Debian Books request (fwd)
-- Forwarded message -- Date: Thu, 6 Dec 2001 22:48:37 +0100 (CET) From: Martin Grande [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Hello If this is the right e-mail adress to send to, I dont't know.. there were so many to choose from :) but I'll give it a try. I just have two simple questions.. here it goes: How many Debian books is there ? I'm wondering if you guys can suggest one for me...(I'm not a _fully_ newbe, i know a bit :) I'm not looking for any web help or something.. just the general, if you know what I mean. Hope you can help me out. Regards Martin
Re: apt-get wierdness: when 6 is equal to 9
On Mon, 24 Sep 2001, Joey Hess wrote: maintains more complex package list information than can be represented in dpkg's available file (especially if you have apt set to use multiple distributons at the same time, etc). Yeap, pretty much. I find this very unsatisfactory. Yes, generating an available file whenever apt-get update is run may not be technically perfect, and the Well, here is my rant.. * tasksel * grep-available * dpkg -p That should _just work_ in a high-quality Debian system. You notice that all but dpkg -p were written after APT? They were written with full knowledge that that available file would be inaccurate on 99% of their users systems. They could have been written to use apt-cache on systems that are using APT and this would be a non-problem. In fact, as long as APT is involved the only way to have things just work is to call 'apt-cache dumpavail' and parse that. That function accounts for all variables and presents the same selections that apt-get would see. The reason these tools don't do this is some IMHO mis-guided aversion to assuming APT is being used. After all, there are 2 people out there who don't use it.. Similarly checking if the APT is selected in dselect would be inelegent :P The only way I can see to compromise around this is if someone changes dpkg. Yes, the limited interface to it's methods is unsuitable and must be extended. I suggest that someone cook up a patch to dpkg that changes it to effectively popen the script /usr/lib/dpkg/methods/*/get-available and parses that instead of the available file. Dpkg would also need an way to dump the available file to stdout.. Jason
Re: apt-utils
On Sun, 27 May 2001, Joey Hess wrote: Jason, if it's doing that, I think that's a dumb heuistic. As you can see, there are valid reasons for ignoring the input and not failing. It is doing that and it has always done that.. It only fails sometimes because the pipe fills up. Probably will take that out.. Jason
Re: Proxy apt cache?;
On Fri, 27 Apr 2001, Alan E. Davis wrote: How can I set up the new machine now to recognize these packages and install, either from this machine or the connected machine? Can I just copy over the /etc/apt/sources.list onto this machine without being connected? Which of the other files in /var/cache need to be copied over? You can copy over /etc/apt/sources.list and /var/state/apt/lists/* (or /var/lib if you are using 0.5 APT's) as well as /var/cache and things will work without needing connectivity. Jason
Re: (OT) - Static electricity grounding device?
On Tue, 27 Mar 2001, Roberto Diaz wrote: I really doubt very much that to connect your electric installation to your pipes installation is even legal.. ask an insurance company. It's common practice around here.. The pipe going out of the house is an excellent ground. However, if you do it wrong, or the pipe is the wrong sort you can do all sorts of nasty things to yourself. I've never seen this sort of connection done using solder, I doubt it wold work. There are special press connectors that are generally used. Be sure to use green wire : That said. Your house should already be well earthed. The 3rd prong on a plug (the roundish one) is a path the grounding block in a fuse box which then goes to either a metal pipe (water) or a metal stake. A well built computer should connect all the metal case parts directly to the ground prong on the outlet, so to dissipate static it is usually sufficient to touch metal on a computer. Jason
RE: Linux Network Security: POP
On Mon, 19 Mar 2001, Joris Lambrecht wrote: I asssumed cable modems were encrypting there communications with some simple built-in algorithm It is my understanding that modern DOCSIS modems use encryption between the cable modem and the cable head end. The motorola cybersufr brand has been doing this forever as well. This prevents someone from using some sort of cable analyser to sniff datagrams after they hit the wire.. I wouldn't count on the encryption being actually super secure, but it is unlikely that someone is going to be sniffing packets by examining the signals on the coax. as you should be, cable modems generally are equivilent to large unswitched lans, which means any bozo with a cable modem can set thier machine to primisquous mode and see every packet sent by any cable modem user. (at least for that segment) This is certianly untrue for modern cable stuff. In general, the bandwidth on the actual coax is far greater than 10mbit ethernet (coming out of the modem), even if the modem wanted to it couldn't spew all packets onto the local lan. Jason
Re: FW: FW: D-Link DFE-530TX+ via-rhine.o module
On Mon, 19 Mar 2001, David Carlile wrote: I have been looking through the archive for references to the card and driver but all I come up with are posts about how easy it is to get this card running. People actually suggest this card to newbies because it is so Dlink changed the configuration of the realtek they were using during later production runs. The particular card you have does not report itself as a realtek, so the driver does not find it. The only fix is to modify the kernel driver. The simplest route is to use the donald becker drivers (http://www.scyld.com/network/) for the 8139 which has an entry, or you can modify the PCI probe of the linux drivers. I would expect that the 2.4.2 kernels have this patch applied, but 2.4.0 did not.. It would be very nice if the next drop of the potato boot floopies included an entry in the realtek driver code for this chip. The update to the 2.4.0 8139too driver looks like this: static struct pci_device_id rtl8139_pci_tbl[] __devinitdata = { {0x10ec, 0x8139, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139 }, {0x10ec, 0x8138, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139_CB }, {0x1113, 0x1211, PCI_ANY_ID, PCI_ANY_ID, 0, 0, SMC1211TX }, /* {0x1113, 0x1211, PCI_ANY_ID, PCI_ANY_ID, 0, 0, MPX5030 },*/ {0x1500, 0x1360, PCI_ANY_ID, PCI_ANY_ID, 0, 0, DELTA8139 }, {0x4033, 0x1360, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ADDTRON8139 }, {0x1186, 0x1300, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139 }, // dlink {0,} }; A fix for the 2.2 kernel should be similar in nature, but the table looks different IIRC. Jason
Re: apt-get problems
On Mon, 12 Mar 2001, John Galt wrote: deb ftp://ftp.debian.org/debian unstable main contrib non-free ^ You're missing the terminal slash: the correct line (the one I use) is: Which is added automatically for you by apt.. Jason
Re: apt-get question
On Sun, 11 Mar 2001, Keith Johnson wrote: On Sun, Mar 11, 2001 at 10:16:35AM -0800, Bill Wohler wrote: Now, we all use `apt-get update; apt-get upgrade' on a regular basis. But you're supposed to use `apt-get dist-upgrade' when moving between distributions. What happens if stable changes without your knowledge and you run `apt-get upgrade'? Okay, now I am a little confused. I have been using `apt-get dist-upgrade' on a regular basis. (It seems to update my system fine). Am I doing something wrong here? No, the distinction is that dist-upgrade removes things and installs new things, while upgrade doesn't ever. Which is why you must use dist-upgrade when moving between distributions. Jason
Re: apt get install of red carpet
On Wed, 7 Mar 2001, Brian Murphy wrote: no package name was given. I tried apt-get install redcarpet but it didn't work. Did anyone else do this? *lol* Try 'apt-cache search red.*carpet', it might find it for you. Jason
Re: D-Link DFE-530TX Probs W. 2.4.2
On Mon, 5 Mar 2001, Marc Moody wrote: I believe that the 530TX does use the via Rhine Chip, but the 530TX+ uses the realtek 8139. Strange that the plus would have such a significant change, but I have the 530TX+, and I though I read this when I was installing mine. Correct. Further, some 530TX+'s have non RTL8139 PCI IDs, like mine: 00:0d.0 Ethernet controller: D-Link System Inc: Unknown device 1300 (rev 10) You need to either use Becker's latest driver, or add the ID's to the 8139too driver, like this: {0x1186, 0x1300, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139 }, Jason
Re: cannot apt-get libc6-dev
On Sat, 24 Feb 2001, Jason N. Price wrote: as there are a lot of dependencies on it. Without it installed, I can't install a lot of other things. I have tried getting it from the cd as well as via HTTP, but the error is the same either way. The error is: Sounds like you might have flaky RAM or kernel or something. It means that the file is corrupt. If the file is not actually corrupt, the it got corrupt as it was read in, flaky RAM. Jason
Re: apt-get strangeness
On Fri, 23 Feb 2001, Frederik Vanrenterghem wrote: I've noticed a strange problem with apt-get: it downloads packages twice. Any ideas what could be causing this? I'm using Woody, and I've noticed this error for some weeks now. It seems to have been reported in bug report #79277, but it's still causing a lot of excess traffic up until today... Send me an strace -o /tmp/apt -f -ff of APT doing this, off list, MIME'd and tar.gz'd with a copy of /etc/apt/sources.list Jason
Re:
On Tue, 13 Feb 2001, Susumu Takuwa wrote: JG All we need is the *full headers* from the message. Our server uses VERP, JG the address he is subscribed with is sent to him in the envelope sender. like this, This address is no longer subcribed. Jason
Re: apt-get: 93 Protocol error
On Sat, 10 Feb 2001, Eric G . Miller wrote: On Sat, Feb 10, 2001 at 08:32:57PM -0600, Mike McNally wrote: The full error is: Could not create a socket - socket (93 Protocol not supported) No idea what protocol 93 is (it's not listed in /etc/protocols). FTP should use the tcp protocol (6). '93' is the errno for 'Protocol not supported'. IIRC this is due to a bug in libc6, and I thought it was fixed months ago. If you grab APT 4 from http://people.debian.org/~jgg/apt I think I put a work around in it.. Jason
Re:
On Mon, 12 Feb 2001, Susumu Takuwa wrote: UB That will sort out their problems, as well UB as ours. I am still receiving mail from UB them. OK, I will try explaining to them again. All we need is the *full headers* from the message. Our server uses VERP, the address he is subscribed with is sent to him in the envelope sender. For instance, the message I just got has this header: Return-path: [EMAIL PROTECTED] Notice that my email address is encoded in the return path. If you send me that information I will personally unsubscribe him. (in fact, he could do it himself, just mail all the messages he is getting now to that address rather than to sender) Jason
Re: apt: http vs. ftp?
On Thu, 7 Dec 2000, David Teague wrote: I am not a networking authority, so I asked a colleague (Mark Holliday) who is. He says http is optimized for relatively small files, mainly web pages, which are not terribly large, (what? 2 or 3 K?) whereas ftp was designed to be optimal files that may be very large. Ask him how you optimize a protocol for file size and when he fails to explain that then you know the truth : Jason
Re: apt: http vs. ftp?
On Thu, 7 Dec 2000, David Teague wrote: I never pretended to know anything, but I find your response amusing, and your discussion and that of others enlightening. I suspected that ftp might not be faster, but did not know why. The true answer is that there is no way one is faster than the other for infinite sized files. They both use TCP, the both use the same kernel and they both use the same TCP options. Any percived difference is mearly statistical error. In practice this is a bit of a lie, HTTP severs tend to be more optimized and use thing like sendfile and pay attention to latency. That kernel http server (TUX?) will likely lay to waste any FTP server when you start talking about real servers with real load. Now you can start talking about non-infinite file sizes and lots of them. At this point FTP starts spending less time transfering data and more time waiting around to decide what to do. In contrast, HTTP -never- stops, the socket buffers remain full through out the etire session and it never sends partial packets. This is because APTs HTTP implemenation is fully pipelined and optimized for this. FTP cannot be. You can see this by running apt-get update twice with http and ftp. The second run when it prints out 'Hit:' is almost instant for HTTP, while it is agonizing for FTP. A typical FTP request may look like this: c: SIZE /foo s: 123 c: MTDM /foo s: 12313 c: get /foo S: Port 123, etc c: send first TCP conn packet s: Reply c: finish s: data Sure is a damn lot of round trips Each one wight take about 25ms to complete on a fost link, so it takes about 100ms to even begin transfering the data. HTTP/1.1 doesn't have ANY round trips once you start going. APT will send 5 requests in a single large packet, the remote will queue them up and start besting out data, each finished request results in another going back to the server. There is never a case when the two have to stop sending and syncronize. Finally you can consider HTTP proxies servers, and HTTP QoS munging. Sadly they exist : HTTP proxies are often poorly implemented, buggy and slower than a router - you rarely get good rates from them. Some ISPs will actually use QoS on HTTP streams to encourage their routers to favor mail and FTP traffic, this of course has a negative effect too.. So, in the real world some people may measure significant deviation, this is always due to their ISP messing with their traffic and not any particular issue with the protocols. Jason
Re: apt: http vs. ftp?
On 4 Dec 2000, Willy Lee wrote: Just curious, is there any particular reason to favor http vs. ftp, or http is faster if you are not proxied. Jason
Re: apt: http vs. ftp?
On Tue, 5 Dec 2000, Ray Percival wrote: That would be backwards ftp is faster but sometimes it is easier to get http through a proxy and with some proxies it would be possible that http might be faster. Er, no it isn't. http is faster and better in all cases where there is not a proxy involved. Jason
Re: apt: http vs. ftp?
On Tue, 5 Dec 2000, Erik Steffl wrote: Er, no it isn't. http is faster and better in all cases where there is not a proxy involved. why would http be faster? how much faster you mean? and what makes it better? AFAIK they are about equally good/fast for purpose of file transfer... http almost no per file overhead, ftp has a substantial amount. You only need to compare apt-get update using both to realize why http is better overall. Jason
Re: apt: http vs. ftp?
On Wed, 6 Dec 2000, Chris Kenrick wrote: itself is less for http. This makes me curious .. why would a hypertext transfer protocol have less overhead on file transfers for one designed for transferring files? Because the design goals of FTP were never to have a low cost file connection? If makes perfect sense really. How many files does your average web browser fetch? Jason
Re: apt-get update ERROR!
On Mon, 4 Dec 2000, Jim Frey wrote: On running apt-get update, I get the following error listing: Upgrade to a nemer version.. Jason
Re: Terminals/dim screens = use bold everywhere but dselect is...dim
On Sun, 3 Dec 2000, Oliver Elphick wrote: S Taylor wrote: Do color monitors die after four years? The boot messages are barely visible; seems they've gotten worse over two months since I installed Where can I find HOWTOs to configure this 'terminal', a magitronic These do :P Amazing you even got 4 years. Jason
Re: Purging a package with apt (was: Re: Q: Why 24 depth/not 32?)
On Thu, 30 Nov 2000, Ignasi Tura wrote: I've grepped /usr/doc/apt for 'purge', man apt, apt --help and I haven't found any reference towards 'purge'! auric{jgg}~/apt2/build/bin#man apt-get | grep -B 1 -i purge --purge Use purge instead of remove for anything that would Jason
Re: Helix-Gnome not installable
On Sun, 12 Nov 2000, Joel Dinel wrote: I can't seem to install Helix Gnome on my woody box. Thats an interesting apt quirk, just list all the packages at once: auric{jgg}~/apt2/build/bin#./apt-get install task-helix-gnome task-helix-core sawfish-gnome rep-gtk-gnome libgdk-pixbuf-gnome2 Reading Package Lists... Done Building Dependency Tree... Done Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. The following information may help to resolve the situation: Sorry, but the following packages have unmet dependencies: libgdk-pixbuf-gnome2: Depends: libgdk-pixbuf2 (= 0.9.0-0.2) but 0.9.0-helix5 is to be installed rep-gtk-gnome: Depends: rep-gtk (= 0.14-1) but 0.14-helix2 is to be installed E: Sorry, broken packages Tada. This appears to be because the helix folks are mixing and matching libraries from debian and them. they were supposed to be using the debian versions :| For instance if I use the new APT I can resolve this.. auric{jgg}~/apt2/build/bin#./apt-get install task-helix-gnome libgdk-pixbuf2=o=debian rep-gtk=o=debian Reading Package Lists... Done Building Dependency Tree... Done Selected version 0.9.0-0.2 (Debian:unstable) for libgdk-pixbuf2 Selected version 0.14-1 (Debian:unstable) for rep-gtk [..] 22 packages upgraded, 100 newly installed, 1 to remove and 238 not upgraded. Need to get 38.1MB/61.9MB of archives. After unpacking 123MB will be used. By picking the Debian versions of the problem libraries.. Jason
Re: apt-get strangeness
On Mon, 13 Nov 2000, Michael P. Soulier wrote: So, apt-get shouldn't be used when something's broken? I used apt-get in the first place, so I take it that means that apt-get didn't catch the problem. I know the packaging tools are much better than my rpm experiences, but I'd love to see these kinds of problems go away. Er no, this is completely silly. There are no problems that APT does not catch. Your problem in that APT has interesting way to complain about missing dependencies. You need to to include more packages on the some install line and you will be able to figure out what is wrong. Jason
Re: Compiled debs are upgraded
On 7 Nov 2000, Martin Bialasinski wrote: Joachim Same here with gmc -- after an 'apt-get --compile source gmc' Joachim another 'apt-get upgrade' will replace my newly compiled Joachim package with the same version. Actually, this is by design and the reason the equivs package was rewritten. Unfortunately, I didn't take notes why apt behaves like this. Therefore, I crosspost this to the gurus. They can say for sure. If it didn't do this then their would be way to make it do this : The only other option is to prefer local packages (which can be done by using a hold) and that makes this usefull behavior impossible. Jason
Re: apt download security?
On Tue, 7 Nov 2000, Bruce Richardson wrote: Unfortunately, while source packages can be checked quite easily, they are not always verifiable. There is no simple mechanism for verifying debs *at all*. Nor even Packages.gz - and the integrity of Packages.gz isn't actually a guarantee of the integrity of any of the packages. Er we can provide a 'ssl-like' assurance for the Packages.gz which will rule out any bad mirror from forging .debs It is much harder to prevent attacks from people with root on our primary box however! Jason
Re: Compiled debs are upgraded
On Mon, 6 Nov 2000, Bruce Richardson wrote: If I compile a deb-src package and install it, apt-get upgrade will over-write it with the precompiled version unless I mark my hand-rolled package as hold. What am I doing wrong? Nothing. This is how it is ment to work. You can change the version number in the debian/changelog file if you want to do something else Jason
Re: dpkg 1.7.0 warning
On Mon, 6 Nov 2000, Gordon Sadler wrote: I followed this advice and upgraded to 1.7.1 just a few minutes ago, however I am still receiving this error below. Looks like you removed debconf.. Put it back, by hand! Jason
Re: apt woody.
On Fri, 3 Nov 2000, Peter Palfrader wrote: Maybe apt should simply retry after all the other stuff got downloaded? It can try forever, glibc always returns that :| Jason
Re: World's largest mailing list?
On 2 Nov 2000, John Conover wrote: Its not exactly a Debian/Linux question, but does anyone know how many email addresses are on the world's largest mailing list, and the OS/HW it runs on? Average messages per day? Well, I can tell you that the debian.org list server has 8 subscribers to all of it's lists. The largest list is debian-announce with 14 thousand. The total traffic we do is on the order of 50 remote deliveries per day, with peak traffic rates of 30-40 remote deliveries per second. Till recently it was running on an older P166 with IDE disks, now it has a PII 400. It runs Debian GNU/Linux and qmail with smartlist (bleck). Jason
Re: dpkg won't install (Internal Error)
On Fri, 27 Oct 2000, Will Day wrote: # apt-get -s install php4-mysql mysql-server Inst libstdc++2.10-glibc2.2 [base-config ] Inst libstdc++2.10-dev [base-config ] Inst g++ [base-config ] Inst libc6-dev [base-config ] Inst locales [base-config ] Inst libc6 [base-config ] E: Internal Error, Could not perform immediate configuration If this is actually an error in one of the packages, rather than dpkg itself, it sure would be nice if it printed the package name it's having problems with! I installed a potato a few days ago, upgraded it to woody shortly thereafter, and was very impressed. Today I installed another potato, and immediately tried to upgrade it to woody, and had a similar problem. As This is a odd interaction between libc6 and APT's sequencing code, there is a bug against libc6 to get the issue fixed so that potato to woody upgrades work again.. In the mean time there are a couple options.. Probably the simplest is to do 'apt-get -s install libc6' then take all the packages listed as 'Inst' and install them by hand using dpkg with a single command. Or you could try the experimental 0.4.1 APT from http://klecker.debian.org/~jgg/apt Or just wait for a new libc.. Jason
Re: apts: something wicked happend resolving 'http.us.debian.org/http '
On Mon, 23 Oct 2000, Russ Pitman wrote: Postfix broke with a stuck mail queue, apt-get broke again and I have just finished reinstalling potato on a separate drive to enable mail. I reckon I will stay down here for mail until I can find out how to repair woody 'cos there is a heap off apps there I don't have room for on this drive. Try installing your own copy of bind setup with no forwarders and set your resolv.conf to only use localhost. If that solves it then it would confirm my theory that this is caused by buggy DNS servers messing with glibc's new logic... Someone really needs to look into this in detail, but I don't know of anyone who can reproduce it at will. If you can do this (by editing your resolv.conf) then I think we can probably find it.. Jason
Re: Apt should be called inapt (rhymes with inept)
On Tue, 17 Oct 2000, Joe Emenaker wrote: It seems that apt-get decided to UNINSTALL: o netkit-inetd o ipchains o netbase and a couple of others. In other words, upon reboot, there was no network connectivity and no way to GET network connectivity without bringing in netbase and it's dependencies via floppy disk. So uh, why did you let it? Is there an option for apt-get to tell it to install extra stuff it needs but to *not* remove anything? In cases like this you need to either use dselect/capt/etc to interactively select packages or specify exactly what to do after the install command. This is only necessary if you have perviosly done something to mess up the dependencies in such a way that the automatic fixing algorithm chooses an undesirable solution. Jason
Re: Apt should be called inapt (rhymes with inept)
On Wed, 18 Oct 2000, Jeff Green wrote: Surely a quick trip to /var/cache/apt/archives and a run of dpkg -i with the right package names would have fixed this, if apt-get doesn't apt-get install with the right package names would have also fixed it, and told you when you finally got the right package names. Jason
Re: Apt should be called inapt (rhymes with inept)
On Tue, 17 Oct 2000, Joe Emenaker wrote: It seems so basic. When you say apt-get install ... the plan is clearly the addition of software to the system. Removal is patently not part of the plan, unless explicitly acknowledged by the user. If apt executed removals you were prompted and you did say yes to that prompt. If you weren't prompted I find it seriously doubtful that apt is the culprit here. Jason
Re: Still cannot acess Debian.org
On Mon, 16 Oct 2000, Russ Pitman wrote: I recall seeing messages attributing these to the AARNET mirror,but my belief was that the http sites were on a round robin so that a busy site would redirect the load. Is this not so?. This is a glibc bug aggrevated by your name server.. Jason
Re: apt woody.
On Tue, 10 Oct 2000, Russell Davies wrote: anyway, after hours of upgrading over a 28.8K modem the process has finished. However something now seems broken with apt-get I did notice a warning about name resolution and the new C library when installing the new packages. Perhaps a reboot might fix the following problem: I've heard this many times.. I have no idea what causes it.. it should not be happening, the getaddrinfo call should never return something that causes the message you see! If you find out what this is please let me know.. x Jason
Re: apt woody.
On Mon, 9 Oct 2000, Seth Cohn wrote: It would be useful, Jason, to understand what the message means: It means an error that is not understood by the code was generated by the lookup. An error that is not 'name not found' or 'service not found'. I've never seen this prior to the new libc6.. Lets try to capture the error code form getaddrinfo and see whats up. http://auric.debian.org/~jgg/apt_0.3.19.1_i386.deb Has a version which ought to print out a number in brackets that can be cross referenced to /usr/include/netdb.h (the EAI_ defines) to determine what code glibc is spitting out, we can then investigate glibc's source to determine why it would spit that out, then figure out WTF is wrong. Jason
Re: Moving cdrom for apt-get
On Tue, 10 Oct 2000, Jonathan Gift wrote: I want to move the cdrom from /cdrom to /mnt/cdrom and even though so noted in fstab, apt-get refuses to see anything but /cdrom to grab apps. Any idea how to change its config? Stick Acquire::CDROM::Mount /mnt/cdrom/; In /etc/apt/apt.conf Jason
Re: Conflicts make APT useless
On Tue, 10 Oct 2000, Thomas Halahan wrote: I have upgraded glibc to 2.1.94-3. During this process I have had the same problems as many. i.e. * libdb.so.3 not found * ldconfig disappears But now my apt (dselsct and gnome-apt) shows many unsolved dependencies which become very confusing and difficult to solve. Most notable is the requirement of libdb2 by many packages. However libdb2 conflicts with my libc6, so there is nothing I can do. How do I persuade my apt programs to work again, as I am fed up of downloading individual packages that turn out to have interdependencies that are too difficult to solve with dpkg? Er generally you just do as it says, apt-get -f install and it will magically figure it out, unless things have become so tangled that a solution is not apparent [this is really rare though]. In that case you must add packages to install by hand after the 'install' until things are better. Sounds to me like you just need to install libc6 and libdb2. Jason
Re: apt stat problem continues
On Tue, 10 Oct 2000, William Jensen wrote: As reported yesterday apt-get update produces errors at the end. Can anyone tell me and the other people experiencing this what the problem is and how we could fix it, or who we should inform if this isn't a user fixable problem. Apt-cache search produces the same problem. For reference, I'm on a standard potato install with vim and iptables from woody. I have never seen or heard of this prior to your posting.. Do the files it mentions exist? It may be that you are behind some kind of horrible transparent http proxy which is doing something strange, try using ftp instead of http, if that works we can perhaps debug the http method and see what is up there. Jason
Re: apt on non-debian system
On Fri, 6 Oct 2000, Piotr Krukowiecki wrote: I'd like to install apt on RH. I don't want to install deb's on RH. I want only use it as it reads in /usr/share/doc/apt/offline.txt (apt-get update apt-get -d dist-upgrade with state file from other, debian system) So, is it possibile ? Sure, just grab the source and do: make startup make cd build/bin export LD_LIBRARY_PATH=`pwd` ./apt-get You'll need to use the configuration options that offline.txt indicates so that you can use a custom sources.list that is not in /etc/apt/. Jason
Re: APT::Force-LoopBreak ??
On Sat, 30 Sep 2000, Thomas J. Hamman wrote: Okay, so I can't dist-upgrade because of a problem that I shouldn't work around because it's A GRAVE BUG and I don't -really- know what I'm doing -- so, what am I supposed to do? In this case you may activate the option, however, I recommend doing 'install libc6' then proceeding with the dist-upgrade, try not to crash your computer during that time because login will stop working for a little while.. Jason
Re: Woody broke my system
On Thu, 28 Sep 2000, Joey Tsai wrote: Need to get 0B/2426kB of archives. After unpacking 2925kB will be freed. Do you want to continue? [Y/n] perl: /lib/libc.so.6: version `GLIBC_2.2' not found (required by /usr/lib/libdb.so.3) E: Sub-process /usr/sbin/dpkg-preconfigure --apt returned an error code (1) E: Failure running script /usr/sbin/dpkg-preconfigure --apt Uh... Yes.. you must comment out the line in /etc/apt/apt.conf that mentions debconf until you can get perl working again. [corban][02:50pm][/var/cache/apt/archives] # dpkg -i libc6_2.1.94-1_i386.deb dpkg: `ldconfig' not found on PATH. dpkg: 1 expected program(s) not found on PATH. NB: root's PATH should usually contain /usr/local/sbin, /usr/sbin and /sbin. You must use the --force-bad-path option to dpkg. You should be very careful to get ldconfig back and then restore libc6 then upgrade the packages that need to be upgraded to get around the db breakage. Jason
RE: apt-get size mismatch with woody
On Tue, 26 Sep 2000, Jason Holland wrote: Is there anyway to bypass the size check? I didn't see anything in the man page. Just curious.. No, technically the file is corrupt and should not be used without that final byte. Jason
Re: What's with at.debian.org?
On Sun, 24 Sep 2000, Robert Waldner wrote: Anyone know WTH is up? No idea, therefore I´ve cc´ed the SOA ([EMAIL PROTECTED]). Judging from the ser# (292305) there was a change in the zonefile yesterday. As ftp.at.debian.org is still in the list at http://www.debian.org/distrib/ftplist I don´t think it vanished on purpose... Our primary name server had a disk crash, the backup isn't quite setup right for delegated sub domains, I'll fix it. Jason
Re: apt-get upgrade size mismatch question
On Wed, 6 Sep 2000, Jason Holland wrote: which is a real pain. anyone know why it fails to install because of a size mismatch?? a workaround for this would be nice. thanks!! This means the resulting .deb is not the correct length. It might be missing a byte or two and still be installable with dpkg, but it is not a valid .deb. Jason
Re: ipv6
On Thu, 31 Aug 2000, Russell Coker wrote: So I have an inet6 address on eth0 as soon as I do modprobe ipv6. What is the way to ping it? I don't think you can ping link addresses very easially. You should install the radvd package and give your site some real IPs, possibly following the 6to4 RFC or something. This will give your box and any other boxes on your lan ipv6 addresses. Jason
Re: apt-get error
Well apt-get upgrade did never finish OK. It started out more or less ok, but after a while it could not properly configure some packages (gimp-manual and libpaperg (here it never accepted any paper format like a4 and there was no list to choose from; therefore forcing a ctrl+c and thereby aborting everything)). After this incomplete upgrade nothing was You must complete this upgrade. Usually this is done by rerunning apt-get upgrade again and again. Any problems you encounter *must* be delt with somehow. Sometimes old systems can be difficult to upgrade because the installation scripts get confused.. Jason
Re: apt-get error
On Wed, 23 Aug 2000 [EMAIL PROTECTED] wrote: When I did apt-get --fix-broken --show-upgrade dist-upgrade, I got the info about what apt wanted to remove, install and update. Then it asked Hm. You have more packages installed than I have ever seen.. This is a bug - can you please send me a copy of /var/lib/dpkg/status [mime+gzip] when this is happening. For your immediate problem.. I recommend you attempt starting with 'upgrade' and then switch to dist-upgrade after that completes OK. Jason
Re: apt-get won't play
On Sun, 20 Aug 2000 [EMAIL PROTECTED] wrote: - apt-get update complains it can't resolve the above address, complains it can't run a stats on the earlier packages on the system, then suggests I run apt-get update! Sounds like your DNS is foobar, unless you made a spelling error. Jason
Re: apt-get update for frozen -- 404 not found
On Tue, 15 Aug 2000, Dave Thayer wrote: Potato is now stable, so you will have to change your sources.list accordingly. FWIW, I prefer to use the release name (i.e. 'potato) to avoid this kind of suprise. I had been tracking potato in its later Yeah, but now you know that potato has been released, independent of the news sites : Jason
Re: Complete local mirror.
On Wed, 16 Aug 2000, C. Falconer wrote: I've been toying with the idea of setting up a complete local mirror for all of my machines here. It wouldn't be public (at least not yet), but it would cut down on my network traffic, and it would cut down on the debian servers that are being pummelled right now. This is so not true. If you installed all your machines, without using any sort of shared local cache it would still produce less bandwidth than 1 week of a full local mirror. You realize it takes about 200meg/day to keep up with our archive? Shared APT, squid, etc are all more banddith efficient solutions. Jason
Re: upgrading helix-gnome stuff
On 25 Jul 2000, Brian May wrote: Hello, when upgrading to the latest helix-gnome debs from within helix-gnome, often weird things happen to my X-Server. Often, for instance my computer constantly beeps at me, until I push Alt+Backspace, and force Did it beep at you only when you press keys or click the mouse? I have seen X do that a few times.. it is very very odd. Jason
Unidentified subject!
unsubscribe
Re: apt-get
On Sat, 1 Jul 2000, Pete Chudykowski wrote: I'm running my ftp service on port other than 21. That confuses apt-get. Is there any way around it? I'm too lazy to edit ftp://localost:54/debian Read man sources.list Jason
Re: package versions in dselect, and Release file
On 25 Jun 2000, Ian Zimmerman wrote: It is not this simple; in fact, I think there's an out-and-out bug. I'll report it when I have some time to waste. It is a bug with whoever typed dpkg-scanpackages because they did it wrong. deb file:/usr/local/src/debs localdebs main non-free Then apt-get update looks for /usr/local/src/debs/dists/localdebs/{main,non-free}/binary-i386/Packages but apt-get install pysol looks for /usr/local/src/debs/main/binary-i386/games/pysol_*.deb. Which is /usr/local/src/debs+[whatever is in Packages for the Filename field] Call dpkg-scanpackages correctly and this will go away. This is related to the thread earlier this week about the semantics of a trailing slash on the deb URL. Apparently the thorough explanation given still haven't exhausted the topic :-( A trailing slash modifies the selection of the Package file, nothing more. Filename: fields are always relative to the URL. Jason
Re: package versions in dselect, and Release file
On 25 Jun 2000, Ian Zimmerman wrote: dpkg-scanpackages binarydir overridefile [pathprefix] Packages binarydir is the name of the binary tree to process (for example, contrib/binary-i386). It is best to make this relative to the root of the Debian archive, because every Filename field in the new Packages file will start with this string. At the very least, this needs a little clarification. What is the root of the Debian archive? I assumed it was foobar/dists/{frozen,stable,whatever}, and it was a natural assumption because it corresponded to the example given. Now we know I was wrong, foobar is the root. But I would never know just from the manpage. When that man page was written its example would have made sense, the structure of the archive has simply changed over time and nobody has updated it. Jason
Re: Multicast Tunnel
On Sat, 24 Jun 2000, Patrick Dahiroc wrote: i've just finished reading the Multicast HOWTO. my question deals with the mrouted tunnel i need to create. my ISP does not support multicast so i can't to tunnel to them. am i just out of luck with multicast or is there a place that sets up multicast tunnels for a small fee (preferably for free). i really would like to have multicast so i can run sdr. I am also interested in the answer to this question.. If you find one please let me know.. Is there anything at www.mbone.net? Jason
Re: apt-get via proxy in Potato
On Wed, 21 Jun 2000 [EMAIL PROTECTED] wrote: I then dist-upgraded to Potato, but now it can't get past the proxy. It says 'connecting to internet' (which happens to be the proxy name), but doesn't get any further. wget on the same machine works fine through the proxy. If it says [Connecting to internet] then it is waiting for DNS to resolve 'internet' Depending on how broken your DNS configuration is you may have to wait a long time, it uses the glibc getnameinfo function which will perform IPv6 lookups first. The way glibc does it is kind of borken so their may be a long delay if your DNS is not set properly. Jason
Re: Keeping up2date with 'unstable' without beeing connected to the internet
On Sat, 17 Jun 2000, Pat Mahoney wrote: The problem is: how do I know _which_ packages are new and which do I have to fetch to update my standalone mirror (without keeping a second mirror at the dial-up/home-cable machine). Run apt-get with the --download-only option. It won't install any packages on your machine, and all the new ones will be in /var/cache/apt/archives. (I think that should work, I've never done it myself.) You could read the /usr/share/doc/apt/offline.text.gz file which describes how to do this. Jason
Re: apt-get
On Sun, 4 Jun 2000, Kerstin Hoef-Emden wrote: deb http://ftp.de.debian.org/debian/dists/ potato main contrib non-free deb http://ftp.de.debian.org/debian-non-US/dists/ potato/non-US main Drop the word 'dists' from both these lines. Jason
Re: Tulip ethernet performance issues
On Sun, 28 May 2000, Damon Muller wrote: I have a couple of tulip-based ethernet cards (I think they are made by Acton, and may certainly be re-badged), one in my Debian box and one in my win98 box. They are connected together by an 8-port 10/100 switch (specifically a LanTech MINI Switch 800 (8 port 10/100 Base-TX Switch) according to the front panel). I've seen this before. The tulip cards are just horrible at NWAY for some reason. Either they get stuck in half duplex or the switch gets stuck in half duplex and you get the problem you saw. I've never actually had a tulip automatically negotiate full duplex with a switch : You can try using the media type forcing options when you install the module, check Becker's web page for information. The carrier+etc errors are your problem, you want to eliminate those. Jason
Re: apt-get, upgrade, /var/cache/apt/archives
On Fri, 19 May 2000, Antonio Rodriguez wrote: 1:) What is the best way to make apt-get use the /archives folder to perform the upgrade and return the system to stage 2. above? cp /archive/*.deb /var/cache/apt/archives/ (or ln -s) Fini. Jason
RE: NIC Cards...is there any difference between them?
On Fri, 12 May 2000, Chris Mason wrote: Aahhh! That's why my realtek cards are not being installed when I boot from cold. In the trash with them. Have you tried the rtl8139too driver? These cards are nice and cheap, perfect for workstations.. Jason
Re: apt [strange error]
On Fri, 12 May 2000, kriptic wrote: the other day i tried using dselect to upgrade as normal, this is what i saw when trying to [I]nstall Reading Package Lists... Done Building Dependency Tree... Done E: The package checker needs to be reinstalled, but I can't find an archive for it. This means the package 'checker' is in a state that is defined to require dpkg --unpack on the original .deb to get out of. APT is telling you that it cannot find that .deb to unpack so checker can be fixed. You will have to do some manual things with dpkg to fix the checker package, probably removing it with a forcing option. Jason
Re: Apt wishlist WAS: Re: crypto patch (OT: ports tree)
On Fri, 21 Apr 2000, Ethan Benson wrote: I guess ultimately, what would be best, would be to keep track of the sources that you have installed, so that you know when the sources have been updated. Or have apt recompile for you. well i just don't understand why apt thinks it should `upgrade' my package whose version number is == to the one its `upgrading' to. The most common case of recompiling is to make something from unstable work on stable, mostly due to library versions. By always upgrading these people get what they want. Everyone else *should* change the version number or put the package on hold. Jason
Re: problems with apt-get upgrade slink - potato
atp-get update-- works fine... apt-get dist-upgrade -- problems... Each time I would only get a few files, but after 15-20 tries, I seem to have almost everything; almost.. Run update again if that fails find out if you are behind one of those defective HTTP proxies :| Jason
Re: apt-get and http authentication
On Tue, 21 Mar 2000, Anton Gyllenberg wrote: I can't get apt-get to do http authentication. It seems sources.list doesn't support http://user:[EMAIL PROTECTED]/ type URI:s. The manuals do mention Heh this got missed, it doesn't support it. You could add it really easially though. Jason
Re: apt behind proxy
On Mon, 20 Mar 2000, Maurizio Boriani wrote: Hi to all, I'm behind proxy, anyone know how i can use apt-get in this case? Yes, read the apt.conf man page. Jason
Re: Apt message Segmentation faulty Tree... 50%
On Sat, 11 Mar 2000, paul wrote: It'd be nice to know what caused it. I'm leaning to hardware problems, as I noticed in my journals that I had some rather funny filesystem and data coruption problems with a similar motherboard (MVP3) and this same HD (Quantum 6.4 gig) last year when the HD was using 32 bit mode. I It is entirely possible that if your hardware corrupted one of those two files that this would happen. They are loaded without any sort of error checking (far too expensive) via mmap. Someday I might arrange to have them auto-erased on segfault. like that. Jason
Re: Apt message Segmentation faulty Tree... 50%
On Fri, 10 Mar 2000, paul wrote: Today, when doing apt-get upgrade on my Potato machine, apt-get exited with the following messages: Usually 'rm /var/cache/apt/*.bin' makes it go away, I dont understand how it is possible to get into a state where that is required Jason
Re: apache, ssl, slink, potato and apt
On Fri, 3 Mar 2000, Mark Symonds wrote: So my question is, what do I do to get ssl back? Am I going to need to get the mod_ssl tarball? auric{root}/usr/lib/cgi-bin#apt-cache search apache.*ssl apache-ssl - Versatile, high-performance HTTP server with SSL support apache-common - Support files for all Apache webservers libapache-mod-ssl - Strong cryptography for Apache ^ Him apache - Versatile, high-performance HTTP server apache-perl - Versatile, high-performance HTTP server with added Perl support libapache-mod-ssl-doc - Documentation for Apache module mod_ssl libssl09-dev - SSL development libraries openssl - Secure Socket Layer and related cryptographic libraries/tools. libssl09 - SSL shared libraries Jason
Re: apache, ssl, slink, potato and apt
On Fri, 3 Mar 2000, Mark Symonds wrote: # deb http://http.us.debian.org/debian stable main contrib non-free # deb http://non-us.debian.org/debian-non-US stable non-US deb http://http.us.debian.org/debian frozen main contrib non-free I did the apt-get update ... why wouldn't it show up? I have another box here which I upgraded to potato awhile ago and tried the same search out of curiosity. It's giving the same output...?! Add deb http://non-us.debian.org/debian-non-US frozen/non-US main contrib non-free Jason
Re: apt-get with a proxy
On Thu, 2 Mar 2000, Jonathan Hall wrote: And I get the following: Err http://http.us.debian.org stable/main wget 1.5.3-1.1 504 Gateway Time-out This is an error directly from squid, there is something wrong on that end. Jason
Re: XDMCP broken after X upgrade
On Wed, 1 Mar 2000, David Karlin wrote: I'm guessing that some new configuration file might have been inserted during the X upgrade. Am I on the right track? Has anyone experienced this, or know how to fix it? Edit things in /etc/X11/xdm - I think it was pretty obvois IIRC. Oddly, it was decided that xdm is some kind of security risk - I personally don't see why.. Jason
Re: quick apt/dpkg question
On Thu, 3 Feb 2000, Ian Alexander wrote: Does dpkg --get-selections report stuff that has been installed via apt? Yes Jason
Re: apt-get probleme
On Wed, 2 Feb 2000, Shaul Karl wrote: I have no experience with proxy auth. However, if I got it correctly then the attached /etc/apt/apt.conf has the skeleton for this to be done. HTTP proxy authentication is very simple: export http_proxy=http://[EMAIL PROTECTED]:proxy.com:3128/ apt.. Your proxy will need to support http/1.1 defined proxy authorization. Jason
Re: dist upgrade the hardway?
On Sun, 30 Jan 2000, Robert L. Harris wrote: silly thing downloaded to my shell, I can ftp it to my office and take it home on a zip disk. Once I get the potato directory restored on ^^ 'on a Large Stack of zip disks' or maybe a CD-R or two. The recursive wget you describe will net several gigabytes. Read /usr/doc/apt/offline.text.gz for the right way to do this. Jason
Re: Debian User mail list
On Wed, 26 Jan 2000, Ehren Wilson wrote: oh and the address I want is [EMAIL PROTECTED] thanx You are gettink Emails Now, Yes? Jason
Re: apt-get and the ftp method
On Fri, 28 Jan 2000, Lindsay Allen wrote: There is no proxy involved and I can use ncftp to connect to the site so it's not a firewall problem. I have removed my apt.conf so it's not that. I also tried Acquire::Ftp true; but could not find any output. The option is Debug::Acquire::Ftp=true it prints to the screen. AFAIK there are no inherent oddities with proftpd servers and APT. I can't duplicate it at least.. Jason
Re: help with dselect (on m68k)
On Wed, 26 Jan 2000, dkphoto wrote: It seems I do indeed also have an issue with mounting the CD. Dselect asks me for the name of a block device. Since I cannot find that term anywhere in any of the documentation, I'm stuck! What is a block device, and how do I get its name? (Should I name it Shirly, Bruce, Albatross?) amber{jgg}~#dmesg | grep -i cd hdb: CD-ROM 36X/AKW, ATAPI CDROM drive hdb: ATAPI 36X CD-ROM drive, 128kB Cache Uniform CDROM driver Revision: 2.55 /dev/hdb is the block device name for the CD. It will be printed during bootup, but you can review the boot messages with the 'dmesg' command. You should edit /etc/fstab and put an entry like this: /dev/hdb/cdromauto noauto,ro,defaults,user0 0 Depending on your m68k, your CD interface may be something entirely different (scsi?) or it may need a special module.. Look at the boot messages. Jason
Re: Problem at ftp.debian.org?
On Sat, 22 Jan 2000, Todd Suess wrote: I did an apt-get update tonight, and then tried to do apt-get dist-upgrade. This is what I get. Little disk space issue there.. The mirror is rerunning now. Jason
Re: apt-get without keepalive
On Tue, 18 Jan 2000, Stuart Ballard wrote: I believe the fix for this would be to configure apt to re-connect to the server every time in a pure http/1.0 stateless sort of way. However, there doesn't seem to be a preference for this. I have tried Acquire::http::Pipeline-Depth 0, which had some effect, but not nearly enough. Yes, that is the best you can do, if this is not enough to fix your problem phone your ISP and complain loudly and tell them to either upgrade or yell at their vendor - these dysfunctional and unadvoidable HTTP proxies must be stomped out :P It is very likely the proxy vendor already has a patch for it, but your ISP is dragging it's feet applying it. Another solution is to simply switch to ftp downloads, but presumably there are disadvantages to this too or it would be the default. FTP is not nearly as efficient as HTTP, but then again, http over a highly defective 'transparent' proxies often looses to FTP anyhow. Jason
Re: dselect/APT problems on potato
On Tue, 18 Jan 2000, Neilen Marais wrote: trick. But what then is the function of the apt.conf file? What am I missing out on? I like fiddling in conf files G You fiddled it wrong.. That config file is not ment for general use, it does 'weird' things. Jason
Re: dselect/APT problems on potato
On Mon, 17 Jan 2000, Neilen Marais wrote: If I choose the apt method in dselect, and do an update, I get the following problem after the package files have been downloaded: 'rm /etc/apt/apt.conf' Jason
Re: apt, dpkg, dselect, and deity?
On Sun, 16 Jan 2000, Chris R. Martin wrote: Forgive me, I've been out of the Debian world for a while. I'm having trouble figuring out the relationship between apt, dpkg, dselect, and deity (which seems to be something in development). I know what dselect deity is APT's development code name, and just they are the same : Deity is also the codename for the APT Native GUI that never got finished. Currently APT proper superceeds all execisting dselect methods and integrates them all under a common interface and lets you mix and match them freely. APT proper can also interface with dselect as a normal dselect method would, it has 5 native front ends (apt-get, Corel Get It, GnomeAPT, Console APT and Aptitute) APT still uses dpkg for the package installation.. Jason
Re: ncurses-base obsolete, but essential
On Fri, 14 Jan 2000, Ben Collins wrote: There is a way to obsolete essential packages. Dpkg itself already handles this. If you install an essential package that conflicts with an already installed essential package, that package will be removed without any sort of --force options. This is on purpose, and was designed for just this purpose. This doesn't entirely work completly. The problem is that if you have two essential packages but can only install one, how do you select which one to install? APT has to make this decision and it is a not an easy problem, it often doesn't get it right either. [It is possible, that is why we have the release files now...] You should be really really carefull with what you do with essential packages, bo-hamm was a real mess but since then we have been getting better. I'd recommend just not getting rid of ncurses-base shrug However, apt does need to realize this, so this bug is valid. Apt will not have to add any --force options to dpkg to get this to work, just recognize the conflict. APT doesn't use dpkg's auto-remove features, it probably should, but that has proven really difficult to get right [and largely unnecessary]. Jason
Re: apt-get with proxy username and password
On Mon, 10 Jan 2000, Gareth wrote: I havtried to use a line like this in the apt.conf http::username:[EMAIL PROTECTED]:port but nothing seemed to happen. You used the wrong syntax.. acquire::http::proxy http://username:[EMAIL PROTECTED]:port; Jason
Re: www.debian.org very slow indeed?
On Sun, 9 Jan 2000, George Bonser wrote: There is no policy for ensuring that the Packages file on 63.209.15.252 matches the files on 207.69.194.216 so failures are frequent. To avoid the Actually 63.209.15.252 recently had some sort of mirroring problem, it should be fixed now. Otherwise the top level mirrors at least stay quite well in sync. Jason
Re: Using dselect With a Proxy Server
On Fri, 7 Jan 2000, Misanthrope wrote: Hello, I have just installed Linux for the first time and am having a few slight difficulties...At present I am running the stable distribution, or slink, and I can't seem to get dselect's asp package selection method working with my home network. My machine is a an i586 166 connected via RJ-45 to a hub (inactive or stupid) and then to my home's proxy server, a 486DX4 100 running Windows 95 (fully updated) which is in turn running WinProxy version 2.1R2h (I want to set up a linux server but my parents won't let me until I can provide a valid reason; a rather difficult task as I can't even manage to get my machine to install packages...). I can FTP out manually by FTPing to my proxyserver, which then provides me with a prompt for [EMAIL PROTECTED]: Munin/~/ftp 10.10.10.2 10.10.10.2 ready: [EMAIL PROTECTED] afterwhich it connects and acts normally...My sole question is how may I get dselect to function with this? Here are a few things which may, in some way help: HTTP Port: 10.10.10.2:80 FTP Port: 10.10.10.2:21 Socks5 Port: 1080 do export http_proxy=http://10.10.10.2:80/; Then edit /etc/apt/sources.list and change all your URIs to http Alternativly, you can setup APT to speak to that ftp proxy, but that is probably lots more trouble than it is worth. Jason
Re: Download once, apt-get install many?
On Fri, 31 Dec 1999, Frank Copeland wrote: Since you are couching that as a question, I assume you haven't actually tried to do it that way. And my answer would be: I don't know. If you ever If you NFS mount /var/cache/apt/archives on all machines APT will magicially do the right thing.. Make sure to use rw,no_root_squash Jason
Re: potato upgrade, probs with perl-base
On Tue, 28 Dec 1999, Pollywog wrote: I am getting into some kind of loop and I need to remove perl-base (temporarily) but it is an essential package. The other way around this is to activate APT::Force LoopBreak but I don't know what that means and the apt man pages do not tell me what that is. They do, it is in apt.conf Force-LoopBreak Never Enable this option unless you -really- know what you are doing. It permits APT to temporarily remove an essential package to break a Con flicts/Conflicts or Conflicts/Pre-Depend loop between two essential packages. SUCH A LOOP SHOULD NEVER EXIST AND IS A GRAVE BUG. This option will work if the essential packages are not tar, gzip, libc, dpkg, bash or anything that those packages depend on. I think you can enable it in this case, the perl situation is very strange. Jason
Docs for the developer DB online
I have written up some documents on how to use the developer DB, they are linked from http://db.debian.org/ In particular, if anyone looses/lost their password I will be directing them to: http://db.debian.org/password.html : If anyone has any questions they would like to see answered let me know and I can update the pages. Thanks, Jason -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-cache dumpavail /tmp/apt-chace; dpkg --update-avail /tmp/apt-chace;
On Sat, 25 Dec 1999, Shaul Karl wrote: I want to avoid the need to run update from within dselect after apt-get dist-upgrade by writing the lines Post-Invoke { apt-cache dumpavail /tmp/apt-chace; dpkg --update-avail /tmp/apt-cache; rm -v /tmp/apt-chace; mount -o remount,ro /usr; }; Is that reasonable? Are there any issues of file locking, breaking the regular backups of the available and status files or other things that needs special care? In theory you can simply do 'apt-cache dumpavail /var/lib/dpkg/available' the dpkg --update-avail step is somewhat redundant. Jason
Re: apt-find
On Mon, 20 Dec 1999, Marco Giardini wrote: After having upgraded (using apt-get) my slink distribution (intel i386) the usefull apt-find doesn't work any more. Use console-apt, which is the what apt-find was renamed too Jason
Re: problems with debian mirrors
On Sat, 18 Dec 1999, Gregg Berkholtz wrote: I have problems like this sometimes too -- any ideas on how someone might get around a transparent proxy or to force the proxy to update itself. I dont have any control over the proxy. No idea, I don't think you can - they are built so they cannot be circumvented. Contact your ISP and complain - really, the vendors need to get a clue and start actually implementing the RFCs. If your proxy is *good* you can tell APT about it directly and APT will set the proper HTTP/1.1 cache control headers to do expires and things. But I know of no proxy that supports them yet. Again, call your ISP and complain. Jason