Re: Sizes and Packages in dselect
Erick Branderhorst writes (Sizes and Packages in dselect): * diskspace requirements (displayed/registered) [...] This is on the wishlist, and I think I know how to support it. This will become more difficult when the different directory's where parts of the packages are going to are mounted on seperate drive. Most of the people will probably have a big drive where /usr is mounted on and it will probably do, but with more and more packages coming up and networking setups are being considered this might be worth a thought. I can't think of a good way to support this, so I don't intend to. There will be one disk usage figure per package; users with several filesystems are presumed to know what they're doing. * dselect creating it's list from several package files dselect is building it's selection list from a file (Packages or something like that). What about making it possible that dselect builds it selection list from different Packages files or optionally from all the Packages files it recursively finds from some root directory. Perhaps this implemented now and I'm not aware. If so please ignore this, otherwise please comment on this one. At this moment a handy solution is probably catting all Packages files into one. This is on the wishlist. Ian.
Re: ld.so (fwd)
Ian Jackson: : Changing package names is usually a bad idea unless there's a good : reason. Or, more generally, inventing new interfaces where existing interfaces will suffice is usually a bad idea unless there's a good reason. : IMO the real solution is to have a real FTP method for dselect that : only gets the first few bytes of each package to check what it is. : This is doable, but someone has to go and write it. Sure, but this doesn't exist yet. Is it likely to soon? [I sure don't have the time -- I practically burned myself out last week.] : If people want to keep a mirror of our FTP archive then they don't : need a program like dftp. dftp is analogous to a mirror -- but takes less storage. : So, in summary, I think dftp is a mistake. Bringing down only part of each package followed by all of most packages isn't necessarily going to be more efficient. Having two distinct package names spaces where one will do isn't necessarily very elegant. It's true that dselect has more feature than dftp, but other than that why would you classify dftp as a mistake? -- Raul
Re: dselect install times out screensaver
Bill Mitchell writes (Re: dselect install times out screensaver): On Tue, 3 Oct 1995, Ian Jackson wrote: What we need is a sequence that means `reset the timeout now'. I could have dpkg output that every time it start processing a new package, if TERM=linux. (I'd rather not hardwire things like that, but if there's no other way ...) Alternatively it might be a better idea to change the screensaver to swallow the first keypress after blanking. I decided not to suggest this when I first thought of it but, since you seem open to having dpkg do some internal processing to work around spurious keystrokes, how about having it swallow all buffered keystrokes prior to running a postinst? No, I don't think that's the right solution. Apart from being a nasty hack, it wouldn't work, because dpkg would probably be sitting there doing nothing waiting at a prompt. Ian.
Re: dselect install times out screensaver
On Tue, 3 Oct 1995, Ian Jackson wrote: Bill Mitchell writes (Re: dselect install times out screensaver): Alternatively it might be a better idea to change the screensaver to swallow the first keypress after blanking. I decided not to suggest this when I first thought of it but, since you seem open to having dpkg do some internal processing to work around spurious keystrokes, how about having it swallow all buffered keystrokes prior to running a postinst? No, I don't think that's the right solution. Apart from being a nasty hack, it wouldn't work, because dpkg would probably be sitting there doing nothing waiting at a prompt. No argument about it being a nasty hack. That's why I didn't suggest it. Of course it'd be necessary to arrange for dpkg swallow the keystrokes without prompting, and detect nonavailibility of keystrokes to be swallowed to stop swallowing. Forget it. We've both got other fish to fry.
Re: xntp-3.4x fails to compile with linux 1.3.30
Torsten Duwe writes: That is a libc problem. I sent a patch to HJL over a year ago to prevent this, it got into the libc-source as README.distributors right away. Obviously some distributors didn't read it :-(. Andrew Ahh well I'm using 4.6.27 so it might be a little too old. Still Andrew haven't switched to ELF yet. It's not a question of version numbers, the README is in since 4.5.26, I think. It only matters if whoever compiled the libc has applied the patch to the kernel includes before. Where did you get your libc from ? Debian distribution, I don't know if they applied the patch or even know about it. I'm ccing this to them now anyway. Talking kernel mailing list: do you still have my previous mail ? did I Cc: it to the kernel mailing list ? I haven't seen it there. No sorry don't have it anymore, I haven't seen it on the kernel list either. Andrew --- Dehydration - 34%, Recollection of previous evening - 2%, embarrassment factor - 91%. Advise repair schedule:- off line for 36 hours, re-boot startup disk, and replace head - wow, what a night! -- Kryten in Red Dwarf `The Last Day' Andrew Howell [EMAIL PROTECTED] Perth, Western Australia [EMAIL PROTECTED]
Bug#1531: syslogd continues to write to old file after savelog
Package: syslogd Version: 1.2-11 As you can see below the syslogd is still writing to daemon.0 after savelog ran on Oct 1st. -rw-r--r-- 1 root adm 0 Oct 1 06:47 /var/log/daemon -rw-r--r-- 1 root adm207110 Oct 3 11:52 /var/log/daemon.0 root97 0.0 0.8 73 132 ? S Sep 28 1:01 /sbin/syslogd After I kill -HUPed the syslogd process it started writing to the proper logs i.e. daemon again. Andrew --- Dehydration - 34%, Recollection of previous evening - 2%, embarrassment factor - 91%. Advise repair schedule:- off line for 36 hours, re-boot startup disk, and replace head - wow, what a night! -- Kryten in Red Dwarf `The Last Day' Andrew Howell [EMAIL PROTECTED] Perth, Western Australia [EMAIL PROTECTED]
Bug#1531: syslogd continues to write to old file after savelog
I just looked into this a little more. It seems that the /var/run file for syslog has changed. /etc/cron.weekly/sysklogd tries to do this. kill -1 `cat /var/run/syslog.pid` but the file in /var/run is now this -rw-r--r-- 1 root root3 Sep 28 13:53 syslogd.pid Andrew --- Dehydration - 34%, Recollection of previous evening - 2%, embarrassment factor - 91%. Advise repair schedule:- off line for 36 hours, re-boot startup disk, and replace head - wow, what a night! -- Kryten in Red Dwarf `The Last Day' Andrew Howell [EMAIL PROTECTED] Perth, Western Australia [EMAIL PROTECTED]
Bug#1537: /usr/X11R6/lib/X11/app-defaults
Package: xbase? Version: 3.1.2-2 There is a directory /usr/X11R6/lib/X11/app-defaults, which contains defaults for various X applications. Are these defaults intended to modified on a system basis? If so, the directory should be under /etc and /usr/X11R6/lib/X11/app-defaults should be a symbolic link. -- David H. Silber [EMAIL PROTECTED] Project: Debian GNU/Linux (dbackup) http://www.access.digex.net/~dhs/ Wanted: Spare time. Programmer for hire.
Bug#823: starting tcsh causes network traffic
I haven't been able to reproduce this bug with my version of tcsh so I'm closing this bug. Andrew --- Dehydration - 34%, Recollection of previous evening - 2%, embarrassment factor - 91%. Advise repair schedule:- off line for 36 hours, re-boot startup disk, and replace head - wow, what a night! -- Kryten in Red Dwarf `The Last Day' Andrew Howell [EMAIL PROTECTED] Perth, Western Australia [EMAIL PROTECTED]
Bug#1187: Various `vi' versions trample
Let's try this again, filling in the Subject field instead of the Cc field. This bug is assigned to nvi, elvis. Since elvis and nvi now use update-alternatives, I've requested that this be reassigned to vim.
Re: Dselect / dpkg interaction.
In the case of the startup script running dselect, there is no interactive shell to take over when dselect gets a stop signal. Thus, you'd simply have to fork a shell at a lower level. Bruce
Bug#1525: root disk missing ls?
I tried KISS a while back. It's a step in the right direction, but doesn't have all of the programs I wanted, and wasn't as robust as I'd like it to be. I then wrote something with the tiny utilities like KISS, but no shell. It's intended to be linked into another small shell like ash. I think Ian has replaced bash on the root disk with ash, because it was Bourne-shell compatible and much smaller than bash. The shell in KISS is not useful for interpreting our scripts. All of these utilities are supposed to run in a fork(), not in the main thread, for robustness. These are the tools I have so far: cat chgrp chmod chown clear cp df false halt init ln mkdir mkswap more mount mv pwd reboot rm rmdir sleep swapoff swapon sync touch tput true umount update I would like to add my own tiny versions of ls, and tar, and link in the existing gzip, dialog, badblocks, and the new fdisk rewrite. I have the tar extractor function I wrote for dpkg, and can easily make a tar writer. To add gzip, dialog, badblocks, and fdisk, I will modify the source packages for these programs so that they will create library versions of themselves. This should give us a reasonably small static-linked environment. It should fit on the boot floppy along with the kernel and modules. Thanks Bruce
symbolic links within /bin
I was just looking around in /bin and noticed some symbolic links that may be incorrect. My system is 0.93R5+++ (I update the base regularly). The links I wonder about are csh - ../usr/bin/tcsh rmail - /usr/sbin/rmail rnews - /usr/sbin/rnews smail - /usr/sbin/smail tcsh (6.05-4) seems the most serious since if someone choses to use tcsh for root, it may not be available when the system is running. The others just seem a little odd. The FSSTND says things in /bin are essential to have the system running and I don't think rmail, smail and especially rnews falls into that category. Maybe its for traditional reasons, but I would think we would try to straighten it out. Are these bugs? If so, I will report each of them individually so that the maintainers can work independently and each defect can be resolved on its own. --- Kenny Wickstrom | gnu - a new generation in s/w devel/support [EMAIL PROTECTED] | Linux - a much improved Un*x clone (708)740-4008 | Debian - a Linux distribution setting the #include std-disclaimer |standard for future distributions
Bug#1532: no revision number with repair
Erick Branderhorst writes: Repair seems to be searching in the file /var/lib/dpkg/available to find some info about the specific package. It however never seems to find the debian revision number. I took a quick look in the script and it seems to search for a sequence package_revision. In the /var/lib/dpkg/available the keyword revision is used. In the control file of a package the keyword package_revision (case insensitive) should be used. Some inconsequence is spotted here. Perhaps this is a bug? in dpkg. The next version of repair, which isn't quite ready yet, calls dpkg to find information about the package rather than make assumptions about the format of internal databases; I'll check that it looks for the right field name tonight. There are one or two other things that need to be resolved before I upload it. Thanks for taking the time to have a look at it! -- Richard Kettlewell [EMAIL PROTECTED] http://www.elmail.co.uk/staff/richard/
Bug#1532: no revision number with repair
Package: repair Version: 0.1-1 (last 1 by hand) Libc: libc.so.4.6.27 Kernel etc: Linux morris 1.2.10 #1 Tue Jun 13 18:37:28 EST 1995 i486 Reporter: repair 0.1 Subject: Repair doesn't give debian revision Repair seems to be searching in the file /var/lib/dpkg/available to find some info about the specific package. It however never seems to find the debian revision number. I took a quick look in the script and it seems to search for a sequence package_revision. In the /var/lib/dpkg/available the keyword revision is used. In the control file of a package the keyword package_revision (case insensitive) should be used. Some inconsequence is spotted here. Perhaps this is a bug? in dpkg. Example: $ dpkg --status repair Package: repair Status: install ok installed Maintainer: Richard Kettlewell [EMAIL PROTECTED] Version: 0.1 Revision: 1 === Conffiles: /etc/repair.conf 1e16d01eb63e9955ba61692321457a20 Description: Report bugs in Debian Repair asks users to describe problems encountered and submits bug reports either to a local administrator or to the debian-bugs submission address. $ dpkg --info /root/debian/repair.deb old debian package, version 0.939000. size 6709 bytes: control archive= 1149, main archive= 5546. 1655 bytes,75 lines * bug #!/usr/bin/perl 17 bytes, 1 lines conffiles 290 bytes, 8 lines control Package: repair Version: 0.1 Package_Revision: 1 Maintainer: Richard Kettlewell [EMAIL PROTECTED] Description: Report bugs in Debian Repair asks users to describe problems encountered and submits bug reports either to a local administrator or to the debian-bugs submission address. The next few lines are found in /var/lib/dpkg/available Package: repair Maintainer: Richard Kettlewell [EMAIL PROTECTED] Version: 0.1 Revision: 1 === Description: Report bugs in Debian Repair asks users to describe problems encountered and submits bug reports either to a local administrator or to the debian-bugs submission address. Erick
New location for ispell-dictionary-devel.tar.gz
Since I cannot get through to ftp.debian.org this morning, I've put the ispell-dictionary-devel.tar.gz file up for anon ftp at /[EMAIL PROTECTED]:/pub/ispell-dictionary-devel.tar.gz If somebody else wants to upload to ftp.debian.org, then please go ahead. Kenny. -- | [EMAIL PROTECTED] http://www.glg.ed.ac.uk/~kenny; Try Linux! | | Portuguese/English/French Translations/Teaching by Native Portuguese | | [EMAIL PROTECTED] http://www.glg.ed.ac.uk/~kenny/helena; |
Dselect / dpkg interaction.
After talking to Stephen Tweedie [EMAIL PROTECTED] last night about my understansing of the current way which dselect and dpkg handle the 'Z' conffiles option, he came up with one possible solution I thought should be aired. At the moment, dselect sets an environment variable, and dpkg then tests this and spawns a new shell if appropriate. Is that correct? Stephen suggests that dselect should trap the SIGSTOP from dpkg, and not treat it as a SIGTERM. dselect should then send itself a SIGSTOP, followed by a SIGCONT. This means the signals are propogated back to the originating shell, altogether cleaner and less resource hungry. Ian, do you think this is a Good Thing? Just a thought for some mulling Kenny. -- | [EMAIL PROTECTED] http://www.glg.ed.ac.uk/~kenny; Try Linux! | | Portuguese/English/French Translations/Teaching by Native Portuguese | | [EMAIL PROTECTED] http://www.glg.ed.ac.uk/~kenny/helena; |
Re: FTP arrangement
: this is great for backwards binary compatability, something Linux : has excelled at. With all due respect, Linux has been worse about this than many other platforms, especially where C++ programs are concerned. FreeBSD, for example, has been more stable over the past couple of years. SunOS and Solaris ditto (but the new C++ compiler does mangle names differently, so that can be a big problem for library vendors). When I released the OI/OB product for Linux, it wouldn't work within less than six months without patches. Not a very good record :-(. And I keep reading about the Motif you have to relink/rebuild problems as well, which sound a lot like the same sorts of problems I had with OI. One problem of some ELF Motif packages is they were compiled for the older, alpha C library. I am sorry it happened that way. I added a stat () in libc 5.2.x for them. -- H.J. Lu NYNEX Science and Technology, Inc. [EMAIL PROTECTED]
Re: Dselect / dpkg interaction.
Bruce Perens writes (Re: Dselect / dpkg interaction. ): In the case of the startup script running dselect, there is no interactive shell to take over when dselect gets a stop signal. Thus, you'd simply have to fork a shell at a lower level. That's right. When the DPKG_NO_TSTP variable is set to `yes' dpkg does exactly this. Ian.
Bug#1525: root disk missing ls?
Bruce, would KISS the tiny shell with common commands builtin be useful to you? Its used in the SAR, search-n-rescue, boot disk. I added code to it to handle df so I got to know it pretty well. It uses the readline lib for command line edit and replay. This think could save a bunch a space. That's why it was created. Want more info or are you all set? Costa
Re: symbolic links within /bin
Kenny Wickstrom writes (symbolic links within /bin): I was just looking around in /bin and noticed some symbolic links that may be incorrect. My system is 0.93R5+++ (I update the base regularly). The links I wonder about are csh - ../usr/bin/tcsh rmail - /usr/sbin/rmail rnews - /usr/sbin/rnews smail - /usr/sbin/smail tcsh (6.05-4) seems the most serious since if someone choses to use tcsh for root, it may not be available when the system is running. Noone in their right mind changes root's shell. Creating an alternative root account with a different login name and tcsh as a shell would be the right thing to do. Perhaps chsh should warn people not to change root's shell ? The others just seem a little odd. The FSSTND says things in /bin are essential to have the system running and I don't think rmail, smail and especially rnews falls into that category. Maybe its for traditional reasons, but I would think we would try to straighten it out. Who put the rmail, rnews and smail links in /bin ? They're not part of the Smail package. Could you try dpkg --search /bin/rmail ? Ian.
Re: ld.so (fwd)
Raul Miller writes (Re: ld.so (fwd)): Ian Jackson: : IMO the real solution is to have a real FTP method for dselect that : only gets the first few bytes of each package to check what it is. : This is doable, but someone has to go and write it. Sure, but this doesn't exist yet. Is it likely to soon? [I sure don't have the time -- I practically burned myself out last week.] I don't think there are any immediate plans, but a rudimentary version that downloads everything would be quite straightforward. : If people want to keep a mirror of our FTP archive then they don't : need a program like dftp. dftp is analogous to a mirror -- but takes less storage. Unfortunately it doesn't work properly :-). : So, in summary, I think dftp is a mistake. Bringing down only part of each package followed by all of most packages isn't necessarily going to be more efficient. Having two distinct package names spaces where one will do isn't necessarily very elegant. No, but (for example) I don't think we can ensure that package names and version numbers are strictly seperable. It's true that dselect has more feature than dftp, but other than that why would you classify dftp as a mistake? I think that was rather harsh of me, actually. It's a reasonably good idea, but it relies on things it shouldn't rely on - which means we're getting bug reports about things that dftp doesn't like. Ian.
Bug#1531: syslogd continues to write to old file after savelog
Andrew Howell writes (Bug#1531: syslogd continues to write to old file after savelog): I just looked into this a little more. It seems that the /var/run file for syslog has changed. /etc/cron.weekly/sysklogd tries to do this. kill -1 `cat /var/run/syslog.pid` but the file in /var/run is now this -rw-r--r-- 1 root root3 Sep 28 13:53 syslogd.pid How about using something like start-stop-daemon --stop --signal 1 --user root --exec /sbin/syslogd ? Ian.
Bug#1533: `w' produces bogus login@, idle time c for xdm login entry
Package: procps Version: 0.97-4 -chiark:~ who ian :0 Oct 3 12:15 [...] -chiark:~ w 1:34pm up 3 days, 1:29, 8 users, load average: 0.57, 0.37, 0.35, 1/69 User tty login@ idle JCPU PCPU what ian :0 12:15pm75days 525:15 [...] Ian.
Re: symbolic links within /bin
On Tue, 3 Oct 1995, Ian Jackson wrote: Noone in their right mind changes root's shell. Creating an alternative root account with a different login name and tcsh as a shell would be the right thing to do. Forgive my ignorance, but I thought that both bash and tcsh would be available at all times. If that was the case, it would not matter if root shell was changed. Why don't I want to change root's shell? Are there security reasons? (This discussion might be valuable on debian-user.) Or is it just that currently no other shell lives in /bin? I guess I need to read my 'System Administrator's' book a little more. Who put the rmail, rnews and smail links in /bin ? They're not part of the Smail package. Could you try dpkg --search /bin/rmail Did that, and there was nothing found. I assume that this means these were put there by me installing some other program. Maybe the manual installation of UUCP (non-debian) did this. I don't know. I will try removing them. I now have the debianize UUCP. rnews may have come from an INN installation. Thanks for setting me straight, --- Kenny Wickstrom | gnu - a new generation in s/w devel/support [EMAIL PROTECTED] | Linux - a much improved Un*x clone (708)740-4008 | Debian - a Linux distribution setting the #include std-disclaimer |standard for future distributions
Re: Dselect / dpkg interaction.
Kenny MacDonald writes (Dselect / dpkg interaction.): After talking to Stephen Tweedie [EMAIL PROTECTED] last night about my understansing of the current way which dselect and dpkg handle the 'Z' conffiles option, he came up with one possible solution I thought should be aired. At the moment, dselect sets an environment variable, and dpkg then tests this and spawns a new shell if appropriate. Is that correct? Stephen suggests that dselect should trap the SIGSTOP from dpkg, and not treat it as a SIGTERM. dselect should then send itself a SIGSTOP, followed by a SIGCONT. This means the signals are propogated back to the originating shell, altogether cleaner and less resource hungry. I think Stephen is trying to solve a different problem here. What the problem was (and I hope it's fixed now) was that, when run automatically when the user logs in as root for the first time, dselect wasn't invoked from an interactive shell. Thus, when it got SIGTSTP there was no shell to take over control of the terminal. dselect doesn't set the DPKG_NO_TSTP environment variable - this has to be set by the script that runs dselect when root logs in for the first time. What dpkg does, if the environment variable isn't set, is to send the whole process group a SIGTSTP. This has approximately the same effect as if the user typed ^Z - it suspends both dpkg and dselect (and any script that might have invoked them), and when the shell notices this it gives a prompt to the user. When you say `fg' it sends both dpkg and dselect a SIGCONT, and things resume where they left off. Ian.
Bug#1535: GPL gone, LGPL exists
Ian, If you installed the base-0.93.6-6 package in pre-release/ over an existing system, it would overwrite a few conffiles. I am uploading a new one that doesn't have this problem, it will be base-0.93.6-7.deb . YOU MUST USE A CURRENT VERSION OF DPKG to install it. Older versions of dpkg would overwrite your passwd file when installing this package! As it is, you have to be careful to give the correct answers to not overwrite your current conffiles, and to not install /sbin/unconfigured.sh . But it will upgrade your documentation files, including your GPL file, if you install it. Thanks Bruce -- -- Attention Radio Amateurs: For information on Linux for Hams, -- read the WWW page http://www.hams.com/LinuxForHams, -- or e-mail the word help to [EMAIL PROTECTED]
Bug#1534: adduser problems with home directories
Package: adduser Version: 1.94-1 Firstly, it doesn't honour the --home option except when creating `system' users. This is clearly silly. If I say adduser --home /u2/fred fred then I want fred's home directory to be /u2/fred. Secondly, it doesn't set the setgid bit on home directories that it creates (when usergroups is enabled - clearly when it's disabled it shouldn't do it). Thirdly, I notice it makes many system calls without checking the error returns. Below is a patch that solves the first two problems. I have not bothered to produce a patch for the manpage (which says that the --home option only works with --system). Fixing the third problem will be quite a lot of work. Ian. --- adduser Tue Oct 3 03:23:52 1995 +++ /usr/local/sbin/adduser Tue Oct 3 13:39:53 1995 @@ -602,7 +602,11 @@ ## add the new user to the passwd file ## print Updating password file... if ($verbose); -$home_dir = $config{home} . / . $new_name; +if ($special_home) { + $home_dir = $special_home; +} else { + $home_dir = $config{home} . / . $new_name; +} add_user_to_file($new_name, $new_uid, $new_gid, @@ -651,6 +655,7 @@ } mkdir ($home_dir, $dir_mode); chown ($new_uid, $new_gid, $home_dir); +chmod ($dir_mode, $home_dir); print done.\n if ($verbose); ##
Bug#1535: GPL gone LGPL exist
I had a weird error on my machine. The file containing the text of the general public license /usr/doc/copyright/GPL is gone, the file /usr/doc/copyright/LGPL is still there. I don't know what did delete this file but I 'm quite sure I didn't do it manually. Is some package perhaps hardlinking/softlinking to this file and when removing the package the GPL is removed too? Am I missing more files on my system? How to check.? Erick
Bug#1539: strace source package does not compile
Package: strace Version: 3.0-1 I needed to recompile strace so that it didn't use /proc/pid/mem to get the arguments to system calls. This is because /proc/pid/mem is only available to root on my system, since having it available to anyone else is a security hole in even the most recent development kernels. (See the thread about my /proc paranoia patch on linux-kernel.) Fine, I thought, I'll just reconfigure it to use some variant of ptrace to get at the data. That bit was easy - I just copied the `old-style SunOS' code, which looked quite plausible and compiled OK. However, most of the rest of the strace source package is a screaming horror. It uses #define __KERNEL__ in several places, obviously to get at constants that probably ought to be in the general namespace. It uses in its own namespace sys_foo for many system calls foo, which clashes with the stuff you get when you say __KERNEL__. It doesn't appear to be compatible at all with my current kernel, which is 1.2.13. Below is a diff of the changes I made to the strace source tree before giving up. Note that I'm not sure whether my portability changes were going in the right direction. I think my ptrace fix was, though. Could the strace package maintaner and the program's author please get together and produce a version that (a) compiles and (b) works on secure systems. (It is possible that the procfs will be made secure in a forthcoming kernel version in a way that doesn't make it - including /proc/pid/mem - a gigantic security hole. Therefore, it may be a good idea to have strace try to open /proc/pid/mem. However, since there is an alternative strategy available it should use it if it discovers - at run time - that /proc/pid/mem doesn't work.) Ian. Only in /usr/local/src/strace-3.0: Makefile Only in /usr/local/src/strace-3.0: config.h Only in /usr/local/src/strace-3.0: config.status Only in /usr/local/src/strace-3.0: desc.o Only in /usr/local/src/strace-3.0: file.o Only in /usr/local/src/strace-3.0: io.o diff -ru /usr/local/src/strace-3.0/ipc.c /u/ian/download/strace-3.0/ipc.c --- /usr/local/src/strace-3.0/ipc.c Tue Oct 3 20:20:09 1995 +++ /u/ian/download/strace-3.0/ipc.cSat Mar 12 23:11:45 1994 @@ -38,33 +38,10 @@ #ifdef LINUX #define __KERNEL__ -#include linux/linkage.h -#define sys_semop x_sys_semop -#define sys_semget x_sys_semget -#define sys_semctl x_sys_semctl -#define sys_shmget x_sys_shmget -#define sys_shmctl x_sys_shmctl -#define sys_shmat x_sys_shmat -#define sys_shmdt x_sys_shmdt -#define sys_msgget x_sys_msgget -#define sys_msgctl x_sys_msgctl -#define sys_msgsnd x_sys_msgsnd -#define sys_msgrcv x_sys_msgrcv #include sys/ipc.h -#include sys/msg.h #include sys/sem.h +#include sys/msg.h #include sys/shm.h -#undef sys_msgget -#undef sys_msgctl -#undef sys_msgsnd -#undef sys_msgrcv -#undef sys_semop -#undef sys_semget -#undef sys_semctl -#undef sys_shmget -#undef sys_shmctl -#undef sys_shmat -#undef sys_shmdt #undef __KERNEL__ static struct xlat msgctl_flags[] = { Only in /usr/local/src/strace-3.0: ipc.c~ Only in /usr/local/src/strace-3.0: ipc.o Only in /usr/local/src/strace-3.0/linux: Makefile Only in /usr/local/src/strace-3.0/linux: errnoent.h Only in /usr/local/src/strace-3.0/linux: ioctlent.raw Only in /usr/local/src/strace-3.0: strace.o diff -ru /usr/local/src/strace-3.0/syscall.c /u/ian/download/strace-3.0/syscall.c --- /usr/local/src/strace-3.0/syscall.c Tue Oct 3 20:07:50 1995 +++ /u/ian/download/strace-3.0/syscall.cSat Mar 12 23:11:57 1994 @@ -475,7 +475,6 @@ tprintf(= %#x, tcp-u_rval); } else if (!(sys_res RVAL_NONE) u_error) { -#ifdef THIS_IS_NOT_IN_LINUX_1_2_13 #ifdef LINUX switch (u_error) { case ERESTARTSYS: @@ -489,7 +488,6 @@ break; default: #endif /* LINUX */ -#endif /* THIS_IS_NOT_IN_LINUX_1_2_13 */ tprintf(= -1 ); if (u_error nerrnos u_error sys_nerr) tprintf(%s (%s), errnoent[u_error], @@ -502,12 +500,10 @@ sys_errlist[u_error]); else tprintf(E??? (errno %d), u_error); -#ifdef THIS_IS_NOT_IN_LINUX_1_2_13 #ifdef LINUX break; } #endif /* LINUX */ -#endif /* THIS_IS_NOT_IN_LINUX_1_2_13 */ } else { if (sys_res RVAL_NONE) Only in /usr/local/src/strace-3.0: syscall.c~ Only in /usr/local/src/strace-3.0: syscall.o diff -ru /usr/local/src/strace-3.0/util.c /u/ian/download/strace-3.0/util.c --- /usr/local/src/strace-3.0/util.cTue Oct 3 20:01:04 1995 +++ /u/ian/download/strace-3.0/util.c Mon Mar 21 05:45:46 1994 @@ -474,30 +474,6 @@ { #ifdef LINUX -#define
Re: Bug#1535: GPL gone LGPL exist
Bill Mitchell writes (Re: Bug#1535: GPL gone LGPL exist): On Tue, 3 Oct 1995, Ian Jackson wrote: Erick Branderhorst writes (Bug#1535: GPL gone LGPL exist): [...] This is very interesting. The same thing has happened to my system. The file /usr/doc/copyright/GPL doesn't seem to be in any of the packages on the src.doc.ic.ac.uk mirror of ftp.debian.org, and it isn't mentioned in any of the /var/lib/dpkg/info/* files on my system. I conjecture that there was once a package that contained this file (by mistake) and that the package maintainer removed it. I audited some old packages I have lying around, and found this file in dpkg-0.93.42-0 and dpkg-0.93.42.2-0. It looks like you've found the culprit, then. My apologies for this mistake. I'm marking this bug as done (more recent versions of dpkg don't have this problem). Bruce Perens writes (Re: Bug#1535: GPL gone, LGPL exists): If you installed the base-0.93.6-6 package in pre-release/ over an existing system, it would overwrite a few conffiles. I am uploading a new one that doesn't have this problem, it will be base-0.93.6-7.deb . OK, thanks. YOU MUST USE A CURRENT VERSION OF DPKG to install it. Older versions of dpkg would overwrite your passwd file when installing this package! As it is, you have to be careful to give the correct answers to not overwrite your current conffiles, and to not install /sbin/unconfigured.sh . Have you considered putting a test for the dpkg version in your preinst ? But it will upgrade your documentation files, including your GPL file, if you install it. Excellent, thank you. I'll grab it when it becomes available. Ian.
Re: Sizes and Packages in dselect
[EMAIL PROTECTED] said: I can't think of a good way to support this, so I don't intend to. There will be one disk usage figure per package; users with several filesystems are presumed to know what they're doing. I think I have the algorithm to do this. Please see if the following makes sense. I can give you pseudocode (or real code) if necessary. For each file in the files list, find the deepest existing directory that matches the pathname. This will often be something like /usr, /etc, and so on. Cache the name of this directory and information about it so that subsequent lookups will run quickly. Get the block device number from stat.st_dev of the directory, to disambiguate this filesystem from other filesystems. Cost the file to the indicated filesystem, and go on to the next file. When done processing the file list, use the block device number to find the block device name, and then use getmntent() to find the mount point for display to the user. Use fsstat() to find the free space count for the filesystem. When replacing packages, subtract the cost of the old package file list before adding the cost of the new one. When overwriting files that are not part of a package being replaced, subtract their cost before adding the new file cost. Use fsstat.f_bsize and fsstat.f_frsize to figure out how many blocks/fragments a file will take up, so that your block count will be more accurate than if you simply counted bytes. Does this make sense? I can elaborate if you like. Bruce -- -- Attention Radio Amateurs: For information on Linux for Hams, -- read the WWW page http://www.hams.com/LinuxForHams, -- or e-mail the word help to [EMAIL PROTECTED]
a new UNIX utility (fwd)
PSHIFT(1) USER COMMANDS PSHIFT(1) NAME pshift - paradigm shift utility SYNOPSIS pshift [-zzeitgeist] [-rragelev] [-v] [-c] [-wn] [+|-n] DESCRIPTION The pshift operator performs a paradigm shift on its input stream within the context of the current or specified zeitgeist. OPTIONS -z Specify the zeitgeist context. May be specified here or from the environment variable $ZEITGEIST. Supported values of zeitgeist are judeo_christian (default), postcommunist, new_age, and when_god_was_a_woman. -r Specify rage level. Acceptable values of ragelev are ennui (default), deep_seated, and consuming. -v Set to verbose mode. Normally pshift operates silently; in verbose mode it publishes a 500+ page bestseller entitled Rethinking [input stream] in the [zeitgeist] Age, and then begins soliciting honoraria until the operator types ctrl-c. On some systems it runs for Congress. -c Set to collective IO. Normally pshift takes its input from stdin and outputs to stdout; in collective mode it takes its input from the Collective Unconscious and writes to the Body Politic. -wn Specify first, second, third or fourth wave. Acceptable values for n are 0,1,2 or 3, with 2 (third wave) being the default. [On Sun systems, the logical waves are 0,3,2,1, which map to physical waves 0,1,2,3; see Sun Technical Manual for details.] +|-nSpecifies the number of times to prepend 'post' to the zeitgeist context, if positive, or 'pre' if negative. The default is 11. EXAMPLES source $DEITY | pshift -zpostcommunist -rdeep_seated -v +1 On most systems, the above command will output a hardcover volume called Rethinking God in the Post-Postcommunist Era, in which the irrelevence of erstwhile religious concepts is seen to have triggered a global, deep-seated rage vis-a-vis traditional sociopolitical norms leading to a premature breakdown of emerging postsoviet infrastructure. pshift -znew_age -rennui The above command produces no output, but privately processes a vague discontent which it will share if its space is honored. May be redirected to /dev/null. pshift -c -w3 -1 Taking its input from the collective unconscious, the above command rejects the failed socioeconomic policies of the last thirty years and replaces them with a futurist, fourth wave polemic of traditional values, the two-parent family, and the supremacy of the private sector that was the foundation of the American utopia of the 1950s. Use a prepend value of -2 to restore the American utopia of the early Industrial Age, a value of -3 to restore the European utopia of the Enlightenment, -4 for catholic hegemony, etc. (note: Requires grass root permission. In verbose mode, it may also require a $4 million advance.) SEE ALSO backlash(1) BUGS You must have root permission to use consuming rage. AUTHOR Robert Drucker ([EMAIL PROTECTED]) copyright 1995 Robert Drucker. Robert Drucker is a trademark of Robert Drucker. --- End of Forwarded Message
Bug#1540: fvwm should require/recommend m4
Package: fvwm Version: 1.24r Revision: 6 The system.fvwm pretty much requires m4 (you get errors in your ~/.xsession-errors, not to mention the fact that it's basically unusable) but it's not even recommended by the package. Mike. -- And I swear that I don't have a gun.