LaTeX, Texlive-luatex, Biber installation procedure
I am about to install LaTeX on debian squeeze, and I'm drawn to luatex and biber. Much has changed over the years, and an effort to be clear about installation procedure has left some issues unclear. Am I correct to assume that if I install texlive, texlive-luatex, and CTAN biber in that order, I should end up with a fully functional TeX Live 2010, with LaTeX2e macros and biblatex support? I run into discussions about setting up some PATHs to support tlmgr. That's easily done, but I don't want to do it if an installation of texlive-luatex sets up the needed paths automatically. Does it? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87bp6n8xmp@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
Allan, thanks for the input. I'm about to change my location, and will not be in a position to take any corrective action, but your points I'll keep in mind when I return. Alan Chandler writes: > I can give my domain an ip-address, and I can also give subdomains > such as www.hartley-consultants.com and mail.hartley-consultants.com An ability to assign an IP address was news to me. Tomorrow I'll be in touch with my web hosting service and will ask them about it. They are Linux-friendly, but my ISP is a hopeless case. > If you look up these two names with anyone of several commands - such as... > > host www.hartley-consultants.com and > > host mail.harlley-consultants.com I get the point, but for ssh the target is not a mail account or web site, but a host. If you could assign an IP to chandlerfamily.org.uk it would seem to solve the problem. In other words, how does a nameserver know to send income u...@domain to a mail server rather than web server? > if I ssh to mail.hartley-consultants.com I should get to where my > mail is sent. When you say, "where my mail is sent", it seems you mean the local host. > Inside my home network I have a number of machines on a lan. I have > happened to have chosen not to call those some subdomain of > chandlerfamily.org.uk (although at one point in the past I have done > it that way, and called one of my machines > kanga.chandlerfamily.org.uk, and gave it an ip-address 192.168.0.12). Not sure I follow. Don't all hosts on your LAN have to have a local domain name such as kanga? On your LAN, don't all your hosts have their own local address, such as 192.168.1.1, 192.168.1.2, etc.? > The problem comes when you want to make INCOMING connections to > kanga.chandlerfamily.org.uk. The program trying to make that > connection looks it up and gets 192.168.0.12 and it doesn't know where > to send this to (its unroutable). Instead, I would ssh to > home.chandlerfamily.org.uk and (if it was a router rather than the > linux box that it is) I would forward the port to 192.168.0.12 - this > is known as reverse NAT. But I can only set the router to do it to > one of my machines per port. In this case I would use port 22 and > forward it on. I'll think about this. If the incoming signal comes to your router, the router sends it to the LAN (192.168.0.0) and the kanga host picks it up. I suppose you are attacking the port to your ssh command. That is, are you forcing delivery to the usual ssh port of 22? Thanks for your input. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87d3si9w3a@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
Morgan Gangwere <0.fracta...@gmail.com> writes: > on Sat, 11 Sep 2010 15:38:04 -0400, brownh > <871v90ax5v@teufel.historicalmaterialism.info> attacked their > terminal with [snip] > > Random Blithering Curiosity... Is the gateway a NAPT? NAT loopback is not enabled on my router. Not sure this answers your question. > I had this problem for a while where I would be fine on the inside, then as > soon as I went to the outside, I'd get the same problem you're experiencing. My googling suggests a lot of people have the problem, but the reasons seem different. > Checking the following fixed my problems: > > 1) don't ssh as root unless you /have/ to. > 2) Check that your NA(P)T allows port 22 on TCP *and* UDP, incoming and > outgoing. > 3) try using an SSH key. This occasionally fixes things. I don't ssh as root; NAT is disabled in my router. As for SSH key, I'll give that a try, but too little time before I fly out of here. > I have a Lenny box sitting on my desk that I SSH to all the time > with the default configuration. Nothing special, just the default > OpenSSH-server configuration. Good to know. I've mailed a query to the ssh list. > For those who were confused: > The setup as it looks like to me is this: > > (Internets)[Gateway/router]--,---[server] > `--[laptop] > this configuration /works/. > [laptop]---(??)--(internets)--[gateway]---[server] > This does not. Now I am confused ;-(. What I'm trying to do is: laptop client -> internet/nameserver -> router -> server on LAN Your first line looks like what I have now; the second line looks like where I'm trying to get. > This leads me to beg that there is either a configuration issue that > says "Dont let anyone who's not on my local network talk to me" or a > configuration issue with port translation. The first problem sounds like it would involve my router, but I've enabled ssh services in it. I don't know if I should enable NAT (my rough impression is that NAT is best avoided). I should think my name server would be able to use port info to send signals to the right place, but I know nothing about it this "translation". Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87hbhva9js@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
David Jardine writes: > I'm afraid I've forgotten - or didn't read - earlier details. Have > you got a public IP address? My router does. It seems it has a dynamic address assigned to it by my ISP. I guess this is what you mean by a public IP address. > Can you ping the domain name from outside your network - from an > Internet cafe or a friend's machine? Have you set up a web server? > Can you access that from outside? Does your IP address point to > your ISP's server or to yours? No way to get readily to an internet cafe or friend's machine, but I find that when I try from my machines on my local router, I get: PING historicalmaterialism.info (216.239.138.216) 56(84) bytes of data. From ge-0.fw-1.torrance.omnis.com (216.239.131.9) icmp_seq=1 Destination Port Unreachable This is the first time I've had this happen. Perhaps my problem would be resolved if I assigned the standard SSH port number (22), but when I try append ::22, the address cannot be resolved. I have not set up a web server, but only have a web hosting account other than with my ISP. I can ftp to this account. The IP address of my router is assigned by my ISP's server. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87lj77aa5l@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
David Jardine writes: >> > Connection closed by 216.239.138.216 > > Haines, I have the feeling you've got this all wrong. Your site is > being hosted and any connection to it goes to the webhoster's site. > All your local hosts are unconnected to this site. You are trying > to log into your site on the webhoster's server. What you say makes sense. The IP address turns out to be my web hosting service's name server. Apparently SSH first reaches this DNS to convert the domain to an IP address. > If you want to log into hosts on your local LAN, you need a dynamic > DNS service (dyndns.com is the best known), which associates an IP > address with your domain name while you are online. Or upload all > your files to the webhoster's server and make sure he allows you > ssh access. All this is new to me, and I'll not be able to implement anything like that before I must shortly leave. But the whole situation strikes me as odd. Surely innumerable people rely on a web hosting service to provide email services as well, and yet it would seem none of them can use ssh to acccess a home or office machine remotely without going through extraordinary gymnastics regarding their IP address. An obvious step would be for me to find out what my hosting service has to say. If I get a simple answer, I'll report it back to this thread. Thanks. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87pqwjafyg@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
Wow, I certainly managed to be obscure! I tried to help David out a bit, and here respond to your uncertainty. Claudius Hubig writes: > I’m sorry I have to ask again. You are trying to connect to a host > "historicalmaterialism.info" and login as user "haines"? Why and how > do you differentiate between client hosts and server hosts and most > importantly > >>the client user account (bro...@historicalmaterialism.info) >>to client user account (hai...@historicalmaterialism.info) > > that mean? Are you trying to connect to "historicalmaterialism.info" > from that very host, with your current account being "brownh" and you > want to login as "haines"? At present, I have three or four hosts on a LAN, and I can ssh from each one to the others. The problem arises because of what I want to do in the immediate future. One of the hosts on the LAN is a laptop, and I'll be taking it into the field and would like to ssh from it via Internet using someone else's access provider. Simply, I want to access files on the machine left running at home on the LAN from a laptop I take into the field. I assumed that a test of this would be to ssh from one machine on my LAN to another on the LAN via Internet by providing an address to ssh in the form $ ssh acco...@domain. These machines all have the domain name historicalMaterialism.info (although I have other domains I could use as well). Each machine has its own local domain name. For example, my current machine has an account brownh on teufel.historicalMaterialism.info. My laptop has an account brownh on langhans.historicalMaterialism.info. The machine I've used to illstrate my problem has an account haines on eng...@historicalmaterialism.info. (You are too young to know who Fritz Teufel and Rainer Langhans were). I've been trying to ssh via Internet from the client account brownh on teufel.historicalMaterialialism.info to the server account haines on engels.historicalMaterialism.info. What I don't understand is why my router, which has SSH service enabled, does not simply redirect a ssh request from one machine on the LAN to another machine having the same domain name directly rather than send it off to my provider. Maybe I've invented by problem by using an impossible situation as my test case. > Also: > >> $ ssh -vv hai...@historicalmaterialism.info > > You tell ssh to log in as user haines on the host > historicalmaterialism.info here. Yes I do in the sense that user brownh on one machine seeks to ssh to account haines on another. Is that incorrect? When I try to log in, I'm given the prompt: "hai...@historicalmaterialism.info's password:", and so I type in the password for the haines account on that other (server) machine. >> ... >> debug1: Next authentication method: password >> eng...@historicalmaterialism.info's password: > > But here, your username is suddenly engels – why? I can't reproduce this (result of an experiment?). What I get is: debug1: Next authentication method: password hai...@historicalmaterialism.info's password: debug2: we sent a password packet, wait for reply debug1: Authentications that can continue: publickey,password Permission denied, please try again. hai...@historicalmaterialism.info's password: debug2: we sent a password packet, wait for reply Connection closed by 216.239.138.216 Interesting. This IP address is one owned by my Interet provider. So my ssh request did enter the Internet, answering one of my concerns. Although my provider closed the connection (timed out or too many password attempts?), I assume that the issue of the specific password is soley the concern of the two hosts being connected and does not involve either my router or access provider. >>I did not edit the ssh_config files on client or server, and what >>follows are the defaults. I gather, then, that you see no problems with the configuration. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87wrqs9fl5@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
Sorry that I didn't make myself clear. I've got two situations: a) my present sitution in which I communicate between hosts on my local LAN, b) a future situation (to which I'd like to arrive in a day or so) of taking a laptop into the field and using ssh to access a home-base host on the LAN. In the former case, to ssh one I only need a user account and the user's password; the the latter case, I have to access the LAN from an internet server, and this requires both a user account, its password, and its domain, so that the server knows where to find the LAN on which the server host is located. I belabor this a bit lest I'm misunderstanding or missing something fundamental. For example, my routing table on both machines has the router as their gateway. Presently my /etc/revolv.conf files simply point to my router-modem-hub as nameserver. This router-hub-modem has SSH service enabled. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/871v90ax5v@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
David Jardine writes: > That ALL: LOCAL entry is there by default. I don't know much about > this myself, but That would explain its presence on my lenny box, but my newly installed sqeeze box has nothing uncommented in that file. So I guess squeeze changed the default. >From the manual, ALL:LOCAL in hosts.allow would admit connections with local user accounts, and I get the sense that it overrides the default ALL:ALL, which would exclude access from hosts not on the LOCAL LAN. I don't expect to use this box as a server for hosts outside my LAN. However, I don't see this as being relevant to my problem, for it is not my (client) lenny machine that has a problem, but when lenny tries to gain ssh access to the server, squeeze, and for this server, the hosts.allow file is empty and so is open to the world. I did comment the ALL:LOCAL on the lenny box, and tried to access it by ssh from the squeeze box (the opposite direction to that I raised in my question), but permission denied here as well. So my problem is broader than just the configuration of one box. Just to be sure, a sanity check: from the account bro...@historicalmaterialism.info on one box I am trying to ssh to the account hai...@historicalmaterialism.info, on a different box on the LAN. I can ssh from one user account to another on the LAN, but can't ssh between them via my provider's server by appending the domain name to the user account name: $ ssh hai...@historicalmaterialism.info Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8762ycb4ku@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
David Jardine writes: > Have you got /etc/hosts.allow and /etc/hosts.deny configured to allow > access from outside your local network? David, good question. I had understood /etc/hosts.allow only as a way to define a selection, and so left it empty for the server, for it should allow any host to access. However, I now discover that I have ALL: LOCAL on the client machine. No idea why I put it there. I fear that this might in fact prevent the kind of non-local ssh connections that I'm having difficulties with. Is that so? Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87aanobcsz@teufel.historicalmaterialism.info
Re: SSH: remote login returns "invalid user"
Claudius, thank you for troubling with my problem. The problem, again: I have no trouble logging a client host (bro...@teufel) [you are not old enough to remember Fritz and Rainer] with a server host (hai...@engels) over the LAN, but not over the Internet, the client user account (bro...@historicalmaterialism.info) to client user account (hai...@historicalmaterialism.info) $ ssh -vv hai...@historicalmaterialism.info ... debug1: Next authentication method: password eng...@historicalmaterialism.info's password: debug2: we sent a password packet, wait for reply debug1: Authentications that can continue: publickey,password Permission denied, please try again. I don't know if this means a file permission problem or if the password was unacceptable. The password I provided was definitely that of the account (haines) on the client machine (engels). Here is ~/.ssh/ssh_config on client: ForwardAgent yes X11Forwarding yes I did not edit the ssh_config files on client or server, and what follows are the defaults. Here is /etc/ssd/sshd_config file, via ssh, from server (I prune commented lines): # Package generated configuration file Port 22 Protocol 2 HostKey /etc/ssh/ssh_host_rsa_key HostKey /etc/ssh/ssh_host_dsa_key UsePrivilegeSeparation yes KeyRegenerationInterval 3600 ServerKeyBits 768 SyslogFacility AUTH LogLevel INFO LoginGraceTime 120 PermitRootLogin yes StrictModes yes RSAAuthentication yes PubkeyAuthentication yes IgnoreRhosts yes RhostsRSAAuthentication no HostbasedAuthentication no PermitEmptyPasswords no ChallengeResponseAuthentication no X11Forwarding yes X11DisplayOffset 10 PrintMotd no PrintLastLog yes TCPKeepAlive yes AcceptEnv LANG LC_* Subsystem sftp /usr/lib/openssh/sftp-server UsePAM yes Here is /etc/ssh/sshd_config on client: # Package generated configuration file Port 22 Protocol 2 HostKey /etc/ssh/ssh_host_rsa_key HostKey /etc/ssh/ssh_host_dsa_key UsePrivilegeSeparation yes KeyRegenerationInterval 3600 ServerKeyBits 768 SyslogFacility AUTH LogLevel INFO LoginGraceTime 120 PermitRootLogin yes StrictModes yes RSAAuthentication yes PubkeyAuthentication yes IgnoreRhosts yes RhostsRSAAuthentication no HostbasedAuthentication no PermitEmptyPasswords no ChallengeResponseAuthentication no X11Forwarding yes X11DisplayOffset 10 PrintMotd no PrintLastLog yes TCPKeepAlive yes AcceptEnv LANG LC_* Subsystem sftp /usr/lib/openssh/sftp-server UsePAM yes Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87iq2cbkeu@teufel.historicalmaterialism.info
SSH: remote login returns "invalid user"
Sorry for a FAQ, but I searched around without sucess. I'm runing debian a local LAN and can ssh between the hosts without a problem. However, I intend to use one host remotely (a laptop that I want to use to access my home machine). I do not have password encryption set up. My LAN is three machines connected to a common hub using the same domain name. I ssh successfully between local hosts on my LAN in this fashion: $ ssh -X acco...@localhostname To prepare for taking my laptop into the field, I try this to access my home machine (server): # ssh -X acco...@domainname In the server's /var/log/auth.log I get: Sep 10 13:04:37 engels sshd[27266]: Failed none for invalid user brownh from 192.168.1.4 port 33279 ssh2 Here the password is "none", which suggests to me that is is not getting the password, although it is typed in at the Password: prompt. Then the user brownh (my account on the client host which is initiating the ssh connection) is not valid, although the its IP address is right. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87mxrpbdvt@teufel.historicalmaterialism.info
Re: authentication of multiple users for one mail server
Camaleon, Thanks for your patience, and I seem to have stumbled on my problem: emim4 configuration. Regarding the value for "system mail name", in retrospect it does makes sense, but not when I was reading the document you cited. First, if I can reconstruct my thinking correctly, I failed to associate the word "mailname," which was not explicitly defined, with "domain". If I asked someone on the street what their "mailname" is, they would give me their email address. I knew that it actually meant domain, but I had to figure out from its function that it was not really the "system" name (i.e., the domain of a host on a LAN), but the alternative meaning of domain, which unfortunately has no dedicated term such as "internet domain". I may be dense and my reading perverse, but it's folks like me that manuals are for ;-) In any case, the source of my problem was my adding my domain name to the list of recipient domains. Again, in retrospect, I understand why this caused exim4 to fail to route a message to another machine that happened to have the same domain name before the LAN was set up. Exim4 naturally looks for the user's of that domain name on its own host or local network and, not finding the user, does not by default ultimately send the message to my provider's mail server. Instead, it simply gives up and returns a routing error message. I guess the assumption is that if a person assigns a local domain name to the host, they intend to make it part of a LAN, and so instinctively know that until such a LAN is actually in place, the local host is not able to send a message to another host having the same domain name. My problem was that I tried to set up and test email before constructing the LAN. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87tylxbw6f@teufel.historicalmaterialism.info
Re: authentication of multiple users for one mail server
Sorry to follow up on my own message, but did some more thinking about the problem. I'm asking Exim4 to send a message to another user having the same domain name. So does this mean exim searches for that user locally rather than ship the message off to my provider's mail server? In my previous message, I went though exim configuration, where at certain points I changed the local domain to the domain. Was this a mistake in one or two cases. A local domain in one or two configuration questions might tell exim that the message is not for a user on its own machine or network, and so it would be routed to my provider's mail server? Just a thought. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87y6babpbz@teufel.historicalmaterialism.info
Re: authentication of multiple users for one mail server
Camaleón writes: > On Wed, 08 Sep 2010 16:52:23 -0400, brownh wrote: > > (...) > >> When it comes to digging into exim4's configuration files, it all goes >> over my head. But I gather from googling that exim4's default is to >> route to only local mailboxes or domains. If that were do, I'd get >> nothing out. In any case, I was unable to locate the variable and value >> DCconfig_internet=1 in all the exim4 configuration files. > > I would take a look into Exim4 manual: > > 2.1. The Configuration System > http://pkg-exim4.alioth.debian.org/README/README.Debian.html#id280581 > > And also "man update-exim4.conf", more precisely to > "dc_eximconfig_configtype" variable. > > Greetings, > > -- > Camaleón Thanks, Camaleón, but no luck. I carefully reread the Debian debconf questions, and yet questions remain. In the dpkg-reconfigure section one question asks whether to hide local mail names. What shows up by default is my _local_ domain name. I gather this hiding refers to what appears in the message envelop, and so my intuition is that I would not want the local domain name there. I changed the default to just the domain name. For example, my current machine is teufel.historicalMaterialism.info, and I changed this to historicalMaterialism.info. Was this a mistake? Likewise, in an earlier question, I provide the system mail name. This is the domain name used to construct addresses. However, here again it the default entry was my _local_ domain name, but since I don't want to have my addresses show the local domain name, I changed the default local domain name to domain name. Again, I hope I did the right thing. On the page for the IP addresses for which exim4 should listed, there has traditionally appeared "127.0.0.1". With squeeze, however, it shows up as 127.0.0.1 ; ::1 without explanation. What is this added "::1" IP address? Or is it garbage that I should cut out? I don't know if any of this has any relation to the SMTP authentication issue. I understand that it is best to rely on TLS authentication, which I'll try in the future, but this issue has already interferred too much with important work, so must leave put off the big challenge of TLS for the next time and cross my fingers in the meantime. By default, exim4 employs TLS authentication, and so disables AUTH PLAIN and AUTH LOGIN. But it does not say what these entities are. Variables given values somewhere? Apparently not. So what are they? I can only guess the first has to do with plain text authentication with the SMTP mailserver, but what is the second? My fetchmail must logs into the mail server, but apparently not exim4. Now, from a reading of the documentation, it appears that for any host that is authenticated in plain text, I need to add the authenticating information into /etc/exim4/passwd.client. That much is clear, but then it says, "If you need to enable AUTH PLAIN or AUTH LOGIN...". Do I need to do so simply because I'm using plain text authentication? Do I need to do one or the other but not both? Since these entities were not defined, I don't know if I need to do it, and further, I'm not told what to do and where to do it. Someone suggested I needed to change value of DCconfig, and you suggested to change the value of dc_eximconfig_configtype, but neither variable, if that's what they are, appear in the configuration files for me to change. So I spent some time googling, and apparently the proper variable is AUTH_CLIENT_ALLOW_NOTLS_PASSWORDS. I had to google yet more to find that the value for this variable should be defined as " = true". But further googling provided only a hint that it should be placed at the top of the /etc/exim4/conf.d/main/03_exim4-config_tlsoptions file. So I tried it, and it didn't work. A messge from my new installation of squeeze on one machine (us...@historicalmaterialism.info) cannot be sent to the provider mail server and delivered from there to another machine which has the us...@historicalmaterialism.info address. Exim4 simply complains I can't route the outgoing message. At this point, I can only communicate from the first machine to the second by sending messages to it to a couple other domain names that the second machine happens to listen to. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8739tid57y@teufel.historicalmaterialism.info
Re: authentication of multiple users for one mail server
I've made some progress simply by defining a set of users+domains on my host mail server. Broadly, now I've got three machines all speaking with each other. However, while the machine running squeeze I'm trying to setup can communcate through an alias on another server and to various addresses, it can't send mail to my current machine to its primary address: exim's error is that it is unroutable. However, I would assume that for an administrator handling dozens or hundreds of accounts which need access to a mail server, the administrator does not create all these accounts in the mail server nor fill up exim4 configuration files with all the information. It seems more likely there would be a simple alias list to grant user account access to the mail server. As for my present situation, from within the account named "haines" on the new machine, I can send mail to old machine, bro...@historicalmaterialism through a forwarding service on another server and I can send mail to my current machine using an alternative domain name for it, but can't send a message directly to bro...@historicalmaterialism.info. When it comes to digging into exim4's configuration files, it all goes over my head. But I gather from googling that exim4's default is to route to only local mailboxes or domains. If that were do, I'd get nothing out. In any case, I was unable to locate the variable and value DCconfig_internet=1 in all the exim4 configuration files. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/877hiwc60o@teufel.historicalmaterialism.info
Re: authentication of multiple users for one mail server
Celejar, sorry, I thought I was asking just a generic question. Yes, I'm running exim4 under debian squeeze, with SMTP authentication required, and I do have an entry in my /etc/exim4/passwd.client file, and my problem probably has to do with the syntax of the entries. For years I've used a wild card for the mail server, entered my UID for the mail server, which is my email address, and finally the password for this account: *:bro...@historicalmaterialism.info:password Now I have a second box with a different user account name (which I'll call user1 here), which wants to be authenticated by this same mail server. On this second box, exim4 creates the From: line by attaching the user account name (user1) to the domain name: us...@historicalmaterialism.info. The aim is to have this sender address authenticated. I try various things in passwd.client: Outgoing mail accesses my mail server account with a UID which is my email address (br...@historicalmaterialism.info) and a password (password). So, for user1 on the new machine to use this email account, I put in the line: *:bro...@historicalmaterialism.info:password This does not work because the From: line of the outgoing message is us...@historicalmaterialism.info, and so the server wants to see if this user1 is legit. So I add this line: *:us...@historicalmaterialism.info:password But this is not the UID of the account name, and so does not work. In desperation I also tried a wild card to cover all users: *:*...@historicalmaterialism.info:password In any ase, when user1 attempts to send a message, it immediately bounces with this error message: recipi...@address.com SMTP error from remote mail server after RCPT TO: : host mail.historicalMaterialism.info [216,239...]: 553.5.7.1 : Sender address rejected: not onwed by user bro...@historicalmaterialism.info I took this to mean that the error message cames from my ISP's mail server even though it seems instantaneous. "mail.historical.materialism" is the mail account on that server. I assume that the account on the mail server with UID bro...@historicalmaterialism.info could not authenticate mail from us...@historicalmaterialism.info. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87y6bc77ui@teufel.historicalmaterialism.info
authentication of multiple users for one mail server
I've never encountered this problem before because I've always used the same user name, but now I'm setting up a machine with different user accounts and I need to have all these users' outgoing mail authenticated by the mail server. I run exim4, but not procmail. Although the error message says it is from the remote mail server, it shows up instantly, and I suspect the error is generated locally. I've struggled with the Exim book, but it goes over my head. I tried listing users in /etc/exim4/passwd.client, but that was a bad guess. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8762yg9a80@teufel.historicalmaterialism.info
Re: .xsession kills X server with sqeeze/fluxbox
Javier Vasquez writes: > Can you try the following? As you already have ~/.xsession, make > sure it has execute permissions. Then make a link to it naming such > link ~/.xinitrc. Once you have that, if it's empty, then I would > expect X to just die, since there's NO window manager specified. > At the end of ~/.xsession, you now need to specify the window manger > you want to run. In my case startfluxbox, you could call fluxbox > directly as well, or any other thing such as startkde, > gnome-session, or any other thing... Yes, now fluxbox starts OK with the .xsession file command. In retrospect, it all makes sense. Startx looks for .xsession, and finding it executes the window manager specified in it, but not finding the file executes the default window manager. However, if there is an .xsession file, but without any exec windowmanager statement at its end, the X server falls through the cracks. I had assumed it looked for a statement in the file, not for the file itself I now removed the executable permissions from .xsession, killed the symlink, and it still works. I'm not clear why you asked me to take these two steps unless it merely served to cover other possibilities for the test. Looks like I'm now all set. Thanks! Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87aanu92fy@teufel.historicalmaterialism.info
Re: .xsession kills X server with sqeeze/fluxbox
Javier Vasquez writes:
> It sounds to me it's not a ~/.xsession issue. Have you tried just
> loading X without window manager at all, to see if it dies? If so it
> might be a xorg.conf problem or some other configuration...
Something just occured to me. Please excuse a bit of senility and the
distraction of other pressing work, but I now believe I recall
having initially created an .xsession file and putting into it these
statements:
xset s 3600
xset -dpms
When I found the time too short, I added a 0, and it did take longer
to blank, but then seemed to revert to a shorter time by itself. So,
if my memory serves, the .xsession file didn't initially kill the
xserver, but my fidding with xset statements may have done
damange. When I lost the X sessions altogether, I removed those
lines. I now try to restore them in .xsession, but it doesn't help.
Going back to elaborate the initial scenario, when I first installed
squeeze, I de-selected the "Graphical Desktop Environment" bundle of
packages, for I startx from command line and never install a graphical
desktop environment such as KDE. Then, After some preliminaries, I
installed xorg and tested it by doing $ startx, and got the expected
graphical mode (blank screen with graphical prompt (?) in UL corner).
I then installed fluxbox, and much to my surprise, it was started
automatically by the X server call. That is, for years I've started
the window manager by putting a statement in .xsession such as:
exec fluxbox || exec xterm -fg red
So I no longer know how fluxbox gets called and how to stop it to
carry out your test. I looked into /etc/X11/ and off-hand couldn't
find a call for fluxbox, although there's a fluxbox directory to
define its default behavior. So I no longer know how to stop fluxbox
except by terminating its process, which will do me no good because
fluxbox is recalled when I start the X server.
I did check root account behavior: I logged in as root and for first
time started X for it. I found again that fluxbox starts automatically
when I startx. I created an empty .xsession file, and when that file
exists, the X server gives up the ghost without any suicide note. That
is, each time I try to start X and it fails because of the presence of
the .xsession file, it only shows up in .xsession-errors as:
Xsession: X session started for ...
However, there nothing like:
$ ps aux | grep X
brownh 3928 0.0 0.0 2864 808 tty1 S+ Sep02 0:00 xinit
/etc/X11/xinit/xinitrc -- /etc/X11/xinit/xserverrc :0 -auth
/tmp/serverauth.bkLgHxoJbD
so apparently the server is not running despite what Xsession says.
There's nothing relevant in user.log or syslog. I did an call trace:
$ strace -o xsession-trace startx
The enormous result was beyond my ken, but for whatever its worth, it
ends like this:
...
stat64("/bin/rm", ...) = 0
clone(child_stack=0 ...) = 3303
wait4(-1 ...) = 3303
exit_group(0) = ?
Ps-aux does not say there's any process 3303 still running.
Haines
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/87lj7f81g9@teufel.historicalmaterialism.info
.xsession kills X server with sqeeze/fluxbox
I'm used to using an .xsession file in ~/. But I find that when I have sqeeze installed with fluxbox, the server crashes if there is an xsession file present, even if it is empty. The .xsession-errors log offers no clue. It does fail to open .Xmodmap, but I don't have this file because I don't remap my keyboard. Then there is a statement: Xmodmap: ... aborting, which I assume does mean the entire X server is aborting. The log also said it had failed to read certain values in ~/fluxbox/init, and was using the default values. But nothing about the X server aborting. Is .xsession still allowed in sqeeze Xorg? If so, how can an empty file bring the X server to its knees? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87pqws822z@teufel.historicalmaterialism.info
Upgrade of squeeze hangs on apt (0.7.26~exp3)
With a brand new installation of squeeze, one of my first steps was to do a package update and a safe-upgrade. The former went well, but when I tried to upgrade the 22 updated packages, the upgrade proceedure terminated with: Reading changelogs ... done apt (0.7.26~exp3) experimental; urgency=low * apt-ftparchive now reads the standard configuration files in /etc/apt/apt.conf and /etc/apt/apt.conf.d and after some more lines ends with /tmp/tmp (END) and it takes a Ctl-C to break back to a command prompt. The upgrade process is terminated and nothing was actually upgraded. I can repeat, and each time only the tmp file id number changes, but the 22 packages are still not upgraded. A google search leaves me unclear, but it seems that some experimental features of apt are being merged into squeeze. Should I just ignore this until such time as a clean upgrade becomes possible? Is there a work-around? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87tym48l1e@teufel.historicalmaterialism.info
Re: Creating a bootable USB key
Jimmy, Just to close the thread and to thank you for your help, I ultimately discovered that my inability to get a netinst or netbook installation from a USB-key to work was not due to an error on the key. It seems a hardware error (new box). That is, nothing boots except occasionally by accident. In one of those accidents, I was able to complete the installation of the operating system, but haven't been able to boot it since, and only once did a Debian Live on USB-key that worked well on another machine boot. BIOS seems to ignore the boot priority and try to boot what it wants. I've already lost too many days of work, and so will return to the problem of USB or BIOS next month. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8739u54ksk@teufel.historicalmaterialism.info
Sqeeze install: partition table not written
I'm restarting a query regarding a grub2 error 15 when trying to do a a boot after a netboot install of Debian squeeze from a USB-key and a boot hang when I do a netinst install from the USB-key. Thinking I needed to fix GRUB2, I booted debianLive from a usb-key, and found that I only have one partition on my HD: /dev/sda1, which is mounted on /boot and is marked bootable. Its content is 628 Mb and has the expected files in it, which suggests that the content for this partion was properly installed. However, # fdisk -l says this is the only partition, althought during installation I had created 14. # ls /dev | grep sda* returns only sda and sda1. The partitioning during installation seemed to go smoothly, as did the installation of files to those partitions. When I do # mount, I find that /dev/sda1 is mounted in /fll/sda1, aufs is mounted on /fll/aufs, and aufs is mounted on the various mount points I had created. # ls reports all these mount points (/boot, /root, /home, etc). When I do # df, I see aufs entries are mounted on the various mount points created during partitioning. No idea what this means. It seems all my partitions were merged during installation and not unmerged from the aufs. What would be a fix? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/877hjk3rwq@teufel.historicalmaterialism.info
Re: squeeze install: Gave up waiting for root devices
I found out by trial and error that, although directions don't mention making the USB-key bootable, I must make its first partition bootable. Now I can boot the installer. However, when it looks for the ISO, rather than look for it on the USB-key (/dev/sdb) (or at least I didn't catch it), it looks for an ISO in the partitions of my target drive (/dev/sda). Not finding any, the installation fails with message, "No kernel modules were found". The most obvious problem would be that my installer and kernel versions are incompatible. I rechecked, and made sure I got my debian 6.0 alpha1 installer from http://www.debian.org/devel/debian-installer/ and my initrd.gz and vmlinuz files from http://ftp.nl.debian.org/debian/dists/testing/main/installer-i386/current/images/hd-media/ . So I do have properly matched versions. The problem seems to be that the installer looks for the ISO on my target (ext3) HD instead of my (FAT) USB-key. While the failure message is being displayed, I do Alt-F2 to go to a bash shell and do a ls, it lists the directories on my target HD, sda rather than the USB-key, sdb. Shouldn't the shell be looking at the USB-key? When I issue the mount command, I find that my USB-key, /dev/sdb1, is mounted on /hd-media type vfat. When I cd to /hdmedia, I find debian-testing-i386-netinst.iso, initrd.gz, ldlinux.sys, vmlinuz, and syslinux.cfg file, which has in it: default vmlinuz append initrd=initrd.gz So the USB-key is mounted, has the right versions of the files, the syslinux.cfg file has the correct information. Why is the installer then not picking up the ISO from the key instantly rather than search for it on the target HD? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87lj814ax0@teufel.historicalmaterialism.info
Re: squeeze install: Gave up waiting for root devices
I'm still struggling to create a squeeze install from USB-key. At this point a netinst installation using the squeeze netinst ISO. When I try to boot the key, I get only so far as "Verifying DMI Pool Data..." My impression this is likely to mean that my USB-key is not bootable. In fact, this raises a question. The reason my key is not bootable is because I followed the directions "4.3. Preparing Files for USB Memory Stick Booting" from http://www.debian.org/releases/stable/i386/ch04s03.html.en very literally. These are directions for installing lenny, but nothing I've seen makes them irrelevant to a squeeze installation except for the choice of installer files. In these directions, there is no step to make /dev/sdX1 bootable, and in fact it isn't bootable. The syslinux command writes a boot sector to the partition and creates the file ldlinux.sys which contains the boot loader code. But I assume that does me no good unless the partition is bootable. I assume I could do, for example, # cfdisk /dev/sdc1 and then issue its "b" command to set the boot flag for current partition. On a hard disk, the /boot partition is made bootable, but with the key, my boot files are in the single / partition, and so I assume I should run cfdisk on /dev/sdc1 rather than /dev/sdc. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87pqxe4i8r@teufel.historicalmaterialism.info
Re: squeeze install: Gave up waiting for root devices
I tried 15 seconds, but didn't help. I suspect my problem was that the /boot partition was not made bootable. I found that the squeeze installer wouldn't allow me to toggle the bootable flag on this partition, and so I resorted to an old /boot partition that was bootable. But I suspect that didn't do as I intended. Incidentally, I created a UNetbootin key and am trying a reinstallation using it. Works up to the above bootable toggle problem. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87y6c542c5@teufel.historicalmaterialism.info
Re: squeeze install: Gave up waiting for root devices
Jimmy Johnson writes: > brownh wrote: > >> I assume that the UUID assigned to my hard disk does not match the >> kernel line in GRUB2. Is there any way to find out what the assigned >> ID numbers are? >> >> Haines Brown > > > Yes, as Boyd said, as root in the console type 'blkid'. Yes, but thar's the rub! If I can't boot, I can't access a console, but only the (initramfs) prompt. That utility does not know the command blkid. When I do (initramfs) cat /proc/cmdline, I see the same disk UUID as that being used by GRUB2. Does this mean anything? Incidentally, setting rootdelay=15 for the kernel didn't help. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87lj86697k@teufel.historicalmaterialism.info
squeeze install: Gave up waiting for root devices
I'm doing a fresh squeeze install from USB key, and it installs grub 0.97 and kernel 2.6.26-2-686-bigmem. A single SATA drive. Boot hangs with Waiting for root filesystem... Gave up waiting for root device. ... I edited grub kernel line to add rootdelay=5, which didn't help. I tried (initramfs) echo $Root, and nothing was returned. When I did (initramfs) cat /proc/cmdline, I get: root=UUID=875fe3ac-ba94-41f8-9ace-23e7e3a0328b ro VGA=773 The grub configuration is: root (hd0,0) kernel /vmlinuz/2.6.26-2-686-bigmem root=UUID=... ro vga=773 initrd /initred.img-2.6.26-2-686-bigmem I assume that the UUID assigned to my hard disk does not match the kernel line in GRUB2. Is there any way to find out what the assigned ID numbers are? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/871v9z7ayl@teufel.historicalmaterialism.info
Bootable Flag: toggle problem with squeeze
I'm doing a fresh install of squeeze from USB key on a disk that had a unused copy of lenny on it. Ran into a problem when manually partitioning the disk. The first partition I went to create was a /boot primary partition, but I found that the partitioning utility wouldn't toggle its bootable status. I see that at one point this was a bug, but my impression is that it was fixed. I gather a solution is to make the /boot partition bootable by running cfdisk on it. I do Alt-F2 to get ~#, but find that cfdisk it is not an available command. So how can I run cfdisk on the disk before the installation finishes and there's a reboot? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/878w477lxf@teufel.historicalmaterialism.info
Re: Creating a bootable USB key
Jimmy Johnson writes: > brownh wrote: >> Jimmy, >> >> I inadvertangly sent this message to you personally, but here send it >> to the list. >> >> The directions for creating a bootable USB key don't mention running >> cfdisk on it to make it bootable. I simply copied syslinux to the key, >> but suspect I have to run cfdisk as well. Is that so? >> >> Haines Brown > > > I use gparted and it works for me, I do as I describe in the post you > are replying to. Jimmy, sorry to follow your kind answer with a question about it. This gnome application, gparted, is only described as a partition manipulator (I don't use any desktop environment, and so I'm not familiar with it). Does this "manipulation" include setting a boot flag? If so, then does your answer imply that this flag does indeed have to be set, despite an absense of its mention in some of the directions for making a bootable USB key? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87fwyg7svj@teufel.historicalmaterialism.info
Re: Creating a bootable USB key
Jimmy, I inadvertangly sent this message to you personally, but here send it to the list. The directions for creating a bootable USB key don't mention running cfdisk on it to make it bootable. I simply copied syslinux to the key, but suspect I have to run cfdisk as well. Is that so? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87k4nt72ax@teufel.historicalmaterialism.info
Re: OTF conversion without OpenOffice
Ron Johnson writes: > On 07/01/2010 08:42 AM, brownh wrote: >> Ron Johnson writes: >> >>> On 07/01/2010 06:11 AM, brownh wrote: >>>> 4. Antiword-for-Office is a perl script, but when I tried to compile, >>>> found I was missing the perl Archive::Zip module. Not knowing what to >>>> do about that and too little time to find out, I did not pursue. >> >>> This indicates that you must install libarchive-zip-perl. >> >> Thanks. This seemed to get through that hang in the compile, but now >> it hangs because it can't find XML/LibXML.pm. I did a search for >> LibXML, and the obvious package, libxml-libxml-common-perl, did not >> help (I already had libxml2 installed). > $ apt-file search libXML.pm > > Interesting. > > $ apt-file search libXML.pm > $ Apt-file search libXML.pm returns: libxml-libxml-perl: /usr/lib/perl5/XML/LibXML.pm If I understand correctly, if I have perl5 installed (which I do), I should find /usr/lib/perl5/XML/LibXML.pm on my machine. In fact it is not in my /usr/perl5/XML/ directory. I have Simple.pm, SAX.pm, NamespaceSupport.pm, and corresponding subdirectories there, but not that file. A search on line shows that the missing module is a common problem in a Windows environment. For linux, I gather the error can occur if there is no path to libxml2. But in my case, I got it finally to work by installing libxml-libxml-perl. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87pqz4v3m8@teufel.historicalmaterialism.info
Re: OTF conversion without OpenOffice
Camaleón writes: > On Wed, 30 Jun 2010 14:48:49 -0400, brownh wrote: > >> I received a .docx file appended in an e-mail, and need to extract and >> convert it to a convenient format such as .html, .pdf, or plain .txt. > > (...) > > If it's a simple file (just plain text) you can extract (unzip) the .docx > into *.xml data for a direct view or convert into another suitable format. Camaleón, I'm afraid you lost me. The file was .docx, which looks binary. As a result, it's MIME'd in the mail message, which makes it plain ASCII. But apparently you mean that I can run unzip on the .docx file to extract *.xml data. This was news to me, for I had no idea that .docx was an archive. But I tried it, and a number of things happened. It created an empty _rels directory; it created a docProps directory in which are app.xml and core.xml, and it created a word/ directory in which there are a number of *.xml files. None of these xml files are understood by abiword. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87tyojuwnj@teufel.historicalmaterialism.info
Re: OTF conversion without OpenOffice
Ron Johnson writes: > On 07/01/2010 06:11 AM, brownh wrote: >> 4. Antiword-for-Office is a perl script, but when I tried to compile, >> found I was missing the perl Archive::Zip module. Not knowing what to >> do about that and too little time to find out, I did not pursue. > This indicates that you must install libarchive-zip-perl. Thanks. This seemed to get through that hang in the compile, but now it hangs because it can't find XML/LibXML.pm. I did a search for LibXML, and the obvious package, libxml-libxml-common-perl, did not help (I already had libxml2 installed). >> 5. Unoconv script is a debian package and seems what I really >> want. However, when I ran it, I found that it depends on JRE, although >> "$ aptitude show unoconv" indicates that it depends on python. In any >> case, I don't happen to have JRE installed in current box, and so did not >> pursue. > > And you couldn't install it? No, the reason is that I'm working with temporary hardware and wanted to avoid doing that, but now I did install the Sun JRE. When I try to use unoconv on a .docx file I get: unoconv: UnoException during conversion: File could not be loaded by OpenOffice The provided document cannot be converted to the desired format. This sounds like it relies on OpenOffice, which I don't have installed. Haines -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87y6dvuxn1@teufel.historicalmaterialism.info
Re: OTF conversion without OpenOffice
Thank you, Matheiu, and others. I ultimately succeeded and here report my experiences with the options. 1. I found several on-line free conversion services. For various reasons such as security and privacy I did not pursue them. 2. Install OpenOffice and OpenOffice.OpenXML Translator. Because this contradicted my desire for command line conversion rather than install big GUI apps, I did not pursue. 3. Abiword can be used to convert the document format from .docx to, say, .pdf. It was my intent to use a command line utility instead, but here report that Abiword did in fact work and automatically detected the input format. 4. Antiword-for-Office is a perl script, but when I tried to compile, found I was missing the perl Archive::Zip module. Not knowing what to do about that and too little time to find out, I did not pursue. 5. Unoconv script is a debian package and seems what I really want. However, when I ran it, I found that it depends on JRE, although "$ aptitude show unoconv" indicates that it depends on python. In any case, I don't happen to have JRE installed in current box, and so did not pursue. 6. Odf-converter. This is a perl script. It requires libtiff.so.3, but by symlinking found that it can use libtiff.so.4 instead. With it I was able to generate an .otf file, which of course required Abiword to convert to PDF since I can't use unoconv. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87630zwj64@teufel.historicalmaterialism.info
OTF conversion without OpenOffice
I received a .docx file appended in an e-mail, and need to extract and convert it to a convenient format such as .html, .pdf, or plain .txt. Apparently .docx can be viewed in Abiword and OpenOffice, but I do not wish to install GUI applications, and so need a command-line format conversion utility. The best I find is OdfConverter (odf-converter_1.0.0-2-getdeb1_i386.deb). I installed this, but it lacks a manual and I have no idea how to use it. I gather that running $ OdfConverter wihout argument will return a list of possible arguments, but I don't seem to have any such OdfConverter executable. Anyone know of a simple command-line convertion utility? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87aaqcwe3y@teufel.historicalmaterialism.info
Re: lprng needed by AUCTeX?
On Sun, Sep 06, 2009 at 01:06:16PM +0200, Jörg-Volker Peetz wrote: > I would like to know how you decided to install the "cups" package, since its > description says it provides a print *server*. Well, three reasons: a) I thought a print server would provide print services, and that's what I wanted. I run plenty of other programs that provide services. b) The printing I have been doing I thought was the result of my having installed cups, but it was actually because I had long before installed lprng and had never uninstalled it. c) Also, "cups-bsd" sounded exotic, for I've no idea what the relation is of BSD and Debian. If I encountered cups-windows, I'd avoid it for the same reason. I mention this to show how ignorance let me down. > Nevertheless the cups package > depends on the package "cups-clients" which contains "client programs (SysV)", > i.e., the lp command for submitting a print job. > > So you don'nt need another package for printing with dvips, just change the > configuration of dvips in /etc/texmf/dvips/config/config.ps line 20 should > then > read (see comments in this file) > > %o |lp Ah, that's very simple - thanks. Now that I've installed cups-bsd there's no point in making the change, but good to know. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Re: lprng needed by AUCTeX?
Wayne, You are quite correct that a google search using your search terms immediately provided the answer to my question, which arose from a bug in the documentation (#512098). I assumed (wrongly) that CUPS was broken, and so googled with the wrong search terms. The tools you point to were of little help, for none of them explicitly said that dvi printing requires cups-bsd rather than the cups package, although now that I know that to be the case, I can see that the documentation might imply that. While the error report said it could not find the lpr command, a) the documentation with the cups package did not tell me it _didn't_ supply lpr, which I incorrectly assumed it did, and b) that dvi couldn't find the printer configuration file when in fact I knew it existed led me to assume that CUPS was broken and for that reason could not find lpr. The Debian Reference was new to me and looks to be very helpuful, but I did a quick look for this issue in it, and it does not explicitly say (at first glance), that the cups package does not supply lpr and so does not support printing with dvi. It simply gives an example of printing using lpr. I understand your frustration with questions raised by folks who do not trouble to figure things out for themselves. But the highest virtue of an educator, as you call yourself, is forebearance. As any educator will tell you, people often don't pursue things on their own because they lack the general framework needed to define the question in a way that implies what is a meaningful answer, lack a sense of direction in which a question needs to be pursued, or do not fully understand the documentation even when the answer stares them in the face. I did struggle with the problem before raising the question, such as googling, looking at the cups package documentation, and looking to see if there was a PRINTER variable. Why AUCTeX/dvi relies on the BSD command instead of vanilla CUPS is still a mystery. Everything else managed to print with the cups package (one app. didn't, but since I normally use it with a2ps, I didn't pursue the reason). I was focused on whether I had installed CUPS properly and whether it was a broken installation. In retrospect, the situation seems simple enough, but that is with the wisdom of hindsight. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Re: lprng needed by AUCTeX?
> The answer is already on your system, if you had done some > research. You seem to want others to research for you. > > Have you installed the cups-bsd package? Not sure what "on my system" means. The information associated with the packages for cups and auctex did not say I had to install cups-bsd (the AUCTeX description only "suggested" cups-bsd, not that it was mandatory if I expected to print anything). If by researching you mean reading the AUCTeX manual, I don't recall it said anything about this, but perhaps I missed it. You imply I should have done something I failed to do, but I'm unsure of exactly what. My debian system is a tool, not a hobby nor has any relevance at all to my work, which takes up all my time, days, evenings, weekends. Are you suggesting that if I cannot afford to make playing with debian a good part of my life I shouldn't mess with it? I did not ask that anyone do research for me, but merely sought simple answers to questions that anyone at all proficient with debian probably can answer off the top of their head. Sorry that this irked you. I'm one of those old fashioned folks who still like to hold doors open for someone else even if they don't appear to need the help. I have not installed the cups-bsd package, but will do so in the hope that it might somehow take care of my problem (you did not explictly say it would). Of course, if that works, that answers only one of my questions. Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
lprng needed by AUCTeX?
Lprng provides lpr and is characterized as a BSD "spooling system". CUPS, on the other hand, is described as a "printing system". I installed CUPS but not lpr/lprng. I cannot print from AUCTeX: Running `Print' on `test' with ``dvips -P hp_Laserjet_1320_series_USB_1 test'' dvips: warning: no config file for `hp_Laserjet_1320_series_USB_1' This is dvips(k) 5.96.1 Copyright 2007 Radical Eye Software (www.radicaleye.com) ' TeX output 2009.09.04:1059' -> | lpr -Php_Laserjet_1320_series_USB_1 sh: lpr: command not found Print broken pipe at Fri Sep 4 11:52:06 However, I do have /etc/cups/printers.conf that has: # Printer configuration file for CUPS v1.3.8 # Written by cupsd on 2009-09-04 11:39 Info hp LaserJet 1320 series Location Local Printer DeviceURI usb://HP/LaserJet%201320%20series State Idle StateTime 1252078723 Accepting Yes Shared Yes JobSheets none none QuotaPeriod 0 PageLimit 0 KLimit 0 OpPolicy default ErrorPolicy stop-printer Does this imply that dvips requires the installation of lpr/lprng in addition to CUPS? Is the warning about no config file for my printer only the result of not having lpr installed? What is implied here by the "broken pipe"? Haines Brown -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
