Attempting a Debian Install on a Libretto 100CT
The Philadelphia Area Debian Society (PADS) (http://www.CJFearnley.com/pads/) Presents Attempting a Debian Install on a Libretto 100CT When: Wednesday 16 April 2003, 8:00 PM - 9:30 PM Presenter: Mike Leone Where: 4416 Osage, Apartment 20 Philadelphia, PA Thanks to our hosts: Samantha and Fred Ollinger Abstract We will attempt to install Debian on a Libretto 100CT. Installing on this laptop is complicated by the PCMCI-Bridge which isn't recognized after the kernel boots. We will explore options for getting this installed in the hopes that Mike will have Debian on his laptop at the end of the evening. Social Dinner Attendees are invited to gather for dinner prior to the meeting at 6:30 PM at Abyssinia Ethiopian Restaurant, 229 S 45th St, Philadelphia, PA 19104-2918, (215) 387-2424. Please RSVP, so that we can call ahead for reservations. -- Christopher J. Fearnley | LinuxForce Inc. [EMAIL PROTECTED] | Chief Technology Officer http://www.LinuxForce.net | Software Solutions / Systems Management
Building Debian Packages by Example: Packaging libgnupg.pm
The Philadelphia Area Debian Society (PADS) (http://www.CJFearnley.com/pads/) Presents Building Debian Packages by Example: Packaging libgnupg-perl When: Wednesday 16 August 2000, 8:00 PM - 9:30 PM Facilitator: Chris Fearnley, Chief Technology Officer, LinuxForce Inc. Where: IQ Group's Technology Lab The Constitution Building, 12th floor, Suite 1200 325 Chestnut Street Philadelphia, PA Abstract We will demonstrate the building of a Debian package by preparing a .deb file and source packages for libgnupg-perl, a Perl Interface for Gnu Privacy Guard (GnuPG.pm). Since this is a perl library module a few extra considerations are applicable. However, GnuPG.pm is still very simple so this session will be appropriate for those new to Debian development. Social Dinner Attendees are invited to gather for dinner prior to the meeting at 6:30 PM at The Nile Restaurant, 120 Chestnut Street, Philadelphia, PA. Please RSVP so we can get an appropriate sized table. -- Christopher J. Fearnley | LinuxForce Inc. [EMAIL PROTECTED] | Chief Technology Officer http://www.LinuxForce.net | Design Science Revolutionary Dare to be Naïve -- Bucky Fuller
Using Debian's make-kpkg to build kernels
The Philadelphia Area Debian Society (PADS) (http://www.CJFearnley.com/pads/) presents Using Debian's make-kpkg to build kernels When: Wednesday 15 March 2000, 8:00 PM - 9:30 PM Speaker: Chris Fearnley, Chief Technology Officer, LinuxForce Inc. Where: IQ Group, 6th floor (its the room with a big Q on the door) 325 Chestnut Street Philadelphia, PA Abstract Debian's kernel-package package provides make-kpkg to help build kernels. We will discuss the advantages and some guidelines for using this convenient tool. Social Dinner Attendees are invited to gather for dinner prior to the meeting at 6:30 PM at Xando, 325 Chestnut Street, Philadelphia, PA. Please RSVP so we can get an appropriate sized table. -- Christopher J. Fearnley | LinuxForce Inc. [EMAIL PROTECTED] | Chief Technology Officer http://www.LinuxForce.net | Design Science Revolutionary Dare to be Naïve -- Bucky Fuller
Re: Migrating to GPG - A mini-HOWTO
On Sun, Sep 12, 1999 at 05:43:21PM -0400, Brian Almeida wrote: How to switch to GnuPG for developers..a very brief mini-HOWTO -- Very nice mini-HOWTO. But I still have several questions: How does one generate an RSA key using the gpg-rsaref package? How does one send the RSA key to [EMAIL PROTECTED] Is this all that is needed to be added to the Debian keyring for GPG? Are the gpg-rsaref and gpg-idea packages both needed? Why? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.CJFearnley.com| Explorer in Universe Dare to be Naïve -- Bucky Fuller
[Philadelphia] Organizational meeting for Debian user's group
Greetings, There seems to be enough interest to form PDG-LUG (The Philadelphia Debian GNU/Linux User's Group). In order to try to accommodate people with families and suburban Debian GNU/Linux users, we will have an optional ``social hour'' at a Center City eatery BEFORE the 8:00 PM meeting. PDG-LUG Social Hour: When: Wednesday, May 19th from 6:30 PM - 7:45 PM Where: Pietro's Coal Oven Pizzeria, 1714 Walnut ST, Philadelphia RSVP: [EMAIL PROTECTED] so I can reserve an appropriately sized table PDG-LUG Main Meeting: When: Wednesday, May 19th from 8:00 PM - 9:30 PM Where: Cyberloft, 1525 Walnut Street, 2nd floor, Philadelphia Topic: Organizational Meeting; QA I hope to see you next Wednesday. Do Enjoy! PS. I'm hopelessly behind in reading Debian's mailing lists. Please copy me on any repies. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.CJFearnley.com| Explorer in Universe Dare to be Naïve -- Bucky Fuller
Re: Serious performance bug in Perl
No, I am not running NIS. Just simple text /etc/passwd and /etc/shadow. On Thu, Jun 18, 1998 at 12:38:45PM -0500, Richard Kaszeta wrote: Christopher J. Fearnley writes (Re: Serious performance bug in Perl): to call it (instead of the default perl - 5.004.04-6). Performance improved several hundred-fold. So I believe the problem is either in perl or libc6. Any suggestions on how to resolve this? As I said before the slowdown seems to occur in the get_current_uids subroutine (and possible get_current_gids). Which has a loop on getpwent (and getgrent). Can anyone else duplicate this behavior? I can duplicate this behavior. Performance gets exponentially better if I move my NIS password records into the local password file. So in my case I am tempted to blame libc6's NIS performance (which in other circumstances I have found to be rather slow anyways) Are you running NIS? -- Richard W Kaszeta Graduate Student/Sysadmin [EMAIL PROTECTED] University of MN, ME Dept http://www.menet.umn.edu/~kaszeta -- Christopher J. Fearnley | Internet21 Network Engineering [EMAIL PROTECTED] | Design Science Revolutionary http://www.i21.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Serious performance bug in Perl
'Darren/Torin/Who Ever... wrote:' Chris Fearnley, in an immanent manifestation of deity, wrote: But yesterday I upgraded a bo system to hamm which has a 3000 line /etc/passwd. Now adduser takes OVER ONE MINUTE to find a UID and GID for the new user. And my staff is complaining about the wasted time. I fear that this perl bug is serious. Something is wrong with your installation or possibly libc. I compiled perl-5.003_07 and perl-5.004_04 on a Solaris box with 5000 users. The 5.004_04 was somewhat faster. Maybe it's a problem with perl on libc6 systems with shadow passwords??? It was a straightforward upgrade from bo. I can't imagine how a misconfiguration could cause this. I have the latest of everything relevant installed: [EMAIL PROTECTED]:~$ dpkg -l libc6 perl passwd adduser Desired=Unknown/Install/Remove/Purge | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ NameVersionDescription +++-===-==- ii libc6 2.0.7pre1-4The GNU C library version 2 (run-time files) ii perl5.004.04-6 Larry Wall's Practical Extracting and Report ii passwd 980403-0.2 Change and administer password and group dat ii adduser 3.8Add users and groups to the system. [EMAIL PROTECTED]:~$ ls -l /etc/passwd /etc/shadow /etc/group /etc/gshadow -rw-r--r-- 1 root root48419 Jun 15 21:41 /etc/group -rw-r- 1 root shadow 35754 Jun 15 21:41 /etc/gshadow -rw-r--r-- 1 root root 191178 Jun 15 21:41 /etc/passwd -rw-r- 1 root shadow 124656 Jun 15 21:41 /etc/shadow Could it be a problem with shadow passwords? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Serious performance bug in Perl
'Wichert Akkerman wrote:'
Previously Chris Fearnley wrote:
But yesterday I upgraded a bo system to hamm which has a 3000 line
/etc/passwd. Now adduser takes OVER ONE MINUTE to find a UID and GID
for the new user. And my staff is complaining about the wasted time.
Are you sure it's a problem with perl? I've had the same problem with
tar becoming dog-slow because I had `compat' in /etc/nsswitch.conf.
Changing that to `nis files' fixed the speed-problem.
Goodness, the libc6 docs on the NSS mechanism are horrible. But in
the interests of experimentation, I tried changing this from compat
to files. Still dog-slow.
You could probably use strace to check what is really causing the delay.
Wichert.
I did try strace it looked like it was doing illions of open's and
then mmap's on /etc/shadow at the point where it was slow. Here is
the recurring part of the strace:
open(/etc/shadow, O_RDONLY) = 5
fcntl(5, F_GETFD) = 0
fcntl(5, F_SETFD, FD_CLOEXEC) = 0
fstat(5, {st_mode=0, st_size=0, ...}) = 0
mmap(0, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x40126000
lseek(5, 0, SEEK_CUR) = 0
read(5, root
...
After 12 reads on /etc/shadow, it repeats. I added print statements
to adduser and it looks like it was the get_current_uid subroutine
where the performance problem occurs. It has a loop on getpwent.
Maybe it's the C library that is buggy?
--
Christopher J. Fearnley | Linux/Internet Consulting
[EMAIL PROTECTED] | Design Science Revolutionary
http://www.netaxs.com/~cjf | Explorer in Universe
ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Serious performance bug in Perl
Hi, Originally I thought that it was OK that bug #19085 which I submitted about poor performance in perl was downgraded from important to normal severity because it only affected one application that I wrote. But yesterday I upgraded a bo system to hamm which has a 3000 line /etc/passwd. Now adduser takes OVER ONE MINUTE to find a UID and GID for the new user. And my staff is complaining about the wasted time. I fear that this perl bug is serious. But I haven't heard anyone else complain (I have been remiss in reading the lists lately). Am I the only seeing severe performance degradation with Perl 5.004_04 ? Can anyone else with large password files duplicate the problem? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: intent to take mawk and gawk
'Santiago Vila wrote:' Sorry, I seem to be perpetually several days behind in reading debian-devel. If nobody objects, I intent to take mawk and gawk. I intend to keep these two. [ There have been no maintainer uploads since March 1997, is one year enough? ]. Sorry. I promise to learn PGP this week (I'm giving a talk to the Philadelphia Area Linux Users Group on PGP so I better learn it this week!). Then I'll be able to maintain mawk and gawk. Debian's failure to keep backward development compatibility with bo and adding the requirement for PGP (which I never needed before) plus my growing involvement in small start-up businesses (mostly Debian Linux based) have consumed all my time. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: first proposal for a new maintainer policy
'Manoj Srivastava wrote:' Well, I think if one is not constrained to follow policy, nor required to do so, I see no reason to actually follow policy. Why is it so bad to require policy to be followed? How would you enforce it? Why require something which your police force cannot enforce? I hope you don't wish to flog or flame violators? Since we are all conscientious people here, it seems that we would be better off using bugs and policy as a means of _persuading_ others to follow us. Not as something required. I think of policy and bugs as a cultural procedure to more formally advise developers and users on the collective wisdom of the Project. I agree that developing impediments to bad packages is important. But I don't see any value in trying to enforce those impediments. Humorous note: Whenever I get called to jury duty, I tell the judge why of course, I will take your pronouncements and the entire history of law under advisement in rendering my decision. For some odd reason they usually dismiss me at this point. I guess the judicial system is not as open-minded as I am :) -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Rescue disk crashes with mem=128M or mem=32M or mem=64M
Hi, Fortunately I got you mail an hour before the install and I was able to try your ideas! Copying ldlinux.sys resulted in not a boot device (or somesuch) error. Running the latest syslinux on /dev/fd0 resolved the problem. Thanks! On Fri, Apr 24, 1998 at 05:38:21PM +0100, Enrique Zanardi wrote: On Fri, Apr 24, 1998 at 11:53:19AM -0400, Chris Fearnley wrote: Hi, The April 11th boot disk crashes when I enter the kernel command line option mem=128M (I also tried mem=32M and mem=64M -- all crashed in the same way (see below)). This system has 128M of RAM, but the BIOS only reports 16M to Linux. When I try to correction this on the boot disk command-line, I get the following crash: linux mem=128M (manually entered at boot disk prompt) Loading root.bin... Loading Linux... (I didn't record number of dots here or above) Decompressing Linux... (exactly 3 dots) crc error -- System halted Using the March boot disks this procedure works (boy am I glad I kept a copy of those March boot disks!!!). And using lilo (with the same kernel as is on the boot disk), the mem=128M line works (I even built a kernel and ran lots of memtests to verify that the memory works fine. It does.) Conclusion: the April boot disks are broken wrt the mem kernel command line option. That may be syslinux fault. Can you copy the file ldlinux.sys from March rescue disk to April rescue disk and try again? Or even better, install syslinux_1.37 (now in incoming), put the April rescue disk in the floppy disk drive, run syslinux /dev/fd0 (or fd1) and try installing with that disk, to see if the bug has been fixed. -- Christopher J. Fearnley | Internet21 Network Engineering [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Rescue disk crashes with mem=128M or mem=32M or mem=64M
Hi, The April 11th boot disk crashes when I enter the kernel command line option mem=128M (I also tried mem=32M and mem=64M -- all crashed in the same way (see below)). This system has 128M of RAM, but the BIOS only reports 16M to Linux. When I try to correction this on the boot disk command-line, I get the following crash: linux mem=128M (manually entered at boot disk prompt) Loading root.bin... Loading Linux... (I didn't record number of dots here or above) Decompressing Linux... (exactly 3 dots) crc error -- System halted Using the March boot disks this procedure works (boy am I glad I kept a copy of those March boot disks!!!). And using lilo (with the same kernel as is on the boot disk), the mem=128M line works (I even built a kernel and ran lots of memtests to verify that the memory works fine. It does.) Conclusion: the April boot disks are broken wrt the mem kernel command line option. -- Christopher J. Fearnley | Internet21 Network Engineering [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
April 11th bootdisks - major failure
Hi, Hoo boy, I've been having bad luck with the bootdisks. This time, After I hit [ENTER] to load the ramdisks, the system hangs with the floppy drive light lit after displaying Loading root.bin... Note the three dots. I take the same boot disk and it works like a charm on another system. The system that failed worked with the older February (or was it March) boot disks. So I have hamm running on that system now and can get more information (it is on the Internet). It is an IBM BIOS. Here is some likely helpful info from /proc and so forth: $ cat /proc/pci PCI devices found: Bus 0, device 14, function 0: Bridge: IBM Unknown device (rev 1). Vendor id=1014. Device id=4. Slow devsel. Master Capable. No bursts. I/O at 0x30. Bus 0, device 13, function 0: MicroChannel bridge: IBM Unknown device (rev 2). Vendor id=1014. Device id=2. Slow devsel. Fast back-to-back capable. Master Capable. No bursts. I/O at 0x58. I/O at 0x50. I/O at 0x0. Bus 0, device 12, function 0: Bridge: IBM Unknown device (rev 1). Vendor id=1014. Device id=1a. Slow devsel. Fast back-to-back capable. Bus 1, device 6, function 0: Ethernet controller: DEC DC21041 (rev 33). Medium devsel. Fast back-to-back capable. IRQ 11. Master Capable. Latency=100. I/O at 0x6100. Non-prefetchable 32 bit memory at 0xbf80. Bus 0, device 11, function 0: PCI bridge: DEC DC21050 (rev 2). Medium devsel. Fast back-to-back capable. Master Capable. Latency=96. Min Gnt=3. Non-prefetchable 32 bit memory at 0x60010100. Non-prefetchable 32 bit memory at 0x22806060. Non-prefetchable 32 bit memory at 0xbf80be00. Non-prefetchable 32 bit memory at 0xbdf0bd80. Bus 0, device 4, function 0: SCSI storage controller: Adaptec AIC-7871 (rev 3). Medium devsel. Fast back-to-back capable. IRQ 10. Master Capable. Latency=100. Min Gnt=8.Max Lat=8. I/O at 0x500. Non-prefetchable 32 bit memory at 0xb000. Bus 0, device 0, function 7: Unknown class: Unknown vendor Unknown device (rev 255). Vendor id=118c. Device id=0. Fast back-to-back capable. BIST capable. IRQ 255. Master Capable. Latency=255. Min Gnt=255.Max Lat=255. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. Bus 0, device 0, function 6: Unknown class: Unknown vendor Unknown device (rev 255). Vendor id=118c. Device id=0. Fast back-to-back capable. BIST capable. IRQ 255. Master Capable. Latency=255. Min Gnt=255.Max Lat=255. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. Bus 0, device 0, function 5: Unknown class: Unknown vendor Unknown device (rev 255). Vendor id=118c. Device id=0. Fast back-to-back capable. BIST capable. IRQ 255. Master Capable. Latency=255. Min Gnt=255.Max Lat=255. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. Bus 0, device 0, function 4: Unknown class: Unknown vendor Unknown device (rev 255). Vendor id=118c. Device id=0. Fast back-to-back capable. BIST capable. IRQ 255. Master Capable. Latency=255. Min Gnt=255.Max Lat=255. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. Bus 0, device 0, function 3: Unknown class: Unknown vendor Unknown device (rev 255). Vendor id=118c. Device id=0. Fast back-to-back capable. BIST capable. IRQ 255. Master Capable. Latency=255. Min Gnt=255.Max Lat=255. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. Bus 0, device 0, function 2: Unknown class: Unknown vendor Unknown device (rev 255). Vendor id=118c. Device id=0. Fast back-to-back capable. BIST capable. IRQ 255. Master Capable. Latency=255. Min Gnt=255.Max Lat=255. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. I/O at 0xfffc. warning: page-size limit reached! $ cat /proc/cpuinfo processor : 0 cpu : 586 model : Pentium 75+ vendor_id : GenuineIntel stepping: 5 fdiv_bug: no hlt_bug : no f00f_bug: yes fpu : yes fpu_exception : yes cpuid : yes wp : yes flags : fpu vme de pse tsc msr mce cx8 bogomips: 39.83 $ dmesg Console: 16 point font, 400 scans Console: colour VGA+ 80x25, 1 virtual console (max 63) pcibios_init : BIOS32 Service
Re: April 11th bootdisks - major failure
On Sat, Apr 18, 1998 at 11:02:51PM -0400, Gregory S. Stark wrote: Chris Fearnley [EMAIL PROTECTED] writes: Hoo boy, I've been having bad luck with the bootdisks. This time, After I hit [ENTER] to load the ramdisks, the system hangs with the floppy drive light lit after displaying Loading root.bin... Note the three dots. I take the same boot disk and it works like a charm on another system. I don't recall having to press enter to continue, i guess you're using the 1.2m disks. Regardless, try another floppy. The instructions talk this, the BIOS disk reading code is a lot less robust than the normal drivers and often has problems reading marginal disks. I guess my report was incomplete. I hit [ENTER] when it asks for kernel command-line options. Then it trys loading ramdisks. That is when the system hung. I tried three floppies. One was bad. Two gave the effect indicated. And the floppy works on another machine. They were 1.44M floppies. -- Christopher J. Fearnley | Internet21 Network Engineering [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Making the libc5-libc6 upgrade to be safe (was: netstd...)
'Santiago Vila wrote:' Please, tell me how much harm does to add a Pre-Depends field on libc6, ncurses3.4 and libreadlineg2 for netstd. I can tell you how much inconvenience does *not* to add it and then we can make a comparison between those two inconveniences. Too much, IMHO. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Upgrading from bo to hamm
'Guy Maor wrote:' LeRoy D. Cressy [EMAIL PROTECTED] writes: I question the purpose of leaving broken symbolic links when upgrading the libraries. For instance libreadline2 leaves the following broken links reported by ldconfig: Those symlinks are part of libreadline2-dev. If you upgrade to libreadline2-altdev, then the links will be fixed. But if I decide not to install libreadline2-altdev am I expected to have ldconfig generate a few pages of error messages henceforth? Why shouldn't the system be kept in a sane state at all times (especially since the -altdev packages are not required)? It is mighty disconcerting for newbie admins to see so many error messages. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Uploaded mc-4.1.28-1 (source i386) to master
'Paul Seelig wrote:' -BEGIN PGP SIGNED MESSAGE- Format: 1.5 Date: Mon, 06 Apr 1998 17:23:15 +0100 Source: mc Binary: mc Architecture: source i386 Version: 4.1.28-1 Distribution: frozen unstable Urgency: low Maintainer: Paul Seelig [EMAIL PROTECTED] Description: mc - Midnight Commander - A powerful file manager. Changes: mc (4.1.28-1) frozen unstable; urgency=low . * New upstream release. No new upstream releases into frozen please! Isn't this policy? (I guess we should allow exceptions in the case of important bugs, but no such issue is mentioned here.) -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naïve -- Bucky Fuller -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: need libc5 non-maintainer upgrade
'Christian Schwarz wrote:' On Fri, 2 Jan 1998, Chris Fearnley wrote: '[EMAIL PROTECTED] wrote:' Actually, I'm not sure there is a problem with libc5-altdev. There definitely is a dependency clash between libc5 and libc6, which David Engel thinks we should patch by producing an upgrade for libc5. This will have to be installed before hamm. It's not yet clear to me that we can make this automatic with pre-depends or not. I don't think we should make it automatic. But it should be documented in the libc5 - libc6 transition FAQ. If it is possible, we should make it automatic, since most users don't read any documentation before trying it first. And try and error really fails here... In general, you would be correct. But pre-depends (or even depends) in libc6 on libc5 (which would be required to make it automatic) would force everyone who intalls hamm to also install libc5 in perpetuity. Which IMHO is equally bad. So I suggest the tradeoff: document the transition for those who want libc5 development support. But don't require it. Unless someone sees a clean solution? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: need libc5 non-maintainer upgrade
'[EMAIL PROTECTED] wrote:' Actually, I'm not sure there is a problem with libc5-altdev. There definitely is a dependency clash between libc5 and libc6, which David Engel thinks we should patch by producing an upgrade for libc5. This will have to be installed before hamm. It's not yet clear to me that we can make this automatic with pre-depends or not. I don't think we should make it automatic. But it should be documented in the libc5 - libc6 transition FAQ. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: revised proposed solution (was Re: Bug#15859: libc6 in stable is horribly broken)
'Remco Blaakmeer wrote:' On Sat, 13 Dec 1997, David Engel wrote: Definitely not! libc5-dev implies that libc5 is the default compilation environment installed in /usr/include. Sorry, I must have been half asleep when I wrote the above. libc5-altdev doesn't have to conflict with either libc6-dev or libc5-dev because it is designed to live together with them. But if a non-conflicting libc6-dev and libc5-dev were installed, which would be the default? If I would write a simple 'hello world' program and type $ gcc hello.c -o hello then which libc would 'hello' be compiled against? Since libc6-dev conflicts (and provides) libc-dev, you won't be able to install libc6-dev (until you make the commitment to install the altdev packages). -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: I'll take radiusd-merit
'Roberto Lumbreras wrote:' Hi! I'd like to maintain package radiusd-merit (orphaned in 1.61 version of prospective-packages), if nobody is working on it yet. Excellent. Maybe you can find the buffer overflow when shadow support is included? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: revised proposed solution (was Re: Bug#15859: libc6 in stable is horribly broken)
'Scott K. Ellis wrote:' On Fri, 12 Dec 1997, Chris Fearnley wrote: Why can't we do the following: In both bo-updates and hamm: libc5: No conflicts, no depends (predepends on ldso, of course) (solves the problem of not being able to upgrade easily) [...] This still forces people installing libc6 to upgrade libc5 past a version that can be used with libc5-dev. This is the problem I'm arguing against right now. Doesn't the quoted suggestion (remove the depends: libc6 dependency) resolve that problem?? I think the only problem is finding someone to compile another release of libc5 for hamm. Or could we use an editor to delete the depndency line and bump the version number (regardless of what it would do to the .diff and source)? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Bug#15859: libc6 in stable is horribly broken
'Martin Mitchell wrote:' Chris Fearnley [EMAIL PROTECTED] writes: Is breaking easy upgradeability really better than corrupting utmp? Yes, it means the system should work properly at all stages of the upgrade. Still, the fact that libc5-5.4.33-7 conflicts with libc5-dev means that I have to break C, tcl, ncurses, and slang development (on my system) in order to install just the libc6 library. So libc5 in bo and hamm need to be fixed. Agreed? In my case my first phase of upgrading is to install mutt only (needs libc6) and leave the development environment the same as bo. I can afford to break utmp, but not libc5-dev. But I think my other post specifies a solution (a solution for utmp AND development). Has anyone found any problems with it?? (Beside David who correctly points out that conflicting with libc5-dev is completely unnecessary as all libc*-dev packages should provide and conflict with the virtual package libc-dev.) -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
revised proposed solution (was Re: Bug#15859: libc6 in stable is horribly broken)
'Martin Mitchell wrote:' If they want to remain with a libc5 development environment, they have two choices, stay with bo, or use altdev from hamm. You regard utmp corruption as a minor issue, I would not, especially if I expected that staying with mainly bo would give me a stable system. No one is forcing them to do anything, however it is not unreasonable to expect them to upgrade some packages, including replacing -dev with -altdev, if they want to have the benefits of some newer packages. No, I think we can fix the packages to support both utmp compatibility and easier upgradeability. Why can't we do the following: In both bo-updates and hamm: libc5: No conflicts, no depends (predepends on ldso, of course) (solves the problem of not being able to upgrade easily) In hamm: libc6: Conflicts: libc5 (=5.4.23-6) (solves the problem of utmp corruption) Always: libc*-dev: Provides: libc-dev; Conflicts libc-dev I think that these two changes fix the problems. Does anyone disagree? Agree? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: packaging agrep
'Sven Rudolph wrote:' G John Lapeyre [EMAIL PROTECTED] writes: I am planning to package agrep, a grep-like tool that allows to We have it already. I think it comes with glimpse . So it should be split into an extra package ? No. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Bug#15859: libc6 in stable is horribly broken
Moved to debian-devel 'Scott Ellis wrote:' On 13 Dec 1997, Martin Mitchell wrote: Wichert Akkerman [EMAIL PROTECTED] writes: Huh? The upgrade path is quite clear: install a newer libc5 (5.4.33-7) from hamm, then you may install libc6. Maybe we can fix this by making libc6 pre-depend on libc5 (=5.4.33-7)? This would mean we always have a libc5 present which is not very nice, but it would make the upgrade more foolproof. No, I think this would be highly undesirable for those who want to purge libc5 entirely. Libc6 should instead Conflict with libc5 (=5.4.33-6). Now that's about the STUPIDEST suggestion I've heard yet. Are we TRYING to make it impossible to run libc6 stuff on a mostly bo system? The only conflict is the utmp format problem, and that is MINOR to many people. Actually, I think Martin is correct. In order to prevent CDROM based 1.3.1 users from corrupting their utmp, libc6 must conflict with older libc5. Modulo my typo (Martin's = is right, not my ), I think my other post suggests the best solution. Of course, upgrading will need to involve upgrading libc5 before installing libc6 for the first time. But this is acceptable to me. The conflict line tells me to find a newer version. But libc5's conflict with libc6 IS totally broken wrt upgrades (it is both untrue and uninformative). -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Bug#15859: libc6 in stable is horribly broken
'Rob Browning wrote:' Scott Ellis [EMAIL PROTECTED] writes: If you don't upgrade anything that deals with utmp to libc6, you don't have any problems). The problem is that maybe *you* know what packages those are, but most users expect to be able to upgrade without major system services breaking if dpkg/dselect doesn't indicate that there's a problem. Your approach would cause silent failures. The current approach causes silent failures. Anyone who purchased an official Debian CD will corrupt their utmp by upgrading to libc6. Has everyone looked at my proposed solution carefully? Does it really help? -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Libc6 progress: 1997-12-06
'Hamish Moffatt wrote:' Chris Fearnley [EMAIL PROTECTED]: dome-4.60-1 Compiled fine but appears to segfault on execution. Hmm, are there problems with g++? I'll be upgrading to hamm RSN and hope to have time before the code freeze to deal with this ... -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Hamm: Exim + Chos standard?
'Tim Cutts wrote:' On 14 Jun 1997, John Goerzen wrote: Christoph Lameter [EMAIL PROTECTED] writes: It might be good if we would replace smail in hamm with exim. Exim should be the standard mailer for hamm: Exim doesn't provide UUCP capabilities *at all*, thus it is rather useless for sites that use UUCP (like me). Right now, I am using sendmail. (What, BTW, is the reason for not using sendmail?) Well, for one thing exim (and smail) are a hell of a lot easier to configure than sendmail. That was what originally moved me towards exim at work - I really didn't want to muck about fixing numerous broken sendmail setups when in far less time I could just switch all the machines Could someone explain why people have so much trouble with sendmail setups? I've used sendmail commercially for several years now and the only misconfigurations I've ever seen are with the Cw directive (something which would need to be specified to any MTA). Virtual domain mail is tricky, but there are several web pages with instructions on configuring this and it is not difficult if you can follow instructions (even easier with 8.5). Virtual domains are the only reason to mess with rule sets and I freely admit to not having a clue about rule sets. Still I've never had problems with sendmail because all the rule sets one needs are included in sample sendmail.cf files. Copy, paste, tweak -- it works! /usr/sbin/sendmailconfig made my life even easier (though as I said I never had trouble with sendmail even though I never studied its config file). OK, I admit to reading through the well commented sendmail.cf and following the instructions in the comments for several parameters, but this is not too difficult. If I might speculate on my winning sendmail configuration strategy: ignore the irrelevant (like rule sets). The answer to all sendmail problems is with the easily configured parts in /etc/mail. grepping through /usr/doc/sendmail also helps. Hence I would appreciate it if the MTA debate could focus on design criteria other than ease of configuration. I'm more interested in performance and design considerations that impact on security and the ability to configure (flexibility). Chris Just flabbergasted that anyone finds sendmail troublesome Fearnley -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Env-varaibles
'=?iso-8859-1?Q?=22Ole?= =?iso-8859-1?Q?J=F8rgen?= =?iso-8859-1?Q?Tetlie=22?= wrote:' Hello, for SmallEiffel (which I am packaging) to work at all, it needs an env-variable to be set. Should it be set with a preinst-script? I wouldn't like that to happen to my system, but I don't see any other way, if it should be set at all. Should I just put a prominent note in /usr/doc/smalleiffel/README.debian saying that this variable must be set to use the package? I'm sure others have had the same problem! What's the standard way? The upstream is broken. It should choose a reasonable default for the Env variable or notice its absence and advise the user on how to proceed. I vote for hacking the source and providing patches upstream. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Humore (no, I'm not serious here (was Re: a.s.r manpages)
'J.H.M.Dassen wrote:' On May 26, Manoj Srivastava wrote Would the doc directory be better for man pages? Why games? Check out the manpages at http://www.bofh.net/man/: lart - Luser Attitude Readjustment Tool - use a lart to adjust lusers' attitudes sysadmin - responsible for everything imaginable that may or may not have to do with the system you're using. Contraction of system and administrator Clearly these belong in section admin, priority important ;) -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: ssh and default home directory permissions (revisited?)
'Carey Evans wrote:' I've removed group write permissions from my home dir because of the programs like qmail and ssh which don't like it. I don't think anything would break because of removing these permissions, so maybe adduser should make home directories mode 755 (or 750)? Or 751. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: rm -r * and the default prompt
'Raul Miller wrote:' '=?iso-8859-1?Q?Nicol=E1s_Lichtmaier?= wrote:' So I say: PS1=[\\u] \\h:\\w\\$ =D On May 21, Chris Fearnley wrote No, PS1='[EMAIL PROTECTED]:\w\$ ' ! I'd prefer PS1='\$ ' However, if you want all that fanciness, a compromise is: PS1='[EMAIL PROTECTED] \W\$ ' How about: PS1='\[\033[40;31m\]pwd: \[\033[40;33m\]\w \[\033[40;[EMAIL PROTECTED] ' I'll repeat my conclusion: leave it as PS1=\\$ and provide a customization app for sysadmins to edit /etc/profile and another for users to edit ~/.bash_profile (and other dotfiles too, of course) and keep the installation's paws away from customizations that EVERYBODY has a different opinion about. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: crons scripts should report status info in the mail
'Karl M. Hegbloom wrote:'
Since the output from cron jobs is mailed anyhow, as it should be, I
think that all cron scripts should report in as they are run, and that
this should be made a standard. Here's why.
But if they complete successfully they should be quiet. Maybe this
would work:
set -e
function onerror { if [ $? -ne 0 ]; then echo error in script $0; fi; }
trap onerror 0
SCRIPT HERE
--
Christopher J. Fearnley | Linux/Internet Consulting
[EMAIL PROTECTED] | Design Science Revolutionary
http://www.netaxs.com/~cjf | Explorer in Universe
ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: dpkg verify mode for security?
'Amos Shapira wrote:' I was asking over Linux-ISP about doing cleanup after breakins and got many use tripwire answers, and one which says that RPM has a verify mode which checks for files which were changed since they were installed. Can the dpkg maintainers consider adding such a feature for Debian? What does the rpm verify give you? As far as I can tell it gives a false sense of security. Nothing more. The rpm database is easily hacked once root access is attained. Tripwire or something similar is the only viable option. -- Christopher J. Fearnley | Linux/Internet Consulting [EMAIL PROTECTED] | Design Science Revolutionary http://www.netaxs.com/~cjf | Explorer in Universe ftp://ftp.netaxs.com/people/cjf | Dare to be Naive -- Bucky Fuller -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Shared libraries in packages. How?
'Yves Arrouye wrote:' Hi, I have recently made the package compface (can be found in unstabe) containg the shared library libcompface. Now I am trying to make a package xfaces that uses this library. But I can't get dpkg-shlibdeps working. This is what is says: # dpkg-shlibdeps xfaces dpkg-shlibdeps: warning: unknown output from ldd on `xfaces': ` libcompface.so = /usr/lib/libcompface.so' dpkg-shlibdeps: warning: unable to find dependency information for shared library libXpm (soname 4, path /usr/X11R6/lib/libXpm.so.4.7, dependency field Depends) dpkg-shlibdeps: warning: unable to find dependency information for shared library libm (soname 5, path /lib/libm.so.5.0.5, dependency field Depends) What's wrong with my compafe package, and to which packages does the other libs belong, and why can't dpkg-shlibdeps find them? You forgot to install the shlibs file into debian/tmp/DEBIAN. It's all in the programmers (or policy) manual. I forgot my first time too! -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Bug#4644: dpkg --compare-versions reality and documentation differ
Package: dpkg Version: 1.4.0 dpkg -h says: Comparison operators for --compare-versions are: lt le eq ne ge gt (treat no version as earlier than any version); lt-nl le-nl ge-nl gt-nl (tread no version as later than any version); = = = (only for compatibility with control file syntax). BTW, that's a typo 'tread' should be 'treat', I think Yet: $ dpkg --compare-versions 1.2.3 '' 4.5.6 dpkg: --cmpversions bad relation Type dpkg --help for help about installing and deinstalling packages [*]; Use dselect for user-friendly package management; Type dpkg -Dhelp for a list of dpkg debug flag values; Type dpkg --force-help for a list of forcing options; Type dpkg-deb --help for help about manipulating *.deb files; Type dpkg --licence for copyright licence and lack of warranty (GNU GPL) [*]. Options marked [*] produce a lot of output - pipe it through `less' or `more' ! Surprisingly, '' works. $ dpkg --compare-versions 5.2.3 '=' 4.5.6 dpkg: --cmpversions bad relation Type dpkg --help for help about installing and deinstalling packages [*]; Use dselect for user-friendly package management; Type dpkg -Dhelp for a list of dpkg debug flag values; Type dpkg --force-help for a list of forcing options; Type dpkg-deb --help for help about manipulating *.deb files; Type dpkg --licence for copyright licence and lack of warranty (GNU GPL) [*]. Options marked [*] produce a lot of output - pipe it through `less' or `more' ! Surprisingly '=' works. $ dpkg --compare-versions 5.2.3 '=' 4.5.6 dpkg: --cmpversions bad relation Type dpkg --help for help about installing and deinstalling packages [*]; Use dselect for user-friendly package management; Type dpkg -Dhelp for a list of dpkg debug flag values; Type dpkg --force-help for a list of forcing options; Type dpkg-deb --help for help about manipulating *.deb files; Type dpkg --licence for copyright licence and lack of warranty (GNU GPL) [*]. Options marked [*] produce a lot of output - pipe it through `less' or `more' ! $ dpkg --compare-versions 5.2.3 'lt-nl' 4.5.6 dpkg: --cmpversions bad relation Type dpkg --help for help about installing and deinstalling packages [*]; Use dselect for user-friendly package management; Type dpkg -Dhelp for a list of dpkg debug flag values; Type dpkg --force-help for a list of forcing options; Type dpkg-deb --help for help about manipulating *.deb files; Type dpkg --licence for copyright licence and lack of warranty (GNU GPL) [*]. Options marked [*] produce a lot of output - pipe it through `less' or `more' ! $ dpkg --compare-versions 5.2.3 'ge-nl' 4.5.6 dpkg: --cmpversions bad relation Type dpkg --help for help about installing and deinstalling packages [*]; Use dselect for user-friendly package management; Type dpkg -Dhelp for a list of dpkg debug flag values; Type dpkg --force-help for a list of forcing options; Type dpkg-deb --help for help about manipulating *.deb files; Type dpkg --licence for copyright licence and lack of warranty (GNU GPL) [*]. Options marked [*] produce a lot of output - pipe it through `less' or `more' ! $ dpkg --compare-versions 5.2.3 'gt-nl' 4.5.6 dpkg: --cmpversions bad relation Type dpkg --help for help about installing and deinstalling packages [*]; Use dselect for user-friendly package management; Type dpkg -Dhelp for a list of dpkg debug flag values; Type dpkg --force-help for a list of forcing options; Type dpkg-deb --help for help about manipulating *.deb files; Type dpkg --licence for copyright licence and lack of warranty (GNU GPL) [*]. Options marked [*] produce a lot of output - pipe it through `less' or `more' ! -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Bug#4627: gawk clean target prevents build
'llucius wrote:' Package: gawk Version: 3.0.0-4 1) debian/rules clean target does not ignore errors when doing the make distclean which causes the build to fail since there's nothing to clean (not even a Makefile). Would you believe I ran into this problem when making my last build. So I just ran debian/rules build. I'm not sure: don't the policy or programmer manual say that the rules file should fail on errors? I suppose could make it more robust. But I probably won't release a new version (nor close the bug) until I can do something more substantial too. Boy you must check Incoming daily :) -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Problems with dselect...
'Michael Dillon wrote:' -- Forwarded message -- Date: Wed, 25 Sep 1996 00:50:32 -0400 From: Jon Lewis [EMAIL PROTECTED] [Clipped generic complaint about dselect's user unfriendlyness.] Wow one of my USENET heroes forwards a mail from another of my USENET heroes about my favorite Linux distribution. I simply must respond :) It is true that deselect is a bit quirky. And the last stable release of Debian had at least one fatal bug (if you select to install cflow, all dies due to a bug in the cflow package). However dselect is under continual improvement and is better each time I've tried it (you might consider downloading the latest version (ftp://ftp.debian.org/debian/unstable/binary/base/dpkg*) and installing it with dpkg -i before running dselect). The on-line documentation explains enough about the system to get by. But my approach is to install whatever dselect chooses (adding a few add-ons maybe -- I've found that trying to add everything can cause problems). Then switch to the low-level command line tool, dpkg. I feel Debian's chief advantages are in its integration between packages, the community development model and its upgradeability. For these reasons I use Debian on most of my ISP clients' servers (and I like it much better than the one client who is sticking with Red Hat). Certainly there is room for improvement in dselect. And the author of dselect reports that it will be easy to develop an alternative user interface to the low-level package management system. If only I had time! In sum, for servers, I really like Debian. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4601: Several small problems in Apache
Package: apache Version: 1.1.1-3 Many of the modules offered are called *required* which didn't seem to me to be required. Sorry, I forget which ones exactly. But I chose not to load common_log_module and got a syntax error due to the TransferLog directive. I also got a syntax error when #LoadModule proxy_module /usr/lib/apache/modules/mod_proxy.so was included. I had to comment it out. Perhaps it depends on another module that I hadn't loaded? Very nice configuration .. just a little incomplete. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4567: gpm upgrade from 1.06-3 to 1.10-1 is not smooth
Package: gpm Version: 1.10-1 I tried installing libgpm1_1.10-1.deb and gpm_1.10-1.deb in that order. The library installed just perfectly, but when gpm's prerm tried to /etc/init.d/gpm stop ... it hung. The init.d script calls gpm -k which may have gotten confused by the new library?? I was able to kill the process manually and re-install. But I couldn't find a way for dpkg to resolve the problem. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Section: and Priority: in shared libs: Questions and PROPOSAL
Greetings, I've prepared two binary packages from one Source: slang package: slang0.99.34 - A C programming library for user interfaces - shared library slang0.99.34-dev - A C programming library for user interfaces - development kit I wanted to set the following Section and Priority: Package Section Priority slang0.99.34devel standard slang0.99.34-devdevel optional There are two main issues to be addressed: First, how to get the correct information into the control and .changes file. I have been unsuccessful in accomplishing this. In the debian/control file, empirical research shows that one can only put Section and Priority fields in the Source portion of the file and hence only identical values can be suggested here for each binary package. I have tried using these two commands in various places in debian/rules to no avail: dpkg-distaddfile slang0.99.34-dev_0.99.34-1_i386.deb devel optional dpkg-genchanges -DPriority=optional Anyone know the answer? Can we document it in the programmer's manual (or maybe it's there and I missed it)? Secondly, my decision to put the shared object in standard and the development kit in optional is /not/ the way tcl74, tcl75, and probably other do it. So I'm really making the suggestion that all shared libraries should go into standard so that most all systems would get these packages installed. But the -dev packages should only be installed by those who go out of their way in dselect to choose such optional packages. For ease of reference here is what the policy manual says: standard These packages provide a reasonably small but not too limited character-mode system. This is what will install by default if the user doesn't select anything else. It doesn't include many large applications, but it does include Emacs (this is more of a piece of infrastructure than an application) and a reasonable subset of TeX and LaTeX (if this is possible without X). optional[4] This is all the software that you might reasonably want to install if you didn't know what it was or don't have specialised requirements. This is a much larger system and includes X11, a full TeX distribution, and lots of applications. Finally, in the debian/shlibs file should I include information about historical versions of the current package? In my case should I add the line: libslang 0 slang-lib to this file, so that the old (inflexibly packaged) version of this shared lib is recorded for possible use later? At first I thought, of course. But now I'm leaning toward only including the current packages information. Because I can't see any reason to maintain that level of support for older versions of the shared libs. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Re: Bug#4550: Build of ae fails since it's statically linked
'Dale Scheetz wrote:' On Sun, 22 Sep 1996, Christoph Lameter wrote: The N option is used to statically link a program. What manpage were you looking at? The man page is for ld. The gcc man page says ld is used to link. The gcc man page also says that -static is the proper option for creating static linked executables. BTW, what architecture are we talking about? It isn't relevant. I strongly remember that -N is for static linking of a.out binaries. The old policy manual said you should only use -N on binaries that stay in core for /very/ small lengths of time as they use memory poorly. It shouldn't be used (certainly NOT for an editor). It probably should never have been used. It probably works under ELF because under ELF -N should be ignored. There may be a bug in m68k libc or ld that doesn't ignore it, but that doesn't lessen the bug of using -N in the first place. Unless you can document some strong reason why you /need/ -N, remove the flag. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4561: dpkg-genchanges puts bad Maintainer: in changes file
'Yves Arrouye wrote:' Package: dpkg-dev Version: 1.4.0 It appears that the Maintainer: field in control is ignored, so the Maintainer: field in the changes file is made with the name of the user calling dpkg-genchanges and the name of the host the file is built on. This is a problem, and in any case, the generated email address should use /etc/mailname (my host name, for example, cannot get mail because his real name and email names are different, even in different domains). Hmm, something else must be going on. The mawk package I built and uploaded didn't use the hostname either (which has been misconfigured on my system as syntropy.OnIt.net). I think the solution is in dpkg-genchanges(1): -Dfield=value Override or add an output control file field. This option is understood by dpkg-source, dpkg- gencontrol and dpkg-genchanges. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4545: dpkg-gencontrol gets it wrong with certain package names
Package: dpkg-dev
Version: 1.4.0
I believe this bug will also affect ncurses3.0 when it gets packaged
with the new packaging standard (perhaps all or most shared lib
packages).
Here is the beginning of my changlog file for slang:
slang (slang-0.99.34-1) unstable; urgency=low
* upgrade to latest upstream source
I also tried several variations on the naming theme. Did I confuse myself
or does dpkg-parsechangelog not support our shared lib naming scheme yet?
I've just rescanned the policy and programmer's manual and I think at
minimum we have a documentation error.
$ dpkg-gencontrol -Pdebian/tmp-lib -pslang0.99.34
parsechangelog/debian: error: unrecognised line, at changelog line 3
dpkg-gencontrol: error: syntax error in parsed version of changelog at line 0:
empty file
I think my formatting is correct. I /may/ have tracked the problem
down to /usr/lib/dpkg/parsechangelog/debian whose regular expression on
line 54:
if (m/^(\w[-+0-9a-z.]+) \(([^\(\) \t]+)\)((\s+[-0-9a-z]+)+)\;/i) {
fails to match my slang version number. [I doubt this analysis, but won't
have time to further investigate until much later today.]
Here is my debian/control (abbreviated descriptions):
Source: slang
Section: devel
Priority: optional
Maintainer: Chris Fearnley [EMAIL PROTECTED]
Standards-Version: 2.1.1.0
Package: slang0.99.34
Architecture: any
Depends: ${shlibs:Depends}
Description: A C programming library for user interfaces - shared library
S-Lang is a C programmer's library that includes routines for the
rapid development of sophisticated, user friendly, multi-platform
applications.
Package: slang0.99.34-dev
Architecture: any
Provides: slang-dev
Replaces: slang-devel
Conflicts: slang-devel, slang-dev
Depends: ${shlibs:Depends}
Description: A C programming library for user interfaces - development kit
S-Lang is a C programmer's library that includes routines for the rapid
development of sophisticated, user friendly, multi-platform applications.
--
Christopher J. Fearnley|Linux/Internet Consulting
[EMAIL PROTECTED] |UNIX SIG Leader at PACS
http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society)
ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary
Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4526: popclient now requires a ~/.poprc :(
Package: popclient Version: 3.05-1 $ popclient mail.host /s2/redhat/home/home.cjf//.poprc: No such file or directory This used to work. I'm disappointed that the new version isn't backward compatible. Since I'm resposible for maybe 10 pop servers on 5 different networks, I prefer to just put the applicable host on the command line and get prompted for the passwd --- just like our old popclient did. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
scp method for dupload?
I was wondering if there is a way to use scp to transfer files to master? This would let me transfer the files without sending the passwd in the clear. I know master supports ssh, but I'm not sure of the procedure. Probably before anyone informs me, I'll have put a new version of mawk into Incoming. But for everyone's sake, I bring it up anyway! -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4449: dselect feedback missing
'Barak Pearlmutter wrote:' Package: dpkg Version: 1.2.14elf When you go through lots of dselect work editing the package selection menu, then you're done (whew!) and you hit the big INSTALL button ... before dselect goes ahead and installs stuff, it should give you a very short description of what it's about to do, and give you a chance to back out if you want. Eg: Wow! a user who can give insightful suggestions on how to improve dselect. I wish I had the presence of mind to translate my frustrations in using dselect into useful bug reports. Thank you very much!!! -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4433: metamail has confounding postinst
Package: metamail Version: 2.7-8 I'm going to have to read the postinst script to determine what to say to this one: # dpkg --configure metamail Setting up metamail (2.7-8) ... New action 'view' for MIME type 'image/*'... -- package=metamailview=showpicture -viewer xloadimage -view -quiet %s 1) package=metamailview=showpicture -viewer xv %s Place at what priority? (1-2) -- -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Re: Do we ever retire packages?
'Michael Meskes wrote:' [EMAIL PROTECTED] writes: I have argued before that a2ps and a2gs are effectively replaced by genscript, and that we should remove them. I think a similar case could be made for xosview as we now have procmeter. Opinions? Remove them. Move them to project/obsolete or some such. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Re: CC's on this mailing list
'Lars Wirzenius wrote:' Spam does make furious, extra Cc's from mailing lists don't. They just annoy me (see signature), and in theory they do cost me a bit. Not enough to make me worry about it, but enough to write kilobyte after kilobyte about it. I do wish that people wouldn't Cc me when I read the mailing list. I feel that it is good netiquette not to do that. I like the CCs because lately I haven't been able to keep up with the list and knowing that someone responded to something (which shows up in my mailbox and not the procmail filter) is useful. This may be a mark that we or I don't think there is enough real problems with Debian. Hm, perhaps the three month release schedule is one: we have one month left for the next release, and we're just about to make a big change in source packaging? Perhaps it would be better to wait an extra month for this occasion only? I don't think we have a good mechanism for code cleanups before a major release. Instead, I think that sticking with stable releases from upstream is the way to keep Debian stable. I like the three month schedule, but feel it is mostly independent of maintainer issues. Though perhaps we should save the source package changes for the release after next month's? -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4103: slrn epends on unavailable slang-lib09931
'Maarten Boekhold wrote:' On Sun, 11 Aug 1996, Joey Hess wrote: Package: slrn Version: 0.8.8.4-1 This version of slrn appears to depend on a version of slang-lib that's not been packaged yet. slang-lib_0.99.23-1 is the newest version of slang I can find as a .deb on ftp.debian.org. Am I missing something? I only put slang-lib0931.deb at http://www.netaxs.com/~cjf/debian/slang-lib09931-1.deb or ftp://ftp.netaxs.com/people/cjf/debian. I'm hoping to get slang-09933 out, but I've been plagued by hardware problems and too much consulting work. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#4064: sendmail should recommend deliver, not depend
'Miquel van Smoorenburg wrote:' You (Michael Shields) wrote: Package: sendmail Version: 8.7.5-4 sendmail depends on deliver. However, in at least two common configurations -- null client, and delivery by procmail -- it will run perfectly without deliver. sendmail should only recommend it. And as a result you can install sendmail without deliver and your mail installation won't work at all.. Sendmail should come with mail.local instead of deliver IMHO. Sendmail should be configurable, IMHO. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#3952: Less annoyances
'Christoph Lameter wrote:' A wish because of the heavy usage of gzipped files under debian: - Add functionality for less to automatically recognize a gzipped file and view it correctly without having to resort to zless. I have code for that (or you can use most): First set some environment variables in .bash_profile or /etc/profile: export LESSCLOSE=lessclose.sh %s %s LESSOPEN=lessopen.sh %s Then somewhere in $PATH put lessclose.sh and lessopen.sh: $ cat /usr/local/bin/lessclose.sh #!/bin/sh rm $2 $ cat /usr/local/bin/lessopen.sh #!/bin/sh # When the environment variable LESSOPEN is set to refer to this file # (e.g., LESSOPEN=lessopen.sh %s), one can do several interesting # things with pipes (see less(1) for details). Namely, # 1. view any gzipped or compressed file(s) anywhere in the filesystem with #the simple command less filename(s). And you can use less' multiple #file searching and bookmark facilities on the gzipped files! # 2. view a list of compressed and uncompressed files from the same #command prompt (using all of less' multiple file features). # 3. zcat file.gz | less still works. if [ $1 != - ]; then # $1 = - when we are reading standard input gzip -t $1 /dev/null if [ $? -eq 0 ]; then # Is it compressed? zcat $1 /tmp/$$.`basename $1` # Use basename for files in echo /tmp/$$.`basename $1` # other directories fi fi exit 0 -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED], [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Bug#3437: fvwm should not recommend fvwm2
Christian Hudon wrote: Package: fvwm Version: 1.24r-24 Fvwm shouldn't recommend (nor even suggest, IMO) another version of itself (i.e. fvwm2). Hmm, I'm not so sure. In principle you are correct, but the fvwm2 package has all the pixmaps needed by fvwm-1.24 and fvwm won't run without them (at least not with my .fvwmrc). So unless the packages are reworked there may need to be a /dependency/ on fvwm2. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Re: 1.2 source archive and packaging issues
'J.H.M.Dassen wrote:' Bruce wrote: Also, we should think about source packaging again. We are welcome to take anything we want from RPM source packaging, if that would help. RPM has the advantage that it include _pristine_ source (identical (cmp or md5sum-wise) to the upstream sources, which are patched during the build process. IMO this is what we should work towards to. These are a venerable goals. But I like that the Debian source packages can be untarred by anyone without dpkg and/or rpm installed. And if we were to force use of dpkg for installing the source code, I'd like more freedom over which directory/partition the source ends up in than rpm allows. Anyway those are the two advantages of Debian's current source packaging that I hope we don't abandon. -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Re: Organizing non-free
'Brian C. White wrote:' Sorry to bother you again, but I thought non-free was precisely for packages which may not be sold on CDs. Now I am confused. You're not the only one. For example, shareware programs can be sold on CD but require payment for use. I'd be more specific, but I can't get to master at the moment. I'd better send Simon a note... This is one of the things I would like to improve about non-free's organization. It would be a Good Thing (tm) in my opinion to at least differentiate between non-free to distribute and non-free to use. Umm, non-free to distribute shouldn't be on /any/ ftp site, right? -- Christopher J. Fearnley|Linux/Internet Consulting [EMAIL PROTECTED] |UNIX SIG Leader at PACS http://www.netaxs.com/~cjf |(Philadelphia Area Computer Society) ftp://ftp.netaxs.com/people/cjf|Design Science Revolutionary Dare to be Naive -- Bucky Fuller |Explorer in Universe
Thoughts on Replaces field (was Re: binary-alpha and binary-sparc directories)
'Ian Jackson wrote:' I think I'll have to support `Replaces' or something, so that old packages can have all their files `taken away' and disappear eventually. Here's the scenario that I hope a Replaces fiels might resolve. I'm working on the S-lang library. Both most and Midnight Commander depend on S-lang, so building a shared library makes sense (and with ELF it's easy to do!). But S-lang is under development and shared libraries aren't compatible from month-to-month (unless I don't understand some subtlety in ELF - very likely, actually). I released most-4.5.0-1 and slang-lib-0.99.23-1. But now I want to release slang-lib-0.99.24-1, but it's incompatible with slang-lib-0.99.23-1 which it should replace. Of course, most depends on slang-lib (version 0.99.23-1 ONLY). most-4.5.0-1 breaks with slang-lib-0.99.24-1. So I need to build another most with slang-lib-0.99.24-1 support. So I envision this specification for the Replaces: field: GIVEN: A, B, and C all depend on package D AND Package E replaces package D. Then When E is installed, it leaves D alone (since things still depend on D). When A, B, or C are upgraded (to versions compatible with E) they are removed from D's dependancy list. Once D's dependency list is empty, D is removed (making sure that E is intact). I think this would give the flexiblity that the S-lang example needs and that isn't presently present. If I remember correctly Ian didn't like encoding the shared library version in the package name (I'm starting to agree). But right now that is the only proposed solution to the problem that I've seen. OK, now I'm waiting for you all to tell me what details I've omitted ... BTW, I think dpkg blazes - performance-wise (especially compared to rpm which is a dog)! Well done. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) http://www.netaxs.com/~cjf |Design Science Revolutionary ftp://ftp.netaxs.com/people/cjf|Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Re: beware sysvinit 2.58 installation
'Miquel van Smoorenburg wrote:' I've changed the postinst script to create a symbolic link in /var/log, so that it will (hopefully) work in all cases. It is also backwards compatible with other programs (UPS watchdogs etc) this way. If I don't get any replies saying this is a bad idea I'll upload the new version this afternoon (MET). You might consider putting the symbolic link in the main installation rather than the postinst (I may be wrong but having dpkg manage the file seems easier administratively). And I think relative symbolic links are to be preferred too. But no objection (from me :) in any case. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) http://www.netaxs.com/~cjf |Design Science Revolutionary ftp://ftp.netaxs.com/people/cjf|Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
dpkg Replaces: field (was Re: binary-alpha and binary-sparc directories)
'[EMAIL PROTECTED] wrote:' Raul Miller writes: Raul It does look like dvips was superceeded by some other package, and Raul that it did originally have some executables in it. Nils switched to the upstream convention of reflecting the 'k' for Karl Berry's kpathsea in the package name. I had moaned about this weeks ago. You have to manually delete dvips after installing dvipsk, xdvi after xdvik etc. A Conflicts: in debian.control might have helped here. Or a new Replaces: field. Yes, this seems to me a good idea. Conflicts involves too much administrator intervention. Ian, can we have this feature in dpkg? -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) http://www.netaxs.com/~cjf |Design Science Revolutionary ftp://ftp.netaxs.com/people/cjf|Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Re: ALPHA release of apache-1.0.0-1 now available
'Michael Alan Dorman wrote:' On Sun, 17 Dec 1995, Chris Fearnley wrote: This is a preliminary release. It seems to work, but I'm disatisfied with my handling of httpd configuration (basically there is none - you have to edit /etc/httpd/* by hand). Hmm. That's what kept me from releasing mine. Maybe we can decide what constitutes a resonable set of things to do at install time? I like the idea of asking a set of questions in the postinst (similarly to cern-httpd). There are too many options to simply impose on the user some default configuration. Now, that I've got it compiled, I'm reading the User Guide! I plan to design some specifications for the /usr/sbin/httpdconfig script (when I can code it is another story, indeed). Like: * Should we create a new user and/or group to control access to the hierarchy of html files? If so, why don't we make it official and get Bruce to include in the base /etc/group and /etc/passwd files. User nobody and group nogroup is either already in there or is it set up by some other package? I suppose user wwwadmin might be better? * Where should we put the html hierarchy? I mean, they could exist on /usr, since it doesn't per se matter if they're read-only or not, but it that right? A lot of places use /home/html or some such---I personally don't think this is appropriate, but someone please tell me I'm wrong. /usr/lib/apache is my choice for serverroot. Where the documents go is site-specific. I'd like to also include an option to chroot httpd to /usr/local/http or somesuch. Can dpkg install a package under some arbitrary directory? If so then the preinst script might be able to get everything into /usr/local/http and run httpd under chroot (for the security paranoid). There are others, obviously. We might also want to coordinate with Ted Hajek who maintains the cern httpd. I looked at his work and stole many ideas from it. Thanks Ted. Also, you might want to rename the package to apache-httpd and the related directories and files, or you need to conflict with cern-httpd. apache-httpd provides httpd (as does cern-httpd) so dpkg won't install one until the other is removed. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Re: ALPHA release of apache-1.0.0-1 now available
'Michael Alan Dorman wrote:' /usr/lib/apache is my choice for serverroot. Where the documents go is site-specific. I'd like to also include an option to chroot httpd to /usr/local/http or somesuch. Can dpkg install a package under some arbitrary directory? If so then the preinst script might be able to get everything into /usr/local/http and run httpd under chroot (for the security paranoid). Uh, why would you want to chroot the httpd? Wouldn't that cause mondo problems, especially if we try and get it to do stuff like dynaloading modules, etc.? For extra security. Like any chroot environment, you need to copy all the shared libs into $chroot. But if a complete list were determined, it could be done in the postinst. Net Access is currently running apache in a chroot environment for extra security. I think it would be nice to add this feature (My only problem is I'm not sure dpkg can handle it - Ian?). -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Bug#2002: Missing manpages
'Bill Mitchell wrote:' On Mon, 11 Dec 1995, Owen S. Dunn wrote: Package: diff Version: 2.7 Revision: 5 This package provides no man pages for any of diff, diff3, sdiff, or cmp. Too true. It comes with info pages, which are not at all the same thing. I think the current custom is to leave bug reports about this particular issue open, so I'm not closing this one. I doubt, however, that I'll get around to writing a man page and contributing it upstream in hopes that it'll be picked up. On my old Slackware I have man pages for diff, diff3, and cmp. So it's not like someone has to start from scratch. I don't know the copyright on these man pages because slakware preformats everything grumble. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Re: symlink in /usr/include (fwd)
'Ian Murdock wrote:' How about installing the kernel headers directly in /usr/include, rather than linking them into /usr/src? I always assumed this was standard kernel practice. Apparently, I was wrong. Are there any opinions on the subject? The only problem I see would be if I upgrade my kernel from non-debian sources. But then I'd be someone who knows what I'm doing, I suppose. And what about systems with multiple kernel trees in /usr/src -- wouldn't you want each to have it's own /usr/include updated by make before each build? A brief look at the kernel source didn't answer these questions for me. What does our kernel maintainer think. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Re: convenience script for building a.out packages
'Bill Mitchell wrote:' Scott Blachowicz [EMAIL PROTECTED] said: Ian Jackson [EMAIL PROTECTED] wrote: I agree. But, now you see that we have a script called /usr/bin/aout and a potential directory called /usr/bin/aout. Hence my suggestion that it ought to be called something else. with-aout perhaps. [...] Actually, I would argue that the _directory_ be someWHERE else. [...] Good point. From FSSTND 1.2, 3/28/95, section 3.1, para three: There should be no subdirectories within /bin. Given the intention of FSSTND compliance, this is an absolute prohibition. But what does FSSTND say about subdirectories of /usr/bin? I would hate to have the mh utilities stored in with the illions of programs already in /usr/bin. And where would the netpbm tools be stored [I dislike Red Hat's solution of all 100 utilities getting mixed up in /usr/bin]? Anyway, I'd like an option to avoid having 500 programs in /usr/bin. Let modularity ring :) -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
mawk(1) , gawk(1) and package control files (questions)
I have debianized mawk(1), a pattern scanning and text processing
language (it's awk, really). I will upload it as soon as I get ELF
installed. For now, I have a few questions on this my first
debianization effort.
o Why is gawk a required package? And is there any reason why gawk
forcibly installs awk to be a symbolic link to awk? Maybe because my
mawk package hasn't been released yet :) Could gawk | mawk be set up
for the Required priority?
o I'm not clear on the provides virtual package as it would apply to
'awk'. Should gawk provide 'awk'? Should mawk provide 'awk'?
Since clearly awk should be a symbolic link to gawk or mawk but not
both, is there a conflict here? Or should the conflict be resolved
by the {post,pre}{inst,rm} scripts (this seems best to me as most
people would have both installed side by side)? Maybe my mawk package
should Recommends or Suggests gawk?
o It occurs to me that many packages should Recommends or Suggests
man and/or info. Yet they don't. Why? What am I missing? And most
packages depend on ldso. So are essential packages too basic to be
listed in the control files? Is it possible that this lack of
consistency in setting dependencies could affect us much like the
current libc, libc4 and libc5 (which as required packages under Debian
0.93R6 weren't mentioned in the depends fields of any packages).
Thanks for Debian. I'm thoroughly enjoying it! I just wish I could
easily order a CD :(
--
Christopher J. Fearnley|UNIX SIG Leader at PACS
[EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society)
[EMAIL PROTECTED] |Design Science Revolutionary
http://www.netaxs.com/~cjf |Explorer in Universe
Dare to be Naive -- Bucky Fuller |Linux Advocate
Re: mawk(1) , gawk(1) and package control files (questions)
'Bill Mitchell wrote:'
The following are my best guesses at answers. If I guess wrong,
someone with better information will correct me.
Thanks for the feedback.
Chris Fearnley [EMAIL PROTECTED] said:
[Deleted first part]
o I'm not clear on the provides virtual package as it would apply to
'awk'. Should gawk provide 'awk'? Should mawk provide 'awk'?
Since clearly awk should be a symbolic link to gawk or mawk but not
both, is there a conflict here? Or should the conflict be resolved
by the {post,pre}{inst,rm} scripts (this seems best to me as most
people would have both installed side by side)? Maybe my mawk package
should Recommends or Suggests gawk?
Yes, both gawk and mawk could provide awk (presuming that mawk is
100.00% backwards compatable with standard awk). Other packages
could then have a dependency on awk satisfied by either of the
packages which provided it being installed.
Yes, mawk(1) is 100% POSIX awk compient. It is GPL'd and still under
development. It is also, perhaps, 3 times faster than gawk. Though
someone could write an awk script that depended on a gawk extension.
[more deleted]
o It occurs to me that many packages should Recommends or Suggests
man and/or info. Yet they don't. Why? What am I missing? And most
packages depend on ldso. So are essential packages too basic to be
listed in the control files? Is it possible that this lack of
consistency in setting dependencies could affect us much like the
current libc, libc4 and libc5 (which as required packages under Debian
0.93R6 weren't mentioned in the depends fields of any packages).
True. Packages install their man pages regardless of whether
/usr/bin/man is installed. Ditto with info pages and /usr/bin/info.
Also, some facilities provided by base system packages are necessary
for installation of (some or all) packages, but no dependency is
declared on the packages which provide those facilities.
Generally, packages presume that the base system packages will be
present without declaring explicit dependencies (though I don't
think it's documented precisely what the base system is guaranteed
to provide), only declare dependencies on other packages which are
needed for their proper installation and operation, and omit declaring
dependencies on packages needed to display their docs.
This strikes me as a bug: it's those little assumptions that could
come back to bite. Perhaps the Maintainers FAQ should suggest that
man, info, ldso, sh, perl, et. al. be added as per appropriate to the
recommends, suggests, and depends fields (just as libc{,4,5} are now
suggested to be explicitely listed).
--
Christopher J. Fearnley|UNIX SIG Leader at PACS
[EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society)
[EMAIL PROTECTED] |Design Science Revolutionary
http://www.netaxs.com/~cjf |Explorer in Universe
Dare to be Naive -- Bucky Fuller |Linux Advocate
Bug#1840: ncftp problems
Package: ncftp Version: 2.1.0 It is very easy to confuse ncftp into refusing to allow ^Z to suspend the process. One way I have done this follows (hard to represent visual mode, sorry): [EMAIL PROTECTED]:~$ ncftp o ftp.debian.org cd /debian/debian-1.0/source Now ^Z won't suspend the session any more. Interestingly, ^Z worked before the last typed command. There are several other sequences of ncftp commands that will also suspend ^Z. Strangely, ^Z generally works just when the program is first started. Also, ^W doesn't work, readline only partially works. Hmm, so much is wrong one might suspect curses or ncurses. I haven't even created a .netrc or .ncftprc config file yet - this is virgin debian ncftp. This is under debian 0.93R6, kernel 1.2.13, and libc 4.6.27. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
Bug#1842: pari's user-manual has a permission problem
Package: pari Version: 1.39 Revision: 3 /usr/doc/pari/user-manual doesn't have global execute permissions on that directory, so it's hard to read the docs as a user. This is under debian 0.93R6, kernel 1.2.13, and libc 4.6.27. -- Christopher J. Fearnley|UNIX SIG Leader at PACS [EMAIL PROTECTED] (finger me!)|(Philadelphia Area Computer Society) [EMAIL PROTECTED] |Design Science Revolutionary http://www.netaxs.com/~cjf |Explorer in Universe Dare to be Naive -- Bucky Fuller |Linux Advocate
