Re: 2 versions of some packages
Christian Hudon <[EMAIL PROTECTED]> writes: > Hmm. I don't really know much about dftp, but when dpkg-ftp sees that two > of the files it has downloaded are different versions of the same package, > it installs the more recent version and discards the older one. Can't dftp > do something like that? Or maybe there's something I'm missing... Well, I'm not sure that installing the most recent one is the right answer. Take ghostscript for example. There's the GNU version in .../text and the Aladdin version in .../non-free. Now the GNU version should always be <= the Aladdin version since GNU repackages the Aladdin version after it's been out for a while. People who want an all free software machine may prefer the GNU version even if it's older. I think that it might make more sense to just disallow two packages with the same name in a given Debian package tree. In this case I think we should have a virtual ghostscript package provided by both gs-aladdin and gs-gnu, or whatever. Then no one gets confused about what you mean. -- Rob -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Can't do setuid
Jim Worthington <[EMAIL PROTECTED]> writes: > It appears that Debian Linux does not have setuid capability. Is this a > feature that I can turn on through a configuration file? I would recommend that you investigate the sudo and super packages. They will allow you to do exactly what you want much more securely. -- Rob -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Latex Help !
Jan Ramon <[EMAIL PROTECTED]> writes: > dvi to word seems difficult. I know a utility latex2wp (or wp2latex ?) > for converting between wordperfect (5.1 ?) and latex (.tex-files). Perhaps > you can use this in combination with wordperfect-word conversion ? I believe that there are also tools to go from LaTeX to Rich Text Format files which Word understands. I'm not sure what the name(s) of the tool(s) are though. -- Rob -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: The *"'($^"'( list... and Xdm login
Buddha Buck <[EMAIL PROTECTED]> writes: > How many utilities would have to be changed to implement the > following password logic: > > If the stored hash is 13 characters long, use crypt(). > If it is 32 characters long, use md5sum(). > Otherwise, fail. > > This would allow us to retain backwards compatability while potentially > increasing the security of the system by allowing arbitrarily long > passwords. I would love to see this happen. Wouldn't it be fairly easy if we successfully migrate to libpam? -- Rob -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: util-linux
"Ervin D. Walter" <[EMAIL PROTECTED]> writes: > I find that the clock binary in util-linux 2.5-6 still makes my real > time clock go nuts. The binary from 2.5-4 works fine. > > Am I missing soomething? It's broken. -- Rob
Re: Scsi errors
[EMAIL PROTECTED] writes: > device, but I don't know... Are there any terminators that would go on > the end of a 50 pin ribbon? Normally, there's a switch, or jumper you have to set that terminates the drive. If the drive's not terminated, and is at the end of a scsi chain. You'll most likely have problems. I didn't think SCSI-III was special in that respect. -- Rob
Re: Graphical package management tool
Tom Fawcett <[EMAIL PROTECTED]> writes: > I'm envisioning something like a cross between Debian's dselect and > Red Hat's glint utility. Before I start in on it, has anyone already > written anything like this? Is anything planned or in the works? > Any warnings or encouragement? I think that Ian may have recently added some functionality to dpkg that would make this easier. You should build it on top of dpkg, and one of the problems was determining something like all the recursive dependencies of a given package. I think dpkg might do this for you now, but I'm not sure. One other place you could look for relevant code is the new perl version of dftp. It may have some useful stuff related to parsing Packages files, installing packages, etc. -- Rob
Re: Unsubscribe from this list...
[EMAIL PROTECTED] (Miquel van Smoorenburg) writes: > Unfortunately not _all_ mailing lists are available as newsgroups, > I guess I'll have to put up with that until somebody writes an > integrated mail/news reader (AFAIAC mail is news is mail etc). That would be any recent version of Gnus. The best mail/news reader around (at least IMHO). -- Rob
Re: hint on handling mail volume: sortmail
"Paul Christenson [N3EOP]" <[EMAIL PROTECTED]> writes:
> Could the list maintainer set up the list so that all subject lines
> start with a unique string, such as [DebUser]? A number of other lists
> I'm on do this; it would make automatic sorting a lot easier. (Message
> number optional.)
You don't need to do this if you have a decent mail program. I use
GNUS, and all it takes is:
(setq
nnmail-split-methods
'(
("duplicates" "^Gnus-Warning:")
("debian-private" "^Resent-From: [EMAIL PROTECTED]")
("debian-devel" "^Resent-From: debian-devel@lists.debian.org")
("debian-user" "^Resent-From: debian-user@lists.debian.org")
("debian-bugs" "^Subject: \\(Re: \\)?Bug#[0-9]+:")
("executor" "^Sender: [EMAIL PROTECTED]")
("guile" "^Sender: [EMAIL PROTECTED]")
("inbox" "")))
I'm sure procmail, etc. have similar capabilities.
--
Rob
Re: Problem with GIMP
Philippe Troin <[EMAIL PROTECTED]> writes: > I have the same problem here. I'm afraid it's a GIMP bug. I'll see about forwarding it upstream, but there's a major upgrade the works so chances are this would be wasted effort. -- Rob
Re: bleh...
[EMAIL PROTECTED] (Ronald van Loon) writes: > This problem goes away when you recompile your kernel *without* extended > Real Time Clock support. > > Ronald van Loon ([EMAIL PROTECTED]) Actually I believe that on the Dell portable we have here, it still segfaults, even without RTC support. If we back off to an older version of util-linux, the problem goes away. -- Rob
Re: static vs. dynamic - gimp
[EMAIL PROTECTED] writes: > gimp-dmotif (dynamic) Use this one if you have purchased and installed Motif on your computer. > gimp-smotif (static) Use this one otherwise. -- Rob
Re: Cross posting per request
[EMAIL PROTECTED] (Bruce Perens) writes: > (like RCS) on /etc . "dpkg" doesn't currently know how to check control files > in and out of RCS - is this a good idea? Currently, it will leave a > "filename.dpkg-new" file around for you to hand-edit if you decline to > over-write a control file. This sounds like a really nice idea, but I bet there could be nasty logistical problems. -- Rob
Re: chat problems
[EMAIL PROTECTED] (John Ross) writes: > The "chat" program cannot find my modem, and as far as I can tell, there > are no command line switches to force it to use /dev/modem or /dev/cua1. Chat only reads from standard input and writes to standard output. pppd just launches it with it's standard input and output connected to the serial port. You can do something somewhat similar with chat <> /dev/ttyS0 -- Rob
Re: perl/tk
Dirk Bernhardt <[EMAIL PROTECTED]> writes: > ii perl5.003-2 > ii perl-tk b11.02-2 > What can I do? I'm pretty sure you have incompatible versions of perl and perl-tk. I believe that perl-tk b11.02-3 is the first version compatible with perl 5.003. You can get this version from ftp.debian.org:/debian/unstable/binary-i386/devel/perl-tk_b11.02-3.deb -- Rob
Re: mailing list
Steve Dunham <[EMAIL PROTECTED]> writes:
> Or, if you prefer, you can use GNUS. For me, procmail dumps
> debian-user into the file incoming/mail.lists.debian-user.spool, GNUS
> automatically adds those messages to a special mail folder,
> 'mail.lists.debian-user', which looks just like a newsgroup. I sort
> about 10 mailing lists this way and read them just like newsgroups.
> Some have auto-expire set, others do not.
You can also do this directly from GNUS without procmail. Something
like:
(setq
nnmail-split-methods
'(
("duplicates" "^Gnus-Warning:")
("debian-devel" "^Resent-From: debian-devel@lists.debian.org")
("debian-user" "^Resent-From: debian-user@lists.debian.org")
("debian-bugs" "^Subject: \\(Re: \\)?Bug#[0-9]+:")
("guile" "^Sender: [EMAIL PROTECTED]")
;"^\\(To\\|[Cc][Cc]\\):[^:[EMAIL PROTECTED]:]*\\(^\\w+:\\|\\n\\n\\)")
("inbox" "")))
works for me. Gnus then just gets and splits the contents of
/var/spool/mail/whoever itself.
> This gives you scoring and threading for free. Without this, I would
> not read mailling lists.
Sure does make it easier anyway.
--
Rob
Re: Transfering system.
Mike Wood <[EMAIL PROTECTED]> writes: > bring the system down to single user mode and mirror/copy(how?) the > entire system onto the new partitions There are many ways, but I'd suggest (assuming that you have the new partition mounted on /mnt, and assuming that you want the entire old drive on that partition) su cd / find -not -regex "^\./proc/.* -not -regex "^\./mnt/.*" | afio -p /mnt This should get the entire contents of the old drive onto the partition mounted on mnt. Note that it will handle devices, links, timestamps, the works. The only thing that I know of that might not be captured perfectly is the timestamps on the directories. If you don't want to cross file system boundaries, add the -xdev flag to the find command. If you need to get a different set of files, just modify the find command to get the ones you want. If you want to list the files being copied, just add the -v flag to the afio command. And of course, if you want to make sure you're getting the right files before actually running the command, just use /dev/null in place of /mnt, or just use less in place of the afio command. Anyone please correct me if I've gotten the command wrong. I wouldn't want someone running something broken as root... -- Rob
Re: tar-1.11.11-2.deb broken?
[EMAIL PROTECTED] (Thomas Baetzler) writes: > Anyways, downgrading to the tar of the buzz release seem to fix this for > me. Is this a tar or a make-kpkg problem? As I understand it, that upstream version of tar is "broken". Debian is going to revert to an earlier version until the problems are fixed. You did the right thing. -- Rob
Re: cron
Al Youngwerth <[EMAIL PROTECTED]> writes: > If anyone really knows how to get cron to stop pestering me, let me = > know.=20 If the command has output, you will get the results mailed to you. I don't know of any way to stop that, but it's easy to arrange for the command to avoid sending results. Just make your command cmd > /dev/null or the more thorough cmd > /dev/null 2>&1 which redirects stderr and stdout. See "man bash" for more details. -- Rob
Re: Kernel headers
[EMAIL PROTECTED] (Miroslav Ruda) writes: > Is any reason why not to use this schema in Debian too? Read /usr/doc/libc5/FAQ.gz -- Rob
Re: does Linux support 2940 Ultra Wide SCSI?
Lawrence Chim <[EMAIL PROTECTED]> writes: > I am considering to buy a 2940 Ultra Wide SCSI card. Does anyone > using it? Does the kernel support it yet? I just started using one with my new machine, and it appears to work fine. You might need to make sure you have a fairly recent kernel. I will note that this card does not appear to cooperate well with the older AHA-2940. I tried using both in the machine at the same time, and it wouldn't quite work. It is supposed to work if you have identical cards. -- Rob
Re: How do I allow users to run a single command as root?
Casper BodenCummins <[EMAIL PROTECTED]> writes: > (a) packages which control superuser execution are not generally > found in distributed commercial UNICES (and how many sysadmins > have the time or the inclination to seek out these packages?); Hmm, I hadn't thought about that. Ok, interesting consideration. -- Rob
Re: Solved: mouse problem (running both gpm and X)
Mark Phillips <[EMAIL PROTECTED]> writes: > Am I right in thinking that the cua devices are now obsolete and that > ttyS devices are the way to go? That's what I was told. I found that switching solved my modem problems. -- Rob
Re: libwww-perl installation incomplete
Bill Wohler <[EMAIL PROTECTED]> writes: > the current version of the libwww-perl package is incomplete. In > particular, it does not install the HTML sub-package. After > obtaining the sources to libwww-perl and installing, I got past > the errors above. Hmm. So did you have to do anything special when you built libwww-perl to get this sub-package, and did you do this with the real upstream sources, or the debian source package? > Could this be user headroom on my part or shall I submit a bug > report? This sounds like a bug to me. Go ahead and submit a bug report so I don't forget to look in to it. Thanks -- Rob
Re: RAID in Linux (was Re: SCSI and EIDE)
Ricardo Kleemann <[EMAIL PROTECTED]> writes: > Few questions: > 1. can raid0/raid1 be done on either scsi or ide or both? The linux md driver is software, so it doesn't care what the underlying block device technology is. You just give it block-device partitions to turn into a RAID device. These partitions can (at the least) be on SCSI or IDE devices. It dosnt' matter to the md driver. > 2. what's the difference between raid0 and raid1? >From /usr/doc/mdutils/README: Raid0 does a classic (and rather efficient) striping on disks (i.e. contiguous blocks on the md device are spread across real devices). It gives rather good performances on SCSI disks, specially with concurrent disk access. There's no limitation on disks sizes (i.e. sizes can be different, md will cope with this). Raid1 adds mirroring to raid0 striping. Note that it is not complete yet (no rebuild tools, error trapping is incomplete). It's also known to be rather slow when writing. >From the same source: Since MD pre0.31, RAID-[15] have been removed from the main distribution, because of pathological unstability... and for the kernel integration of linear and RAID-0 modes. > 4. How does one go about "creating" an md device? Would it automatically > mirror a "non-md" drive into the multiple devices? Get the debian mdutils package install it, and read the docs. -- Rob
Re: [Fwd: Virus Alert]
[EMAIL PROTECTED] (Terry Eck) writes: > I remember reading about this several months ago. It was the opinion > of most people at the time to be just a joke. Anyone know for sure > if it is anything other than a joke? Well, I'd like to see the code for the infinite loop that'd melt down my processor. I've written a few in my time (accidentally), and the machine's still here. I think this is probably a joke. I actually have received a "GOOD TIMES" message, actually got one a couple of times, way back when I used AOL (feel free to flame, it was pre-linux for me, too). Anyway, the message was a virus in some sense, but it required active participation. It was a message saying that it had been around the world X number of times and that anyone who would do what it said and continue the pyramid, tacking on their own name, would get great wealth and voodoo luck. If you failed to do this in Y number of days, all manner of evil catastrophies would befall you. I ignored it with no ill effects at the specified instance of doom. -- Rob
Re: How do I allow users to run a single command as root?
Casper BodenCummins <[EMAIL PROTECTED]> writes: > What you need here is to set the setuid bit. Run this command as root: > >chmod +s filename Not to be nasty, but this is generally a *REALLY BAD IDEA* unless you know *exactly* what you are doing. If "filename" was not designed with extremely careful attention to the fact that it's going to be run suid root, you can be opening up your system to all kinds of security attacks, or accidental disasters by enabling suid root. I'm talking about attention to things like explicitly setting the PATH, checking and setting IFS, etc. If it's a perl script, using the "taint" checks helps, but you have to know enough perl to be able to fix the problems it reports. Don't do this. Use sudo, super, or some equivalent. [end preach mode] -- Rob
Re: Isn't it a security hole...
Casper BodenCummins <[EMAIL PROTECTED]> writes: > Fewer than that. The range of ASCII characters used in passwords is > quite small: perhaps ~= 110, optimisticly taking into account control > characters and punctuation marks. Then, many people don't use the full 8 > characters, so we have to reduce the _average_ number again. I'd > >tentatively suggest 6 characters. It's too bad we can't support longer passwords. Not only are longer ones easier to remember sometimes, but they are harder to break. Something along the lines of the PGP passphrase. I'm sure there's some historical, or backward compatibility reason why we can't... -- Rob
Re: How do I allow users to run a single command as root?
[EMAIL PROTECTED] (Tim O'Brien) writes: > Any ideas on how to accomplish this? I tried reading the man pages, but > su (which I think is what I would need to use) is not exactly what I'd > call well documented. su will only work if you want to give the users the root password (not a good idea). What you want is either the sudo package or the super package. I use sudo. It lets you specify which users are allowed to run which commands as root. Then, to run the command, the user types sudo some_command maybe with args and sudo will ask them for *their* password, and if it's OK, run the command for them as root. "man sudo" has all the details. -- Rob
Re: SCSI and EIDE
Ricardo Kleemann <[EMAIL PROTECTED]> writes: > Now will Linux implement anything greater than RAID0? > Would you say your performance is significantly increased with striping? It also currently handles RAID1, and linear; higher levels have been temporarily disabled because they are still too buggy. I don't know about performance. The machines certainly *seem* faster, but I didn't do any real testing. I would have done hdparm -t, but it turned out it wouldn't work for md devices. -- Rob
Re: SCSI and EIDE
Ricardo Kleemann <[EMAIL PROTECTED]> writes: > That's wonderful! > > Now will Linux implement anything greater than RAID0? > Would you say your performance is significantly increased with striping? > > How many drives can be striped? That I don't know. I'm not sure there's a (small) limit. I didn't find one in a cursory glance of the manpage for mdadd, or in a similarly cursory look at the stuff in /usr/doc/mdutil. md just groups a number of physical disk partitions into one logical one, /dev/md*. Then you treat /dev/md* as the actual device, and you can mkswap it, mke2fs it, whatever. It doesn't state any limitations on the number of partitions that can be added to a /dev/md*. Note though that with RAID0, you can't add more partitions after you have data on an md device, and you have to create the device in exactly the same fashion every time. Normally this happens at boot from /dev/mdtab. -- Rob
Re: How to debianize packages?
[EMAIL PROTECTED] (Volker Ossenkopf) writes: > To maintain a well organized system I would like to install them > with the Debian package installation tools and not by hand or at > least to register them when being installed as debian packages. > > To focus the problem: How do I convert tar.gz archives (when > no installation scripts are needed) into debian packages? > How do I create a debian archive from installed files ? This is really not supported well, if at all. You should probably just install these things in /usr/local. For the most part Debian doesn't touch that dir. It's specifically for this sort of thing. Then, if there are really great things that you'd like to see as debian packages, you might consider packaging them yourself, and adding them to the distribution. -- Rob
Re: SCSI and EIDE
Ricardo Kleemann <[EMAIL PROTECTED]> writes: > Good point about RAID0. > > I know linux's implementation of RAID0 is with the MD program. Do you > know if that's fully functional, and if so, is there a Debian package for > it? How is it used? I'm using it on two machines in the lab. One has two identical Quantum Fireballs, and the other has a Fireball and a Western Digital Caveat (ahem Caviar). On each machine most of the partitions are striped, but at least one is not. It works fine. You need to compile a kernel with bultiple device/RAID0 support and you need to install the debian mdutils package. Then read the stuff in /usr/doc/mdutils. That's it. The md package doesn't care if the drives are the same type, or even on the same kind of bus. That's the advantage to a software implementation. Granted I would expect performance to be best when you have 2 identical drives on the same kind of bus... -- Rob
Re: SCSI and EIDE
Al Youngwerth <[EMAIL PROTECTED]> writes: > Sorry, I just had to get in my two bits about this... No problem. > The drive's CPU can always make better decisions about reordering = > requests because it knows the details of the drive's parameters and > its = current state. A drive's CPU can reorder requests based upon > seek AND = latency; the system's CPU cannot. I see your point. I was mostly just voicing a general preference about avoiding embedded systems. In this case using the embedded control might be warranted, but in general, embedded systems are a real pain to deal with, and most people can't do much to fix them when they're broken. -- Rob
Re: SCSI and EIDE
Sherwood Botsford <[EMAIL PROTECTED]> writes: > I agree that you can do this in the OS, but I don't think that it > *should* be done in the OS. > 0.In general smarts should be at the point they are used. We had a > VAX that was about as speedy as a 12 MHz 286 with 287 co-processor. > However, that vax could handle 8 simultaneous 19 KB terminal lines. > How? Smart serial cards. > The CPU should do those tasks that are either too general for > dedicated hardware, or that don't happen often enough to take up much > time. > 1.The OS shouldn't have to care about the layout of the disk. Not trying to start a flame war here, but if the overhead to the cpu is minimal, which I would guess it is for seek algorithms, you are often better off doing it with the cpu. Mainly because changing the algorithms, fixing bugs, trying out new approaches, etc. can happen with a simple kernel recompile, and improvment's in CPU speed (which happen all the time, as opposed to the somewhat slower embedded controller market) immediately translate to improvements in algorithm speed. Now if the disks just had some flash ROM and a 386 (or equivalent that gcc knows about) on board, things might be different. Either that or a couple of cheap 386's on the motherboard for stuff like this... I'm mostly just speculating. I don't know that much about drive hardware. -- Rob
Re: Real Audio
Christian Lynbech <[EMAIL PROTECTED]> writes: > I upgraded to 2.0.0 and found much to my pleasant surprise that my > soundblaster card had started to work. Then I upgraded to 2.0.6 and it > stopped working again (I haven't had time to properly report it yet, > so I haven't the error message at hand but it was something about > allocating an IRQ). Downgrading back to 2.0.0 fixed things. Same problem here with a CrystalLake card, but it was again fixed in 2.0.11. -- Rob
Re: About Triton DMA IDE vs. SCSI performance
[EMAIL PROTECTED] (Bruce Perens) writes: > I didn't say wide fast. On my system I am getting about 5 MBPS on IDE from > a Quantum Fireball drive, which is as fast as my narrow fast SCSI drive > on a narrow fast Adaptec AIC-7850 SCSI on the same motherboard (an IWILL > P54TS). The speed seems to be limited by the disk media, not by any > difference between IDE and SCSI. What have you used for testing? I'm curious to test my own system. Something more sophisticated than hdparm -t? > No doubt SCSI works better for multiple devices on a single controller. > However, for my configuration this is probably moot - the Triton > provides two DMA capable IDE busses each with its own PCI bridge, and I > have one disk on each. So from this it sounds like if you have two IDE drives, you are better off to put one on each IDE bus. Is that true? I looked for some info about this, but never found any. Also, I started using the md (RAID0) driver on a couple of machines, and it seems to work well. However, you can't test performance with hdparm since it doesn't understand md devices. Is there a good alternative? One one machine I striped 2 identical Quantum fireballs, and things seem good, but the other machine has drives of different makes/sizes. The md HOWTO says that you can stripe drives of different makes and sizes, but it doesn't say when it makes sense (performance wise) to do this. I did it (just to try it out) with a 1GIG Fireball and a 500MB Caviar. Could this actually be a win? I know that you have two mechanisms doing the fetching, but if one drives is slower than the other, it seems that you might have a problem... Thanks -- Rob
Re: Serial ports...
Sherwood Botsford <[EMAIL PROTECTED]> writes: > CAVEAT I'm an ignoramlepuss about Linux/Debian. [ good info about serial ports deleted ] I do recall a discussion recently where someone posted text from the maintainer of the Linux kernel's serial port code that essentially said that cua* devices were no longer really supported, and that you should always use ttyS* when possible. I was having some trouble with cua* at the time (for dialout), and sure enough, switching to ttyS* fixed the problem. -- Rob
Re: chat bugs
Randy Gobbel <[EMAIL PROTECTED]> writes: > I installed Debian Linux 1.1 on my new Pentium Pro a couple of weeks ago. so > far everything mostly works great, but I've noticed that the chat program > seems to have a number of problems: Just a personal experience: I had all kinds of trouble with the modem that came with my Gateway 2000 (the Telepath internal -- almost a Sportster, made by USR for Gateway). I had dropped connections, chat problems with recognizing REPORT and ABORT strings, etc. When I recently replaced it with an external Courier, *all* the problems vanished. Anyone want a cheap internal modem :> -- Rob
Consensus about /bin/perl.
So what was the decision about /bin/perl, and the packages that depend on it right now (kernel-*, etc). Should it be a symlink, or should these packages just be fixed? Just wanted to know so know whether or not I should create the link or patch the scripts on my systems. Thanks -- Rob
Re: How can I submit packages to be included in the distribution?
Dermot Bradley <[EMAIL PROTECTED]> writes: > Started work on it last night - just a bit confused with libgr/libgr-dev > versus libtiff3/libtiff3-dev/zlib1 needed for building it. libgr is dead. Use the individual libs. -- Rob
Re: a memory + SCSI error??
Sounds to me like you might have unseated one of the cards, most likely SCSI, or loosened a SCSI cable. You might want to check that. Also make sure all your termination switches (if any) are right. -- Rob
Re: rename a machine
[EMAIL PROTECTED] writes: > Where is this info written ? /etc/hostname at the minimum. -- Rob
Re: GCC frepo patch don't work.
Jimen Ching <[EMAIL PROTECTED]> writes: > Has anyone ever tried the -frepo feature? It is actually patched in? > Using the switch to g++ doesn't tell you anything since g++ ignores > arguments it doesn't understand. This feature is very important. I can't > continue my project without it. Please help, thanks in advance... Yes, we use it here, but you have to go to cygnus (ftp.cygnus.com/pub/g++), get the patch, and the debian source package and rebuild gcc with the patch yourself. -- Rob
Re: mail-delivery-agents?
Rob Browning <[EMAIL PROTECTED]> writes: Several people have asked me for my Gnus setup, so I'll post it here. Hope no one's offended, but it has a fairly high signal to noise ratio if you're interested in gnus for mail with debian. This is stuff I've sent to a couple of people before. Feel free to ask if you have any questions. Gnus supports a number of backends for mail. It can handle RMAIL, mh, etc, but I would recommend that you switch to using the nnml back end. It's much faster since it generates NOV overview databases (the same ones used by news servers) for all the mail folders. Other than that it stores mail just like mh, one message per file. Makes scripting over messages easy, but it can be hard on the inodes :> I think that nnml is compatible with mh, but only if you don't use both at the same time. I think that gnus won't notice changes made by mh with this method until restart and vice versa. I'f you're already using mh, you don't have to do anything special to use nnml other than tell gnus to use it. It will then run over your mh heirarchy and automatically generate the databases. That's what I did. All my code assumes you are using nnml, but that should only be an issue in a couple of places where I set up the mail group servers. I'm going to assume for now that you don't want to use the bleeding edge release, and that the one that came with your emacs is OK. Is it emacs-19.31.1? If not, you may need the newer version of Gnus (or emacs). I will have info below about how to use the bleeding edge if you're interested. I'm using 5.2.38 to write this and it works great. Gnus treats mail groups (as opposed to news groups) specially with respect to deleting articles. The normal rule is: "if it's a mail group, never delete anything unless specifically told to, and then only after the expiration period." You expire mail articles with "E" in the summary buffer. The default expire period is two weeks. If any article is two weeks old and has been expired, it is deleted. This is kind of nice if you later (within two weeks) decide you really did need some old message. Also, expired articles (until their deletion) still show up in threading histories, etc. You should also look in the info pages (gnus' are quite complete, Lars is insanely prolific) at the section on "Topics". This allows you to organize your groups in collapsible outline form. With respect to file system layout, I have my groups set up like this: ~/Mail/Incoming/inbox(incoming mail that doesn't go anywhere else) ~/Mail/Incoming/debian-bugs (debian bug list) ~/Mail/Incoming/... (other assorted mail groups) and for outgoing stuff: ~/Mail/Outgoing/misc-mail ~/Mail/Outgoing/misc-news this used to be the default in gnus, but in the latest version, it saves no mail by default, so I explicitly enable this in my .emacs. For news (don't forget to check out the info pages on persistent articles and asynchronous fetching -- too cool): ~/News/... (all Gnus related stuff for *incoming* news goes here) Note that it might make more semantic sense for misc-news to go under ~/News, but it turns out that would be pretty awkward to manage with the way Gnus wants to handle things right now. I don't think you could do it without more hassle than it's worth, and the way I have it now, all personal writings are in one subdir. A couple of other things of note. If you just want to read mail, without going through the time consuming (at 28.8) process of contacting the news server and getting the current news group status, you just use M-x gnus-no-server instead of M-x gnus to launch gnus. One caveat, you might notice that all of the mail groups vanish along with the news groups when you do this (i.e. when you run gnus-no-server). That's just because groups have levels of "activeness", and gnus-no-server only shows groups of level 2 or lower. Unfortunately all groups, including mail groups, are created at level 3 by default, so they vanish with gnus-no-server. To see them again, you just need to launch gnus the old fashioned way (M-x gnus) and then use the set level command to set the level of all your mail groups to 2. You can set the levels via "S l" with the cursor on the relevant group in the Group buffer. In order to actually get your mail, you might be able to use pop directly from gnus (I think that's supported now, but I'm not sure), but I just use popclient to grab my mail from all the relevant machines, and then let gnus suck up the mail from my normal system mailbox, which it does by default. Gnus automatically gets your new mail when you launch it, but if you want to incorporate all the new mail after launching gnus, you can use "g" from the group buffer which updates all the mail and news groups,
Re: LILO and W95
"James D. Freels" <[EMAIL PROTECTED]> writes: > If it were that bad, why are there so many people using it? I'm not commenting on anything but this line. IMO this may be a reason why you have to put up with something, but it is never a good argument for the merits of an item. -- Rob
Re: mail-delivery-agents?
[EMAIL PROTECTED] writes: > I can really recommend emacs with the vm mode. Really. My personal favorite is Gnus. It does a *wonderful* job with mail. Threading, filtering, deleting duplicates, etc... Getting it set up just the way I wanted was a little tricky, but anyone who's interested can mail me, and I'll give you an intro. -- Rob
Re: gcc and libg++27
Dale Scheetz <[EMAIL PROTECTED]> writes: Are you using STL? If so, I was completely unable to compile certain bits of STL code until I modified the bottom of /usr/include/g++/std/bastring.h like this: #if !defined (_G_NO_EXTERN_TEMPLATES) //#include breaks stl (RLB) -- ambiguous instantiations // same as complext.h including cinst.h #endif and complext.h like this: //#include breaks stl (RLB) -- ambiguous instantiations. // same problem as bastring.h including sinst.h Don't know if this is your problem, but I thought it might help. Hopefully libg++ will get this straightened out upstream soon. -- Rob
Re: ftp.debian.org problems
Martin Konold <[EMAIL PROTECTED]> writes: > On Mon, 17 Jun 1996, Bruce Perens wrote: > > > We suggest you use the same mirror script we use, and set it so that if > > more than 10% of files go away at once it won't delete them. > > Which one do you use? Where to get it? 10% seems reasonable. Isn't it the one in the Debian mirror package? -- Rob
Re: debian.93 and inn
Chris Rhodes <[EMAIL PROTECTED]> writes: > Rather than install Debian 1.1, I would prefer to stay with Debian .93 > because my reason for changing from Slackware is it's general unreliablity > (and I have already spent a lot of time getting this far!) Don't take this the wrong way, but you are wasting your time with 0.93. 1.1 has been a long time in the making and is more stable than 0.93. I highly recommend you bite the bullet and go with 1.1. Besides all the bug/security fixes, new packages etc. will be for 1.1. -- Rob
Re: Meta/Hierarchical packaging?
[EMAIL PROTECTED] (Amos Shapira) writes: > Got the idea? Any comments about it? This kind of thing has been discussed a little, but the discussion was postponed until after the 1.1 release. -- Rob
Re: critical :) nethack question
[EMAIL PROTECTED] (Bruce Perens) writes: > I work for two years to build the system and this is what people use it > for :-) Who do you work for :> Anyway, what else would you do with a computer :> :> -- Rob
Re: manpage compression
Derek Lee <[EMAIL PROTECTED]> writes: > is there some way ask the install scripts to do so? I cannot just go This issue has been discussed, but not resolved. It was decided to wait until after the 1.1 release to deal with it. For now you would probably be better off to try and just live with the uncompressed pages. Compressing them yourself would, as you accurately surmised, probably cause more problems than it solves. -- Rob
Re: .xinitrc
Yves Arrouye <[EMAIL PROTECTED]> writes: > When one uses xdm, the .xsession file is used instead. You can just have > one that looks like: > > #! /bin/sh > > sh $HOME/.xinitrc Or just ln -s .xinitrc .xsession which has worked fine for me. I do the same for .Xdefaults -> .Xresources, although someone recently told me that that they weren't so sure about that, but it seems to work... -- Rob
Re: package conflicts
Branden Robinson <[EMAIL PROTECTED]> writes: > 3) Are there Debian packages for the following (as of 0.93R6, there > weren't)? > kermit Don't think so. > mpeg viewer Yes, ucbmpeg. > pdf (acroread) Yes, xpdf. > psfonts Don't know what you mean. > vrml Don't think so (feel free to make one :>). -- Rob
Re: installation notes
Scott Barker <[EMAIL PROTECTED]> writes: > Where is the option for this ftp method? I'd love to give it a try, but it's > not listed in the access methods of dpkg-1.2.6elf You have to install dpkg-ftp. -- Rob
Re: More groff weirdness
Branden Robinson <[EMAIL PROTECTED]> writes: > Now poke yer nose into groff.list and just see if you can find a directory > called /usr/doc/groff. Nope, hain't there. > > Again, I'm using 0.93R6. If this has been fixed since, I'd love to hear > about it. I'm poor and would have to suck the new distribution through a > 14.4 line, so please cry for me. :) I have all the docs, but I'm using the new Debian 1.1 package. Time for the kleenex... -- Rob
Re: .xinitrc
Rick Hawkins <[EMAIL PROTECTED]> writes: > and now emacs is telling me that it doesn't recognize a -geom option. > huh? so how do i get it to place itself automatically? Are you using -geom or -geometry? I use the latter and it works. -- Rob
Re: .xinitrc
Rick Hawkins <[EMAIL PROTECTED]> writes: > one more, probably minor thing: with it set up for xdm, it ignores the > .xinitrc file in my home directory. am i missing something obvious? Is it's executable bit set? /etc/X11/Xsession checks this to decide whether or not to run the user's script. -- Rob
Re: Strange IP address...
Branden Robinson <[EMAIL PROTECTED]> writes: > In a nutshell: this is normal. You don't have a problem. I'm not sure if this is related, but there was one problem with the *configuration* of the loopback interface that was fixed in the latest package. The original /etc/init.d/network read: # Configure the loopback device. ifconfig lo 127.0.0.1 route add 127.0.0.1 It should read: # Configure the loopback device. ifconfig lo 127.0.0.1 route add -net 127.0.0.0 dev lo This is mentioned in the Changes file for the new kernel (among other places). -- Rob
Re: Debian 1.1beta problems
David Gaudine <[EMAIL PROTECTED]> writes: > This does work. Since I don't know how to find the configuration > programs for some package, I use dselect to remove the package and > then to reinstall it. I assume there's a better way, but this does > work. There's nothing too terribly wrong with that, but often a package will tell you about it's config programs during the install process. Failing that, or if you forget what it told you, you can also either try reading the docs in /usr/doc/packagename, or more directly, try "dpkg --listfiles packagename". -- Rob
Re: kernel header files
Lawrence Chim <[EMAIL PROTECTED]> writes: > Is it necessary to install lib5-dev in order to compile programs. Yes > I don't know why there is a libc5-dev. If I install a new kernel, > it comes with its only kernel headers. Install libc5-dev and read /usr/doc/libc5-dev/FAQ.gz. > Also, if I install libc5-dev, can I still install a newer version of > linux kernel whenever it is available. Yes, use the kernel-package package if you want to roll your own. It's still in the experimental section because the user interface may change, but it works great. You can also just use the debian kernel-source pakages if you are happy recompiling the version that debian has available. -- Rob
Re: POP forwarding
Douglas Bates <[EMAIL PROTECTED]> writes: > Does anyone have suggestions on > a) does this seem like a reasonable approach? > b) if so, what would be good tools for creating such a POP forwarder? > I have some experience with sockets under perl if that seemed to be a > good way or I could treat this as an opportunity to get more exposure > to Java. Both of these methods would take me a bit of time. Is there > some tool in the Debian distribution that could do a quick-and-dirty > job on this? There are probably better or more sophisticated solutions, but I think this one will work. I think you can just give each of them an account on your Debian box, then use popclient (part of netstd) to get their mail from the external server to their Linux account. Then they will be able to use pop to get their mail from your Linux box. I've done this before, but I have sendmail running on all the machines involved. I'm not sure if that's necessary. Perhaps someone else can say one way or the other. -- Rob
Re: xdm problem
Zachary DeAquila <[EMAIL PROTECTED]> writes: > I can't seem to log in through xdm into my newly upgraded from .93 to 1.1 > system. I log in, the screen freaks out as it swaps video modes to > the correct one that I'm running (or is it restarting the X server? > whatever) and then... it comes back to the xdm prompt. Check out the contents of ~/.xsession-errors after an unsuccessful login (use a console). xdm's whole setup changed from .93 to 1.1, so you may need to edit the files in /etc/X11/xdm to bring your stuff up to date. -- Rob
Re: what/where is kernel-package-1.01?
Derek Lee <[EMAIL PROTECTED]> writes: > I saw some messages about kernel-package-1.01 in the debian-user list. > Where is it? I cannot find it in tsx-11.mit.edu or master.debian.org I think it's in debian/experimental. Don't know why. -- Rob
Re: Debian 1.1beta problems
Mark Phillips <[EMAIL PROTECTED]> writes: Not sure about all your problems, but... > * Why do I automatically get an xterm coming up when I start X? It > didn't happen with Slackware. This is the default for debian when you don't have your own ~/.xsession. /etc/X11/Xsession is where this happens. If you have your own ~/.xsession then whatever's there will have control. > * When using dselect to install some software, I got a conflict > between cpp and gcc. This doesn't make sense does it? Surely you > need both? gcc provides cpp, so installing cpp if you already have gcc is redundant, that's what dselect is telling you. cpp is provided as a separate package for systems that don't need to compile, but need the preprocessor for other things. (Some programs use cpp to preprocess their own config files). > * Is it true that dselect automatically updates to newer versions of > packages (if a newer version appears in the directory structure and > you run dselect)? What do you do if you want to reinstall the same > version of a package - can you do it? (for example you might want to > re - set it up) I'm not real familiar with dselect (although I've been meaning to look at it), so it's quite possible dselect will handle this, but you can also always use the lower level command line tool, dpkg, for manual tweaking. See man dpkg. You can install a single package via "dpkg --install foo.deb". -- Rob
Re: loss of routing info with 2.0.0
Rick Hawkins <[EMAIL PROTECTED]> writes: > I have compiled the 2.0.0 kernel (successfully), but lost my routing > info in the process (coming from 1.3.95). "route" only shows the > machine itself. > > If i manually add the missing router & gateway, it works fine, but i > lose this on reboot. I don't know if this is related, but I lost all net connectivity on one of my ethernet connected machines today when I upgraded to 2.0. It turned out that the difference was that the machine had NETMASK="255.255.0.0" instead of NETMASK="255.255.255.0". When I changed this in /etc/init.d/network, the problem went away. -- Rob
Re: linux2.0
Derek Lee <[EMAIL PROTECTED]> writes: > I am afraid I am not familiar with the kernel-package package. > Is this a package to build any kernel version and install it in > a debian way? Bingo. I'm using it as we speak. I'm not sure if it has made it to the unstable tree on all the mirrors yet, but it should soon. When you get it, the docs are /usr/doc/kernel-package, as expected. -- Rob
Re: stty vs setserial
Dale Scheetz <[EMAIL PROTECTED]> writes: > I can point setserial at a particular port, but can't seem to tell it to > set the stop bits to two. > I can tell stty to set the stop bits, but can only seem to use it while on > the tty in question. Is there any way to point stty at the port I want > changed? Unless I'm mistaken, the paradigm is stty options < /dev/tty-whatever Hope that helps. -- Rob
Re: linux2.0
[EMAIL PROTECTED] writes: > Would the release of debian 1.1 use the new stable Linux (v2.0)? I'm not sure if it'll be packaged in time for the initial 1.1 release, but it'll certainly be available shortly thereafter. Anyway, using the new kernel-package package, it's pretty trivial to do it yourself. -- Rob
Re: Packages need libc.so.4
"Larry Loos" <[EMAIL PROTECTED]> writes: > I'm doing my first install of Debian Linux using the 1.1-Beta. I have > the initial system installed and now I'm trying to install the > packages that I want. When I tried I got many errors that say it > needs libc.so.4. Where is it? Where did you get the packages you're installing? Most of the new packages depend on libc5, the ELF libc. It should have been installed when you upgraded to Debian 1.1. If you have old packages which depend on libc4, and you still need to use for some reason, you can get the libc4 (not libc4-dev) package from the development tree. It provides libc4 binary compatibility. -- Rob
Re: pthreads and libc version for debian 1.1
[EMAIL PROTECTED] (David Engel) writes: > > The MIT pthread library 1.60 beta2 seems to work. It passed most of tests > > in the MIT pthread package. The pthread libraries are not installed by [...] > Would someone like to test the pthread library that is built? You'll > need to get the libc5 source and recompile the entire package > yourself. If libpthread proves to work, I'll include it in either the > next libc5 package or a separate libpthread package. We've been using pthreads in the lab for a while, but we had to get 1.60 beta 4 (I think there's an even newer one now) because beta 2 still had a number of bugs that made it unusable. It'd be really nice to have the library included in Debian's libc (or as a separate package). Maintaining a separate version of pthreads on top of Debian's libc5's been a pain. Note that we may need a separate pthreads package since pthreads includes wrappers for gcc/g++ to make sure you compile with the right headers and libraries. Or perhaps that wouldn't be an issue if Debian's libc were already compiled with the support. I'm not sure. -- Rob
Re: Socket destroy delayed (r=0 w=1340)
Gerd Bavendiek <[EMAIL PROTECTED]> writes: > on a quite recently installed Debian-1.1 box I sometimes encounter the > following. > > koko:/root>>> ping us1by-6 > PING us1by-6 (156.53.107.26): 56 data bytes > > --- us1by-6 ping statistics --- > 5 packets transmitted, 0 packets received, 100% packet loss > > Socket destroy delayed (r=0 w=1340) > || > > I never saw this before ! Anything to worry about ? > My netbase is version 2.02-1. We had this on the machines in the lab for a short while. I believe it went away when we changed to a newer kernel version. (I can't remember for sure though, it could have been some newer debian package.) -- Rob
Re: Help with auto-pgp (emacs)
Yves Arrouye <[EMAIL PROTECTED]> writes: > With a default installation (1.1 from scratch), I can't use auto-pgp as > documented. When I do C-x y s to sign, I get the following error: > > /usr/bin/pgp-auto: connect to pass-socket for passphrase: No such file or > directory Did you set the environment variable PGPPATH? Do you have a .pgp directory? See the auto-pgp info pages for details (C-h i or "info auto-pgp" from the command line). It's also important you understand all the warnings associated with auto-pgp detailed there (and of course in the pgp manual). Good luck -- Rob
Re: Tex and metafont
> "D" == Dale Scheetz <[EMAIL PROTECTED]> writes: D> I have been playing with lyx lately, so I thought it would be a D> good idea to install latex (lyx uses it). Well latex needs tex D> which among other things declares a dependance on metafont. I can D> find no metafont package. I assume it is a virtual package? Which D> *real* packages provide metafont? You are looking for the mf* (mfbin, mflib, etc.) packages. -- Rob
Re: Must pppd be run by root?
> "C" == Craig Sanders <[EMAIL PROTECTED]> writes: C> I suppose that would do the job, but what if a sysadmin wants to C> allow users to dial in using ppp, but NOT allow them to dialout C> with minicom or send faxes? C> I'm absolutely certain that I wouldn't want to add users to a C> dialout group just to let them dial in with a ppp account. If you're going to do something, ppp/slip seem sufficiently different to warrant their own group. Something like transient-network-connection (but shorter :> ). I'm still not sure that ppp will really work properly when run by someone other than root (unless it's suid root) if you want to use the "default" option that makes the routing changes automatically when the connection goes up and down. Even if you handle the routing outside the call to pppd, you'll still have to do the routing as root, so I don't see what trying to get pppd to be runnable by non-root users buys you. I still think that sudo's the best solution, but I probably just don't understand the situation well enough. -- Rob
Re: uninstalled packages in 'dpkg -l'
>What can I do about this? dpkg --list 'ncurses*' | egrep -v '^un' Desired=Unknown/Install/Remove/Purge | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ NameVersionDescription +++-===-==- ii ncurses-base1.9.9e-1 Video terminal manipulation: Minimum set of ii ncurses-bin 1.9.9e-1 Video terminal manipulation: associated prog ii ncurses-term1.9.9e-1 Video terminal manipulation: additional term ii ncurses3.0 1.9.9e-1 Video terminal manipulation: shared librarie ii ncurses3.0-dev 1.9.9e-1 Video terminal manipulation: Developer's lib ii ncurses3.0-pic 1.9.9e-1 Video terminal manipulation: Shared-library -- Rob
Re: which..
> "R" == Richard Dansereau <[EMAIL PROTECTED]> writes: R> Is the command "which" available in any of the debian releases? I R> haven't seen it yet in 0.93R6 though I don't have everything R> installed so I may have missed it. I would say that Debian needs a FAQ for this, but after the new release is out it'll be irrelevant. The upcoming 1.1 release has which. It's just a bash shell script that calls bash's built in type command: #!/bin/bash type -path $* -- Rob
Re: 3D Graphics Cards?
> "B" == Bruce Perens <[EMAIL PROTECTED]> writes: B> Is anyone out there using 3D graphics cards with Linux? The only B> one I know about is the Matrox Millenium that works with B> Accellerated X. I need X support and Open GL, and as much B> rendering power as possible. I'm using the Millenium, but I have no idea how to access the 3D capabilities. I have Accelerated X, and the new version claims to support PEX, but I havent delved enough to figure out how to try it out. As you can probably tell, I'm pretty 3D standard naive. If you'd like me to compile and test some 3D code, I'd be more than happy to. I'm kind of curious about the 3D support myself. -- Rob
Re: Must pppd be run by root?
> "R" == Richard Kettlewell <[EMAIL PROTECTED]> writes: R> pppd has to do various messing around creating network interfaces R> and so on, so running it as (not root) is a bit of a non-starter. R> Why do you want it to be able to run it not as root? Right, pppd really needs to be run as root. You don't want just any user killing your net connection. If you want to let a selected set of users bring the net connection up and down, check out sudo. It's designed for this. -- Rob
Re: Checking if the network is up
> "M" == Maarten Boekhold <[EMAIL PROTECTED]> writes: >> I just have my ip-up script (it's a ppp config script, see man >> pppd) touch a file in /usr/local/etc/ppp when I connect, and >> ip-down remove the file when I disconnect. Then I can just check >> for the existence of that file, and know if I'm connected. >> Alternatively, you could M> doesn't work if the connection is dropped by a bad line or such M> things. It does if you also have these set. lcp-echo-interval lcp-echo-failure Actually, I think hardware handshaking should handle the problem, but it doesn't always seem to (grrr). -- Rob
Re: Checking if the network is up
> "e" == eckes <[EMAIL PROTECTED]> writes: e> Actually if your modem is configured correctly (CD drop on Carrier e> loss &C1), and if the pppd is given the option "modem" the Process e> will hangup if the line goes down. This is not the actual hardware e> handshake, but on of the RS232 signals. Right, I have this set up, but I still got somewhat flaky behavior. I haven't tested it again recently, but I think it might be my modem. I have one of the internal Sportsters (an older one of the 28.8 flavor), and I had heard some rumbling about problems with those. I'm going to get a modem ROM upgrade soon. Maybe that'll fix it. Thanks -- Rob
Re: 1.1 X setup default: bs/del don't work with motif apps.
> "G" == Gerry Jensen <[EMAIL PROTECTED]> writes: G> It's interesting that people are having problems now with Motif G> apps, because suddenly, my problems with Motif apps have been G> magically solved. That is, the backspace key now works correctly G> for both Motif and non-Motif apps. Previously, if I wanted the G> backspace key to work correctly with Motif apps, I would have to G> map it to backspace (where the default was mapped to delete). But, I've always had some magic in my Xdefaults for Motif. *XmText.translations: #override\n\ osfDelete: delete-previous-character() *XmTextField.translations: #override\n\ osfDelete: delete-previous-character() Without this, things were kind of screwy. I've never investigated what it does or if I still need it. Someone just told me when I was getting started with Linux, "You need this in your Xdefaults for Motif." so I added it and went on. -- Rob
Re: 1.1 X setup default: bs/del don't work with motif apps.
> "K" == "Keith Beattie[SFSU Student]" writes: K> I don't think there is a .deb package for it but it compiled "out K> of the box" for me. FYI: Actually, there is a debian package for xkeycaps. -- Rob
Re: Checking if the network is up
> "L" == Luis Francisco Gonzalez <[EMAIL PROTECTED]> writes: L> Is there any way of checking whether the network is responding so L> that I only run popclient when it is? I just have my ip-up script (it's a ppp config script, see man pppd) touch a file in /usr/local/etc/ppp when I connect, and ip-down remove the file when I disconnect. Then I can just check for the existence of that file, and know if I'm connected. Alternatively, you could check the output of "netstat -r". -- Rob
Re: upgrading from 0.93R6 to 1.1 beta
> "I" == Ian Jackson <[EMAIL PROTECTED]> writes: I> I don't know if dftp (or dpkg-ftp for that matter) support I> dependencies and/or pre-dependencies (the latter are required for a I> smooth upgrade to the ELF system). The current dftp doesn't (I haven't used dpkg-ftp). Currently dftp's better for incremental upgrades (or getting all the new packages so you can do it by hand with dpkg), rather than a complete overhaul. If dpkg supported something like dpkg --dependencies foo.deb (does it?) then it would be fairly easy to handle the simple cases of package install reordering. -- Rob
Re: ncurses problems with xterm/rxvt
> "C" == Carlos Carvalho <[EMAIL PROTECTED]> writes: C> One is already mentioned here, that after a program that uses C> ncurses is called, lines longer than the width of the xterm/rxvt no C> longer wrap. This happens with nvi, but also with ncftp. Is there a C> character sequence that can be sent to restore the wrapping? This is a known problem, and in fact has been fixed (I was so happy). Unfortunately the new version of ncurses that fixes this has not made it to the public ftp sites. In the interim you can do the following (probably dumb and overkill, but it works): reset vt100 reset xterm clear C> The second problem is that when you're going to search for C> something inside less, and you want to erase some characters, the C> backspace just prints ^H instead of erasing. The only way out is to C> type ^U and retype everything. I think that's also fixed. In the interim try using an actual Ctrl-H key sequence, or the "delete" key (under "insert") instead of backspace. That might work. -- Rob
Re: Verifying that download was successful?
> "M" == Martin Rheumer <[EMAIL PROTECTED]> writes: M> Rob, Any hints on where I would find dftp Check the contrib directory on your favorite Debian ftp site. It's /pub/Linux/Debian/contrib/tools on the site I use. Contrib is at the same level as the stable, unstable, etc. directories. I see that version 1.4's there now. It had a couple of bugs. The new 1.5 version has been released, but apparently hasn't migrated to its proper directory yet. Brian (the maintainer) put an extra copy somewhere where people could get to it in the interim, but I can't remember where that was. I cc'ed this to him, so if he's listening, he can remind us :> M> I have just mirrored the unstable 1.1 release and want to now use M> dftp, but can't find the correct one... One thing I'll point out is that you don't need to mirror the whole release to use dftp. It'll get the pieces you need for you. -- Rob
Re: How to remove a user?
> "A" == Andreas Wehler <[EMAIL PROTECTED]> writes: A> After searching a while for a program or script to completely A> remove a user I didn't found something and did it per hand. Is A> this neccessary? Thanks. Right now, this is the only way to handle it. You need to remove their /etc/passwd, and /etc/group entries, but first, you should probably search the filesystem for any files they own and do something with them. Then, as a final step, you way want to remove their home directory. You can find all a user's files by using find / -user username -or -group usergroup You'll need to run this as root so that you don't get blocked from directories without read permission. I haven't actually done this recently, so this may not cover everything. -- Rob
Re: uudecode?
> "F" == Fundamental <[EMAIL PROTECTED]> writes: F> what package does uudecode come in? sharutils -- Rob
Re: Verifying that download was successful?
> "M" == Mark Phillips <[EMAIL PROTECTED]> writes: M> Hi, I am downloading debian 1.1 beta using ftp. I want to check M> that the files have been downloaded correctly. One way of doing it M> would be to manually do a md5sum on each file and check it against M> the "Packages" file, but this would take a long time. Is there a M> quicker way of doing it? dftp will do this automatically via "dftp verify". -- Rob
