fetch/ftp problem
I am having an odd problem with fetch. I cannot fetch an ftp:-type address but I can ftp to it and get the files just fine. This is problematic for ports and downloading via sysinstall. I am running the new 5.2. I am firewalled though (I did try both passsive and active ftpmode's). Any advice on how to further debug? -r ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
fixing motherboard sound
OK, I found one of the problems that the festival app was having playing on my system. Apparently, the authors assumed that a SNDCTL_DSP_SYNC ioctl() was not necessary before a close() because the close() is supposed to sync. I have found, however, that it does make a difference. Is close() supposed to cause a SYNC, as far as anyone knows? I am not subscribed. -r ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
why does motherboard sound suck?
I just got new hardware and I installed 5.1 on it (not tht it matters, because what I am asking about is observed in 4.X). Why is it that so-called the built-in soundcard that my motherboard comes with does work with everything? By that I mean that I can play the usual multimedia stuff without a hitch, xmms, mpg123, mplayer, artsd all work just fine. The card comes up as some AC97 under pcm. But if I try to use some of the more esoteric ports like stella or festival, I get no sound. With festival I can do a trick like filter the output through the command 'sox -r 1600 -t sw $FILE -r 4800 -t ossdsp /dev/dsp' because it provides hooks to do this, but I don't see why I have to (btw the port can't build festival on 5.1 yet, but you can on 4.X, and I am working on patching it for 5.1). stella, the atari emulator, has the same trouble. I have observed this on another motherboard with a so-called built-in soundcard as well (mpg123 plays, sox plays, stella doesn't) about 2 years ago (problem solved by buying a SB16 for it). The solution has, in the past, been to just keep using an old soundblaster card and disabling motherboard sound in favor of the SB card. But I am tired of using up a PCI slot just to keep a handful of my favorite ports running. I've looked at the sources to festival (and a bit at stella, though I do not understand them as well), specicially the voxware.cc file which seems to talk directly to /dev/dsp. It doesn't seem to do anything fancy, just sets the same rate and copies an array of sound bytes to a FILE*. So I ask, where does the fault lie? Is stella/festival soundblaster centric, and just won't talk to AC97 cards? Is my motherboard AC97 deficient? And why are motherboard cards deficient? Should I focus my energy on my hardware (replace card)? My OS (debug AC97 driver or find a sox-like workaround for stella)? Or my apps (debug stella and festival)? I'll bet there are other ports that I don't use that have this same trouble. -r ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
umass device incomplete contents
Hi, ever since we upgraded to 5.0 we have had problems with our SDR-31 compact flash reader. No interesting messages come up in dmesg to describe the problem. This is what happens, the device mounts just fine: mount -o ro -t msdos /dev/da0s1 /mnt/ but while we can see all the files with an 'ls' we cannot actually get the contents of some of the files: % ls % md5 * the ls works, but the md5 command gives an input/output error on some of the files. Really the md5 command is just a quick way to test the 'open' and 'read' calls; 'cp' likewise fails on these files. To try and diagnose this further, we did % dd if=/dev/da0s1 of=/tmp/img and took a look at /tmp/img. It is a 256MB card. The img file was only 8MB in size. When I page through the img file with vi or emacs, it is clear that the FAT is intact, and a few jpeg's are present, but not all of them are there. Thus, I conclude that somehow or another, this device appears to be 'truncated'. This is consistent with 'ls' working and md5 only working on some of the files (the ones which haven't been truncated away). Has anyone else seen this? I didn't see a PR about it in the gnats. -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
devfs and /dev/fd/3
I just installed freebsd 5.0. One gotcha that got me was in my lpif filter, which used to use a clever set of redirects to send only the outputfile output of gs to stdout while the stdout output of gs was sent to stderr: exec 31 12 /usr/local/bin/gs -dSAFER -dNOPAUSE -q -sDEVICE=cdj550 \ -sOutputFile=/dev/fd/3 - exit 0 However, with the new devfs, I seem to be short a file descriptor, namely, fd/3. So I must do /usr/local/bin/gs -dSAFER -dNOPAUSE -q -sDEVICE=cdj550 \ -sOutputFile=- - exit 0 Which will work as long as gs doesn't decide to write something to stdout, at which point something bad will come out of my printer. Is there a better way to accomplish the above? -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
DSL, verizon, FreeBSD
Hi. I'm trying to get a verizon DSL service started on my FreeBSD box. I don't have windows running at home, so I cannot run their install software. The nice folks there, have been helpful enough to walk me through the install. I'm happy they are taking a shot at it. They have given me the install username and password, which will expire in 24 hours. If I can get the modem working in that time, then we can proceed to registration and the rest. The modem is a westell, and I have connected everything up (subquestion: can PPPoE be run on the same NIC as my LAN? Right now the other computer is disconnected from the hub so it is just my gateway and the modem.). I have used the example from the handbook: default: set log Phase tun command # you can add more detailed logging if you wish set ifaddr 10.0.0.1/0 10.0.0.2/0 name_of_service_provider: set device PPPoE:xl1 # replace xl1 with your ethernet device set authname YOURLOGINNAME set authkey YOURPASSWORD set dial set login add default HISADDR with rc.conf ppp_enable=YES ppp_mode=ddial ppp_nat=YES # if you want to enable nat for your local network, otherwis ppp_profile=name_of_service_provider (and all the LAN related stuff in rc.conf commented out to prevent any possible interferences, see subquestion). I am getting dial-carrier Hangup! in the ppp.log file (I won't be back home til this evening, if someone wants more lines from the logfile). Which seems bad. I have tried manually started ppp and using the 'term' option (I used to have very informative conversations with my phone-modem this way), but the DSL modem doesn't say anything (no happy 'OK', but I am not sure if it is supposed to). Anyhow, does anyone out there have this setup? Suggestions on changes to my rc.conf or ppp.conf or anything? -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
mapping ports to packages
I am writing a little utility to help me get an instant-workstation (the instant-workstation port is fine when you have high bandwidth, but not when you are low bandwidth and have CDs with some stuff already, blah blah blah, there are other reasons why I believe this would be a nice thing). Anyhow, I'm trying to make it flexible enough that it will fetch from ftp or the CDs or build the port, depending on user input. However, this bring up a bit of FreeBSD ugliness: port vs package names. Three bits of functionality are needed here, as I see it: 1) figuring out whether a specified package/port is installed 2) figuring out what package name to use with pkg_add -r 3) figuring out which port name to CD to to make install With the user specifying either a port name or a pkg name. If the user specifies a port directory, than #1 and #3 seem to be easy to take care of, since #1 can be check by looking for ORIGIN comments in /var/db/pkg. (note: if I go with package names, then I have to deal with their brittleness due to version information encoded in them in non-uniform way, which requires more patter matching than I really think I know how to do.) #2 is trickier. I know that somewhere in the Makefile of each port is a set of variables, such that when I concat them I can the package name, but port makefiles are a bit icky, and I was wondering if there was a quick way (like 'make package_name') to spit out the portname--packagename map I want? So, how do I get a package name from a port nicely? -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
doh! I found my answer
Don't bother to respond to what I just wrote. I have discovered that make -V PKGNAME will do the mapping I want. -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
trying to get a fresh install
Hi. I'm trying to make a fresh install on a computer which had 4.6 on it (installing 4.7) and for some reason, although the NEWFS=Y in every row of the disklabel dialog, I am seeing clear signs that sysinstall is reading my old /var/db/pkg info among other things. This indicates to me that the newfs was somehow incomplete. It is certainly possible that the disklabel partitions are just identicle enough that it would line up correctly. I do find it very disconcerting that this happens. How do I get newfs for real? This is an ugly mess. -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: trying to get a fresh install
I don't have an msdos disk. I have no windows or msdos at home. -r On Wednesday, 23 October 2002 at 20:04:12 -0400, JoeB wrote: Use msdos fdisk, repartition disk and format it. -Original Message- From: [EMAIL PROTECTED] [mailto:owner-freebsd-questions;FreeBSD.ORG]On Behalf Of Ross Lippert Sent: Wednesday, October 23, 2002 7:50 PM To: [EMAIL PROTECTED] Subject: trying to get a fresh install Hi. I'm trying to make a fresh install on a computer which had 4.6 on it (installing 4.7) and for some reason, although the NEWFS=Y in every row of the disklabel dialog, I am seeing clear signs that sysinstall is reading my old /var/db/pkg info among other things. This indicates to me that the newfs was somehow incomplete. It is certainly possible that the disklabel partitions are just identicle enough that it would line up correctly. I do find it very disconcerting that this happens. How do I get newfs for real? This is an ugly mess. -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: /usr/share/examples
Some of us never make world. Or rather, there must be an easier way to get this directory populated than having to learn how to make world. make world kind of makes me nervous. If this is a recognized problem, then why isn't it reported in the errata? Is it too much to ask for someone to just make a tarball and put it on the errata page? No, they're in /usr/src/share/examples (or something like that--go into /usr/src and look and you'll find them.) It gets fixed the first time you do a make world. -r To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: MS Dynamic DNS problems
OK, here is /etc/resolv.conf: search celera.com ad.celera.com rkv.ad.celera.com rkv.celera.com applera.com pe-c.com ssf.ad.celera.com fc.celera.com nameserver 172.20.7.10 nameserver 172.20.7.11 The search line there appears to be the result of adding a supercede line to the dhclient.conf, which is as follows: interface ep0 { supersede domain-name celera.com ad.celera.com rkv.ad.celera.com rkv.celera.com applera.com pe-c.com ssf.ad.celera.com fc.celera.com; request subnet-mask, broadcast-address, time-offset, routers, domain-name, domain-name-servers, host-name; } The above contents are just a total guess based on what the Win-support person seemed to be doing. I visited the win2k side again, to check my name list and so forth, and did CMD ipconfig /all The primary name is lipperra-p1, and the connection specific one is rkv.ad.celera.com. I am not sure if that is reflected in my resolv.conf or my dhclient.conf. Finally, on the boot back to FreeBSD, I ran a ping from my desktop machine and watched the disappearance from DNS: cglwadministrator@LIPPERRA-W1 ~ $ ping lipperra-p1.rkv.ad.celera.com Pinging lipperra-p1.rkv.ad.celera.com [172.20.168.104] with 32 bytes of data: Reply from 172.20.168.104: bytes=32 time10ms TTL=128 Request timed out. Request timed out. Reply from 172.20.168.104: bytes=32 time10ms TTL=64 Request timed out. cglwadministrator@LIPPERRA-W1 ~ $ ping lipperra-p1.rkv.ad.celera.com Bad IP address lipperra-p1.rkv.ad.celera.com. which may be of interest because it shows the DNS entry being persistent right up until just after FreeBSD's dhclient starts, and then the DNS entry disappears. -r From: David [EMAIL PROTECTED] To: Ross Lippert [EMAIL PROTECTED], [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] References: [EMAIL PROTECTED] Subject: Re: MS Dynamic DNS problems Date: Thu, 18 Jul 2002 08:11:59 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2600. X-Mimeole: Produced By Microsoft MimeOLE V6.00.2600. - Original Message - Thanks for the quick reply, Steve. I suspect that since they were about to modify my laptop's win2k install so that it started showing up in DNS, that there must be something doable on the client-side alone, if only I knew what that was. I have since rebooted to win2k to look at the configuration panel that was changed to get the fix. It is the advanced popup under the DNS tab in the TCP/IP properties (gosh why can't people just use directories). The suffix list used to be empty, and unchecked, and now it is populated and checked, as previously described. Secondly, I have noticed a checkbox at the bottom of this panel of the form use suffix in registration which might have also been checked during the fix. While win2k was up, I was able to ping the laptop. Then I rebooted to FreeBSD (getting the same IP addr from DHCP) and pinged again and my DNS entry had disappeared, no ping. Though pings by raw IP addr work fine. There are a couple palces in your /etc directory you may want to look. For example, the settings you indicate they added are usually stored in the /etc/resolv.conf file. Make sure you have the proper domain specified there. Also make sure your fully-qualified domain name includes the appropriate domain, and the same name as the Win2k side of your machine. David To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message