CDs. What is the best way of configuring the kernel so I can use both my CD-ROM (Creative 52X IDE/ATAPI) and CD-RW (Yamaha CRW2100E IDE/ATAPI) ? I have seen similar docs saying that you can't have regular ide-atapi and scsi emulation support at the same time, but it isn't true. So go back an include ide-atapi support for your creative, and scsi-emulation for your yamaha (I would just do it modularly). Then set up lilo or grub or whatever you use so it knows you want to use emulation on that specific drive... I was never able to make the above fix work (i.e. have both ATAPI/IDE CD-ROM and a ide-scsi CD-RW). However, the penalty is not so awful: it's just that the CD-ROM is now a SCSI device. I used to have the CD-ROM as /dev/hdc, but after installing ide-scsi support, it and the CD-RW are /dev/scd0 and /dev/scd1 respectively, for mounting disks. The CD-RW is refered to as /dev/sg0 though, when I write to it. In general, scd# is a SCSI CD Device and sg# is a SCSI Generic Device. The original poster can probably solve his problem by changing the /dev/cdrom symlink to point to the correct device. Good luck. -- Terry Hancock [EMAIL PROTECTED]
Re: cvs - howto? (aviplay)
Michael Mertins wrote: I'm very interested in the development of avifile. Now I heard it's going on and is updated by 'cvs'. How can I use it on my Debianbox? Do I need a package to get it all started? Btw: Is anyone going to build deb-packages of avifile? CVS = Concurrent Versioning System This is used by programmers to continuously maintain a working version of the software as they work on it. If the program is only available in CVS, then there have been no releases, and so the software is pre-alpha, which means it's probably awfully unstable. If you are only up to installing from deb packages, you'll probably find this extremely frustrating. It's basically one step harder than installing from source: i.e. installing from source _with_ _bugs_. :) I have done this only for software I was actively involved with, and it isn't really that much fun. Stick to releases if you have the option (this means you may have to wait for them to make a release). Debian packages usually follow only after a release has been made. I understand that one of their lists can be used to post a request for package or RFP you might want to look for that (I haven't used that though). -- Terry Hancock [EMAIL PROTECTED]
ISA SCSI card in Python
I have an Adaptec 1502 SCSI card which is of the pre-PnP ISA jumpered variety. Refreshingly, it is documented on the silkscreen, so I know it's set up for I/O=140h and IRQ=9. But, in installing Debian 2.2, I didn't see where you can give it this information. It seems like there should be a driver module for it, but there's no specific mention of this card, and scsi-generic doesn't seem to take any arguments (that is, there's no dialog box to enter arguments for it). Am I missing something obvious here? Surely this doesn't require a custom install disk or anything -- SCSI is popular and this card is listed in the Hardware HOWTO as a supported one. An (unrelated?) problem involves the booting process: when I try to boot directly from the hard drive, it stops after printing the LI (from LILO) and hangs up indefinitely. I've seen this before, but I can't remember where. Anyway, it worked with Debian 2.1, so I don't really understand why it would mess up now. What is it trying to do at that point? Thanks in advance, -- Terry Hancock [EMAIL PROTECTED]
Stock kernel config info?
Is it possible to determine what the configuration settings were for the stock kernels that Debian packages? I need to recompile the kernel (1st time), and I would like to change as little as possible, since the stock kernel is working well for the most part. (I just need to enable ide-scsi module support for a CD-RW install). Also, it appears that you cannot simply compile the modules and install on the existing kernel binary (I tried this and got a bunch of error messages at start-up, and lots of stuff stopped working, including my parallel port). I am surprised by this, expecting that the modules would operate more independently, but I suppose not? Please CC direct to me, as I am only on this list intermittently. Thanks, -- Terry Hancock [EMAIL PROTECTED]
Re: Compaq 5300
Jens B. Jorgensen wrote: I'm not sure quite what you mean. Do you mean you don't have any room on The Millers wrote: Dear Wonderful Deb_Ian People, I would like to move forward with exploring a Linux system using a Compaq LTE 5300 laptop. The method by which I can load the kernel/OS is limited to diskette. What would be the best way way to proceed from this point using Debian's Linux? Thank You Very Much for Your Time, Jerry Miller Well, assuming that he means that he only has the ability to install using floppies (i.e. he is intending to install to the hard drive as normal). This is my recommendation: 1) Download and create the rescue disk boot floppies according to Debian installation guide. 2) Install the base system on the laptop, with the appropriate configuration options for whichever of the following is easiest for you: A) Install a laptop modem (PCMCIA type, for example). Should cost about $50. (?) Of course, if you've already got a modem, you're in luck. B) Ethernet card (assuming your laptop has PCMCIA you can probably pick one of these up for $50-$100). Of course, this is only useful if you have an internet-connected ethernet to which you can attach the laptop, but this was the method I had selected. C) Parallel port CD-ROM install. You'll need one of those laptop CD-ROMs that install through the parallel port. This is probably $100. D) *FREE* Install SLIP or PLIP and network using the parallel or serial ports on the laptop. (But, sorry I don't know how, I just know you can do it). I'd guess this is the hardest, but it has the advantage of not requiring new hardware. 3) Proceed with the complete install (via dselect) using the source method that you chose above. I HAVE NOT TRIED THIS, and I'm no expert, but I did work this out as a plan for myself, when I was considering buying a 701 Thinkpad, and I thought it might be useful to you. Best of luck. P.S. -- Mr. Miller, is your username really untitled ? :) -- Terry Hancock [EMAIL PROTECTED]
Re: printcap problems, please help
Hi, I had a look at your printcap file and compared it to mine. I can't find anything wrong, so let me tell you how I tested my printer: 1) Test hardware + basic devices: ls -l /dev/lp0 This sends the directory printout to your printer device directly (i.e. without any filtering). Unless you have one of those Windows-only (host-based) printers this will usually produce some kind of output. In any case, if it doesn't you have a more basic problem than your printcap. Maybe you've done this already, though. Also makes sure you're using the right device name. 2) Make sure that your magicfilter file exists and is executable. Of course, if it isn't, you shouldn't hear anything from the printer. 3) Sometimes if the printer reacts, but doesn't print, it means it's waiting for a form feed. On some printers you can give it one from the printer's front panel. Otherwise, you can feed one straight to the device to check. (I don't remember how to do this offhand, but I know you can). ALSO, I just noticed that with the Linux 2.2 kernel, your first parallel port (~LPT1 in dosspeak) is /dev/lp0. However you mentioned Slink, which is based on the 2.0 kernel. With that one, I had the first printer port as /dev/lp1 in my printcap (I had to fix this when I upgraded the kernel). Well, maybe one of these suggestions will help. Steve Winston wrote: I need a little help getting a printer working with Devian. On slink, I mistakenly erased my printcap while setting up an Epson Stylus 740. Because I also have Linux mandrake on the same computer with the same printer, I installed the mandrake printcap and then did /usr/sbin/magicfilterconfig --force to convert the printcap to Debian. But when I try to print something (e.g., lpr something?), the printer makes a noise and then stops. Enclosed is my printcap file. Can anyone tell me what I should change? Thanks in advance. Get your own FREE, personal Netscape WebMail account today at http://webmail.netscape.com. Name: printcap printcapType: unspecified type (application/octet-stream) Encoding: base64 -- Terry Hancock [EMAIL PROTECTED]
ALSA Mixer Trouble
Hi, I just installed the ALSA driver on my system, which appeared to go perfectly (after a lot of sweat, anyway :) ). However, when I run amixer to unmute the card, according to the INSTALL instructions and the online FAQ, I get: peregrin:/proc/asound# amixer set Master on Can't access mixer 1/0 Failed to open mixer device Clearly I'm missing a step somewhere. Also, possibly of use, here is /etc/modutils/alsa_sound which I created after the install. Running update-modules sticks this text into the /etc/conf.modules file (I checked to make sure): # Modules setup for ALSA Sound driver # Set up by Terry Hancock, 2000-3/26 # based on ALSA install instructions #ISA PnP support options isapnp isapnp_reserve_irq=9,10,11,12,13 #ALSA native device support (Avance Logic ALS120 -- Avance Logic ALS100 driver) alias char-major-116 snd options snd snd_major=116 snd_cards_limit=1 alias snd-card-0 snd-card-als100 options snd-card-als100 snd_index=1 snd_id=ALS120 #OSS/Free setup alias char-major-14 soundcore alias sound-slot-0 snd-card-0 alias sound-services-0-0 snd-mixer-oss alias sound-services-0-1 snd-seq-oss alias sound-services-0-3 snd-pcm1-oss alias sound-services-0-12 snd-pcm1-oss Also maybe useful: peregrin:/etc/modutils# lsmod Module Size Used by ppp_deflate38660 0 (autoclean) bsd_comp3600 0 (autoclean) snd-card-als100 4020 0 snd-mpu401-uart 1472 0 [snd-card-als100] isapnp 22949 0 [snd-card-als100] snd-midi 12836 0 [snd-card-als100 snd-mpu401-uart] snd-seq-device 2896 1 [snd-card-als100] snd-sb16-dsp 15408 0 [snd-card-als100] snd-pcm1 17256 0 [snd-sb16-dsp] snd-mixer 29308 0 [snd-card-als100 snd-sb16-dsp] snd-pcm 8632 0 [snd-card-als100 snd-sb16-dsp snd-pcm1] snd-opl31760 0 [snd-card-als100] snd-hwdep 2412 0 [snd-card-als100 snd-opl3] snd-timer 8004 0 [snd-pcm1 snd-opl3] snd34080 1 [snd-card-als100 snd-mpu401-uart snd-midi sn seq-device snd-sb16-dsp snd-pcm1 snd-mixer snd-pcm snd-opl3 snd-hwdep snd-time unix9884 47 (autoclean) soundcore 2372 2 [snd] serial 19404 2 parport_pc 6964 1 (autoclean) lp 4956 0 (unused) parport 7060 1 [parport_pc lp] ppp20460 2 [ppp_deflate bsd_comp] slhc4232 1 [ppp] ne2k-pci3840 1 83905828 0 [ne2k-pci] rarp2864 0 (unused) umsdos 24192 0 (unused) ufs52736 0 (unused) sysv 23672 0 (unused) smbfs 25016 0 (unused) vfat8816 0 (unused) nfs28120 0 (unused) lockd 30888 0 [nfs] sunrpc 52132 0 [nfs lockd] autofs 8932 0 (unused) Anyway, there's no man page for amixer, and documentation for the alsautils package seems to be extremely scarce, so I'm hoping somebody knows what's going on, as I'm starting to get stumped. Thanks. -- BTW, it would be nice to CC this to me direct, I know some people on this list feel that it is inappropriate to post questions without taking the list, but it's pretty much like drinking from a firehose. Has anybody considered maybe breaking this list up by sub-topics or something? (e.g. debian-user-install, debian-user-drivers, etc.) I've joined this list about three times and finally given up after less than a week each time, just because it really chokes up my mailbox. (Not that the content is bad, mind you, just copious). -- Terry Hancock [EMAIL PROTECTED]
Re: debuser- Debian on Olivetti PCS 42P
Johan Andersson wrote: Hello, I'm trying to install debian on an old Olivetti computer. It has no network access but a quad speed cd-rom. But this this cd (Panasonic CR-562) uses a strange controllercard and therefore I can't boot of the cd. I have tried floppies but the computer only halts with the message 'Boot failed'. Since floppies seems to be the only alternative for me to get the base system on the computer so I can compile and install the correct driver for the cd I was just wondering if anybody know how to solve this. The computer only has a i486 50MHz, 200Mb harddrive, 16Mb RAM and windows 95 right now. Probably, the floppies are your best bet, still. The Linux floppy format is not as robust as the DOS system though, I don't exactly understand why, but it has something to do with tolerating bad floppy media (I think the Linux system assumes perfect media -- no bad sectors). Anyway, you have to pick really good disks: probably buy new ones, and not bulk. Another thing you can do is take your existing disks, and run chkdsk (from DOS) on them -- only select disks that have no bad sectors. Then use rawrite and follow the instructions for creating your rescue disk and base-system install disks. Of course, if the disk drive is actually bad, there's nothing you can do but replace it. Test this by trying to boot your Debian Rescue floppy on another computer (assuming you can get access to one). It won't overwrite your system so long as you stop it after the startup message comes up. I'm also assuming you don't have a problem with the BIOS letting you boot off of the floppy. On my system, the BIOS menu lets you determine which drives are tried for booting. If you were booting off the hard drive, I'm assuming your Windows would come up, or it would be otherwise obvious that it wasn't trying to boot off of the floppy. You might want to check the Linux hardware HOWTOs to make sure that drivers exist for your CD drive since you say it's unusual (and a proprietary interface). Actually being able to boot from CD is also a function of your BIOS as well as the drive, of course. Most 486 BIOS's can't do it. Nicely, Debian has a lot of CD interfaces supported in the base system, so you might get lucky that way. Anyway, good luck. Best regards Johan -- Terry Hancock [EMAIL PROTECTED]
Bus error crashes, different programs
Hi, This is one that's got me stumped (I am still fairly new at this, however). While using Netscape 4.51 (its the X11R5 version, if that makes any difference) under Debian Slink I tried to point the browser at one of the *.snm files in a new mail directory (I get the impression you're not supposed to do this, but how else do you learn?), and got some really disturbing behavior: every open xterm window started spouting error messages from SYSLOGD (accompanied by terminal bells, but I suppose that just adds drama) -- and I mean a lot of them, they looked like HTML code, but I'm not positive as they were going by too fast and I don't seem to have a permanent log of them anywhere. Since then I've had erratic program crashes with (so far) Netscape and XEmacs 20, both programs that use a lot of memory, but I don't know of anything else they have in common. The SCARY thing is, this is also happening to OTHER USERS (well, there's only one, but still). The crash behavior varies: sometimes I just lose the window (if run from an xterm, I get a Bus Error message on the terminal afterwards, at least with Netscape). Other times, the X-server apparently goes down dumping me out onto a text screen (I can recover by restarting xdm). Or it may just throw me back to the xlogin prompt. That says to me that whatever happened has damaged something in one of the system areas (i.e. not just files owned by me). I've already deleted Netscape's HTML caches, and eliminated cookies from websites I was looking at right before the incident. I've also used ... find * -mtime ... to identify files altered after the incident (this found that the files I mentioned were altered, also several /var/log/... files, but nothing that seems interesting). Now, if this were a Windows machine, I'd say I'd run out of memory, or swap space on the hard drive. However, none of the filesystems is more than 61% full -- (100s of megabytes are still available on /, /var, /tmp, /usr, and /home). Now, I don't actually know how to thoroughly check the swap partition, but top shows it as essentially unused. One thing, I've considered is that the swap mechanism might be broken in some way. The problem does tend to occur when system memory is mostly used up and no swap space has been. Examining syslog.conf, I find that messages that match *.emerg are sent to all users logged in (which appears to be the behavior I experienced), but they are not logged to disk, so I don't have a copy of what the errors were. I have not tried recreating the circumstances that (may have) caused the problem (I hesitate to do this, lest it do more damage). Its also possible that this was just a coincidence, and that I just happened to have some system resource problem at that time. Maybe I'm missing some basic system administration procedure. Oh yeah, I did of course reboot a couple of times (this apparently had no effect). I forced fsck on one of them, apparently it didn't find anything, though. Any ideas, references to documentation I ought to read, things I ought try, etc would be really handy. I would prefer to avoid the Windows solution (i.e. re-install EVERYTHING :) ). Sorry about the length, but I wasn't sure what details would be useful. Thanks a lot! -- Terry Hancock [EMAIL PROTECTED]
KDE/Qt Status in Debian dists
Hi, I've installed the Hamm (=2.0) distribution of Debian Linux from CD, and I intended from the start to install KDE as quickly as I could, but Debian has since gone to Slink (=2.1). On the Hamm CDs, the Qt library required by KDE is not included for licensing reasons. However, the Troll Tech website claims the newer versions of Qt are free software, apparently resolving previous licensing conflicts. (?) The Qt library debian package _is_ on the FTP site, but unfortunately, when I try to update the package from there using dselect, I get a cascade of dependencies -- basically dselect wants to upgrade my whole system to the Slink distribution. This would be okay, except for the downtime. I'd prefer to get the Slink CDs, but will the Qt library be on them? I have little time to work with this computer, and I fear that if I get Slink, Potato will be out before I get the Qt package, and I'll be back where I started. :) Also, can you get dselect to _downgrade_ your packages? I've now got some kind of mixture of Hamm and Slink packages on this system, and I'm not sure how stable that is. I wish dselect gave you a little more control, but I haven't quite figured out how to do everything with dpkg. Would appreciate any help at all, thanks! Terry Hancock [EMAIL PROTECTED]
ATAPI CDROM -- Aargh! Newbie problems
don't think I've had memory problems with it -- I couldn't quite pop for the 168-pin DIMM module, maybe next month :). Would another Linux distribution work? (Maybe the NEC-260 would be recognized correctly, for example?). I am pretty new at Linux, though I'm a moderately experienced unix user (actually I'm a system admin for a Solaris system at work, but this is only because we can't afford a real one :) ). I do plan to put this system to pretty heavy work once I can get it to work at all: for office applications, CAD, and software development. I need to be there in about a year or so, which is why I need to start learning now! If it comes to buying another CDROM drive, are there any recommendations? Having two different types not work does not instill me with confidence about others on the market! Anyway, I appreciate any help at all, thanks! Terry Hancock [EMAIL PROTECTED]
ATAPI CDROM problem (NEC-260)
Hi everyone, I'm really green, so pardon me if this is in a FAQ, but I did search the archives for this problem to no avail. I have an NEC IDE/ATAPI CD-ROM. It says CDR-250 on the case, and the BIOS reports NEC ATAPI CDROM: 260 (or something similar to that). I'm having trouble getting it to be recognized by Debian Linux (Hamm release, version 2.0.2 on CD from Linux Systems Labs). Using a DOS boot floppy with the NEC driver and MSCDEX, I am able to read the CD, which convinces me that it is not a drive or cable problem. It is jumpered as a slave on the primary IDE controller (on the motherboard). Examining the dmesg output I found the following, which would seem to be the problem: ide: i82371 PIIX (Triton) on PCI bus 0 function 57 ide0: BM-DMA at 0xf000-0xf007 ide1: BM-DMA at 0xf008-0xf00f hda: SAMSUNG VG38404A (8.40GB), 8056MB w/490kB Cache, CHS=1027/255/63 hdb: NEC CD-ROM DRIVE:260, ATAPI cdrom or floppy?, assuming FLOPPY drive ide-floppy: Packet size is not 12 bytes long ide-floppy: hdb: not supported by this version of ide-floppy ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 Floppy drive(s): fd0 is 1.44M FDC 0 is a post-1991 82077 [complete dmesg output after message] Apparently the OS is making an incorrect assumption about the identity of the device and then failing because of it. (Ever heard of an IDE floppy? I've certainly never seen one). Anyway, after booting, there is no /dev/hdb device to mount. I don't know what I can do about this. Surely there's some way to nudge Debian in the right direction? (I have installed the basic part of the OS from floppy disks, so it is possible to tinker with the configuration files). - As an alternative, I have a Creative Labs Infra-1800 CD-ROM which is a 12X with a gimmicky infra-red remote. I figured it would be better to leave that on the Windows machine. It's also an ATAPI IDE CD-ROM. I would expect it to have the same problem, but I'll probably try it if all else fails. -- Many thanks! Terry Hancock ** Complete DMESG Listing Follows ** Console: 16 point font, 400 scans Console: colour VGA+ 80x25, 1 virtual console (max 63) pcibios_init : BIOS32 Service Directory structure at 0x000fabc0 pcibios_init : BIOS32 Service Directory entry at 0xfb040 pcibios_init : PCI BIOS revision 2.10 entry at 0xfb070 Probing PCI hardware. Calibrating delay loop.. ok - 80.08 BogoMIPS Memory: 6020k/8192k available (1124k kernel code, 384k reserved, 664k data) This processor honours the WP bit even when in supervisor mode. Good. Swansea University Computer Society NET3.035 for Linux 2.0 NET3: Unix domain sockets 0.13 for Linux NET3.035. Swansea University Computer Society TCP/IP for NET3.034 IP Protocols: IGMP, ICMP, UDP, TCP VFS: Diskquotas version dquot_5.6.0 initialized Checking 386/387 coupling... Ok, fpu using exception 16 error reporting. Checking 'hlt' instruction... Ok. Linux version 2.0.34 ([EMAIL PROTECTED]) (gcc version 2.7.2.3) #2 Thu Jul 9 10:57:48 EST 1998 Starting kswapd v 1.4.2.2 Real Time Clock Driver v1.09 tpqic02: Runtime config, $Revision: 0.4.1.5 $, $Date: 1994/10/29 02:46:13 $ tpqic02: DMA buffers: 20 blocks, at address 0x277800 (0x2777d8) Ramdisk driver initialized : 16 ramdisks of 4096K size loop: registered device at major 7 ide: i82371 PIIX (Triton) on PCI bus 0 function 57 ide0: BM-DMA at 0xf000-0xf007 ide1: BM-DMA at 0xf008-0xf00f hda: SAMSUNG VG38404A (8.40GB), 8056MB w/490kB Cache, CHS=1027/255/63 hdb: NEC CD-ROM DRIVE:260, ATAPI cdrom or floppy?, assuming FLOPPY drive ide-floppy: Packet size is not 12 bytes long ide-floppy: hdb: not supported by this version of ide-floppy ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 Floppy drive(s): fd0 is 1.44M FDC 0 is a post-1991 82077 md driver 0.35 MAX_MD_DEV=4, MAX_REAL=8 Failed initialization of WD-7000 SCSI card! ppa: Version 1.42 ppa: Probing port 03bc ppa: Probing port 0378 ppa: SPP port present ppa: PS/2 bidirectional port present ppa: Probing port 0278 scsi : 0 hosts. scsi : detected total. Partition check: hda: hda1 hda2 hda5 hda6 hda7 hda8 hda9 hda10 VFS: Mounted root (ext2 filesystem) readonly. Adding Swap: 130748k swap-space (priority -1) Module inserted $Id: cdrom.c,v 0.8 1996/08/10 10:52:11 david Exp $ Serial driver version 4.13 with no serial options enabled tty00 at 0x03f8 (irq = 4) is a 16550A tty01 at 0x02f8 (irq = 3) is a 16550A tty03 at 0x02e8 (irq = 3) is a 16550A lp1 at 0x0378, (polling) VFS: Disk change detected on device 02:00