Re: Non-debian running DD's (Was: Re: stop abusing debconf already)
On 24-Apr-03, 13:31 (CDT), Tollef Fog Heen [EMAIL PROTECTED] wrote: when did windows become the native OS on x86? Just to be difficult: The original OS for the IBM PC was DOS[1], and if you track the lineage, I think it's fair to call Windows the native OS for that architecture. And when did MacOS become the native OS on PPC? By my above argument, the native OS for PPC is AIX. :-) Steve [1] Yes, I know about (and used) the UCSD P-System, but it wasn't shipped with the IBM PC. -- Steve Greenland The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net
Re: Non-debian running DD's (Was: Re: stop abusing debconf already)
On Fri, Apr 25, 2003 at 09:08:03AM -0500, Steve Greenland wrote: On 24-Apr-03, 13:31 (CDT), Tollef Fog Heen [EMAIL PROTECTED] wrote: when did windows become the native OS on x86? Just to be difficult: The original OS for the IBM PC was DOS[1], and if you track the lineage, I think it's fair to call Windows the native OS for that architecture. Hey, what about the in-ROM BASIC the early IBM machines had? -- You grabbed my hand and we fell into it, like a daydream - or a fever. pgp6DToTmuVB7.pgp Description: PGP signature
Re: Non-debian running DD's (Was: Re: stop abusing debconf already)
* LapTop006 | I use both Mutt and OE to read my E-mail (mostly mutt). The one feature | OE has (on both mac and windows) that NO other client I've seen matches | (Mozilla 1.0 came close, haven't tried since then) is its support for | offline IMAP. use isync or something similar? Or try gnus (which I believe has offline-imap support, but I don't use IMAP, so I'm not sure). | [..] but on x86 and PPC I also run the native OS' (Mac OS and | Windows (windows stays on my laptop because I can kernel-panic linux | with a keystroke...)). when did windows become the native OS on x86? And when did MacOS become the native OS on PPC? AmigaOS is just as native on many PPC boxen (if native in this context mean «what the box comes with from the manufacturer»). -- Tollef Fog Heen,''`. UNIX is user friendly, it's just picky about who its friends are : :' : `. `' `-
Re: stop abusing debconf already
On 21-Apr-03, 21:16 (CDT), John Hasler [EMAIL PROTECTED] wrote: Don Armstrong writes: I (apparently incorrectly) presumed that debconf was also intended to allow for the eventual automation of replicated Debian installations. I distinctly remember reading exactly that. I remember that this came along later, but I could be incorrect. It's misguided, though. For this to work, we need *EVERY* configuration variable needs to be under the control of debconf, and we can just get rid of conffiles and support of editing files directly. Otherwise, you still need an alternative way to distribute configuration data/files (e.g. cfengine), and then there's no need to use debconf, except as a standard frontend for the non-defaultable options. Don't get me wrong, I'm *NOT* against debconf. I think it's great to have a standard UI for installation questions. I'm against being asked questions about things that should just default, and I'm against converting conffiles to non-conffiles for no reason. If I want to change a default, I'll edit the file, and you damned will better not overwrite it. Now, if ucf (or some other tool) can make non-conffiles work as well as conffiles, and people use it correctly, and it integrates properly into dpkg and apt, then that will be cool, and about 90% of my objections will go away, and I'll be reduced to filing bugs against packages that think Should I make your cursor blink blue or green is a high-priority question. I look forward to that day. Regards, Steve -- Steve Greenland The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net
Re: debconf review of cvsd (was Re: stop abusing debconf already)
* Joey Hess | Hmm, you might have to do something mildly tricky with the limits stuff; | if the user did not turn it on you would have to manage | commenting/uncommenting the lines in the config file. Still seems quite | doable. db_get mailman/gate_news || true if [ $RET = no ]; then # Comment out any lines containing the words gate_news awk '/^[^#]*gate_news/ { print # $LINE;next }; //' /etc/cron.d/mailman /etc/cron.d/mailman.$$ else # Remove commenting of any lines containing gate_news awk '/^[ \t]*#.*gate_news/ { sub(/^[ \t]*#[ \t]*/, ); print; next }; //' /etc/cron.d/mailman /etc/cron.d/mailman.$$ fi is what I use in mailman for commenting and uncommenting lines in the postinst. It seems to work, but kinda sucks. (If anybody feels like reviewing the debconf usage in mailman, feel free. :) -- Tollef Fog Heen,''`. UNIX is user friendly, it's just picky about who its friends are : :' : `. `' `-
Re: Non-debian running DD's (Was: Re: stop abusing debconf already)
On Mon, 2003-04-21 at 08:58, LapTop006 wrote: I use both Mutt and OE to read my E-mail (mostly mutt). The one feature OE has (on both mac and windows) that NO other client I've seen matches (Mozilla 1.0 came close, haven't tried since then) is its support for offline IMAP. Try Mac OS X's Mail program. It has at least as good offline support as OE, and is much nicer, too. /me wishes Mail were free. signature.asc Description: This is a digitally signed message part
Re: debconf review of cvsd (was Re: stop abusing debconf already)
On Sat, Apr 19, 2003 at 12:12:38PM -0400, Joey Hess wrote: cvsd.conf is a trivial config file to parse and modify from what I can see. port=`sed -n 's/^Port *\([^ ]*\).*$/\1/p' /etc/cvsd/cvsd.conf` That's a reasonable way to get any value from it. I'm glad you do this What about writing values? I imagine that changing the value of port might be as simple as: sed 's/^Port .*$/Port xyz/g' /etc/cvsd/cvsd.conf /etc/cvsd/cvsd.conf.new mv /etc/cvsd/cvsd.conf.new /etc/cvsd/cvsd.conf (not tested). However, what if (and I haven't looked at cvsd; some of my examples may not make sense for cvsd): - There is more then one Port setting? (its probably illegal here). Do you change all of them, or only the first one, or what? - There is no Port setting and one needs to be added? Is it OK to blindly add at the end of the file, or should a script try some black magic to work out the best spot? - Adding a port setting destroys the config, because the adminstrator deliberatly deleted that setting for some reason (I am assuming this would mean something, eg. only bind to a UNIX stream socket, for instance). - The system adminstrator accidently or deliberately commented out the port setting, and puts it back in and finds it doesn't do anything, because a new port setting has been added to the end and has overriden the first setting without giving any errors or warnings? Now thats a lot of ifs. ;-). I am just curious though on what how other developers feel that these situations should be handled. -- Brian May [EMAIL PROTECTED]
Re: Non-debian running DD's (Was: Re: stop abusing debconf already)
On Wed, Apr 23, 2003 at 04:01:35AM -0400, Anthony DeRobertis wrote: On Mon, 2003-04-21 at 08:58, LapTop006 wrote: I use both Mutt and OE to read my E-mail (mostly mutt). The one feature OE has (on both mac and windows) that NO other client I've seen matches (Mozilla 1.0 came close, haven't tried since then) is its support for offline IMAP. Try Mac OS X's Mail program. It has at least as good offline support as OE, and is much nicer, too. /me wishes Mail were free. Please hack on GNUMail.app. It is in the gnumail package. Simon
Re: stop abusing debconf already
On Mon, 21 Apr 2003 21:16:08 -0500, John Hasler [EMAIL PROTECTED] wrote: Don Armstrong writes: I (apparently incorrectly) presumed that debconf was also intended to allow for the eventual automation of replicated Debian installations. I distinctly remember reading exactly that. And I have greatly simplified my potato = woody updates by seeding the debconf database with answers to anticipated questions. With database overloading, there is even a mechanism in place to allow things like that, but it didn't work when I tried. Had to brute my answers into the real debconf database with a small perl script :-( Greetings Marc -- -- !! No courtesy copies, please !! - Marc Haber |Questions are the | Mailadresse im Header Karlsruhe, Germany | Beginning of Wisdom | Fon: *49 721 966 32 15 Nordisch by Nature | Lt. Worf, TNG Rightful Heir | Fax: *49 721 966 31 29
Re: debconf review of cvsd (was Re: stop abusing debconf already)
On Sun, Apr 20, 2003 at 10:24:01AM -0400, Joey Hess wrote: This looks like it may be due to a bug (or incompatibility) in zsh. Do you have /bin/sh set to zsh? I have some strange results if I use zsh to process the postinst. I'll do some more testing. Somehow the result of the 'GET cvsd/limits' is passed to the next command. My /bin/sh is dash. Looks like there must be some problem in the for loop that is not splitting the RET on words. How much of POSIX compliant is dash? I have not been able to reproduce your abcde bugs [1] [2] by using ksh/sh/bash in POSIX mode. J [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=185640 [2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=187788 -- Jesus Climent | Unix SysAdm | Helsinki, Finland | pumuki.hispalinux.es GPG: 1024D/86946D69 BB64 2339 1CAA 7064 E429 7E18 66FC 1D7F 8694 6D69 -- Registered Linux user #66350 proudly using Debian Sid Linux 2.4.20 I never drink ... wine. --Dracula (Dracula) pgp6abw2A7OHb.pgp Description: PGP signature
Re: debconf review of cvsd (was Re: stop abusing debconf already)
Jesus Climent wrote: How much of POSIX compliant is dash? I have not been able to reproduce your abcde bugs [1] [2] by using ksh/sh/bash in POSIX mode. Well I think the abcde bugs are probably dash bugs. But aside from bugs it's as posix compliant as anything else in debian, as far as that goes. -- see shy jo pgpM7PyujfXX6.pgp Description: PGP signature
Re: debconf review of cvsd (was Re: stop abusing debconf already)
I demand that Joey Hess may or may not have written... [snip] You may choose to remove the chroot jail but you will also loose all the repositories inside the chroot jail. If you have not | backed up your repositories you want to keep, do not remove it now; | manually remove it later once your repositories are safe. One other small change: s/loose/lose/. -- | Darren Salt | nr. Ashington, | linux (or ds) at | woody, sarge, | Northumberland | youmustbejoking | RISC OS | Toon Army | demon co uk | This space reserved for future expansion Nothing is as temporary as that which is called permanent.
Re: stop abusing debconf already
Apologies, 'reply-all' is not clever enough in Outlook Express to evaluate the sender preference on being copied on list emails. Any suggestions for a MUA that can perform this feat are appreciated. Any mailer that honours the Mail-Followup-To: header that I set would do nicely. There are plenty of these in Debian, such as the one I'm using. As another developer has pointed out, I'm not running Debian on this box. There are plenty of other email clients for Windows, but I suspect that all of them are going to be somewhat lax in following follow-up headers in email. Those that potentially do are most likely not going to support IMAP at the same time. Despite is parentage, OE is quite acceptable for day to day use. Again, if anyone knows of another client that supports both requirements I'll give it a go. Matt.
Re: stop abusing debconf already
True, however it seems clear that he is not running Debian. This is the case as you have noticed. (the irony is almost too much to bear) Why? I have 5 PC's here (at home) and 4 of them run Debian (mixture of stable, testing and unstable). I have one Windows box that I use for email and web browsing as I happen to like OE and IE for these tasks. All the backend services (DNS, NFS, POP, IMAP, SMB, HTTP) are run on the Debian servers. It works well and I see no reason to change the way I use it. Should I have to run Debian (to continue to be a developer?) on every box I have or can I make an informed decision on the best way to support my work? Matt.
Re: stop abusing debconf already
No offence taken. I joined when Debian wasn't run by anal retentives. Sure there was the whole free software part - but not the SS Nazi version of free software that is being prompted recently. I have to say that I'm beginning to think that your assessment is right and I should find a more liberal bunch of Linux fanatics to join with. Oooh, is that a promise? Anything we can do to facilitate that? Get some policy in place that allows you to throw out people who you don't agree with or don't like would do it one way or another. Matt.
Re: stop abusing debconf already
On ma, 2003-04-21 at 13:03, Matt Ryan wrote: As another developer has pointed out, I'm not running Debian on this box. There are plenty of other email clients for Windows, but I suspect that all of them are going to be somewhat lax in following follow-up headers in email. I would like to point out that support for the Mail-Followup-To header is not required. It is sufficient that the mail client lets you edit the headers before sending the mail. This works in all mail clients I'm aware of, even if some of them make things a bit awkward.
Re: stop abusing debconf already
I would like to point out that support for the Mail-Followup-To header is not required. It is sufficient that the mail client lets you edit the headers before sending the mail. This works in all mail clients I'm aware of, even if some of them make things a bit awkward. Trouble is I need to know what the sender of the email, I'm replying to, wanted in regards to getting copies of the response to both list and direct. One could manually parse the email headers and set the reply appropriately but this is a rather onerous task when replying to a large number of emails. Matt.
Re: stop abusing debconf already
On Mon, Apr 21, 2003 at 12:07:24PM +0100, Matt Ryan wrote: Trouble is I need to know what the sender of the email, I'm replying to, wanted in regards to getting copies of the response to both list and direct. One could manually parse the email headers and set the reply appropriately but this is a rather onerous task when replying to a large number of emails. If you're not sure, you should just follow list policy and not Cc. It's also nice to attribute your quotes, btw. Richard Braakman
Outlook Express (Re: stop abusing debconf already)
On Mon, Apr 21, 2003 at 11:08:04AM +0100, Matt Ryan wrote (and CC'd me): (the irony is almost too much to bear) Why? I have 5 PC's here (at home) and 4 of them run Debian (mixture of stable, testing and unstable). I have one Windows box that I use for email and web browsing as I happen to like OE and IE for these tasks. All the backend services (DNS, NFS, POP, IMAP, SMB, HTTP) are run on the Debian servers. It works well and I see no reason to change the way I use it. Should I have to run Debian (to continue to be a developer?) on every box I have or can I make an informed decision on the best way to support my work? This need not be a philosophical discussion, and it need not involve your software preferences. The difficulty is that the MUA that you choose to use is lacking (and/or broken) in ways which inconvenience others on this mailing list (carbon copies, threads, etc.). If it can be fixed, please fix it, and if it cannot, please consider an alternative. -- - mdz
Non-debian running DD's (Was: Re: stop abusing debconf already)
I use both Mutt and OE to read my E-mail (mostly mutt). The one feature OE has (on both mac and windows) that NO other client I've seen matches (Mozilla 1.0 came close, haven't tried since then) is its support for offline IMAP. This e-mail comes to you via putty on my laptop being NAT'd via my debian proxy, to my debian server at work, both running woody which runs mutt that then sends via the postfix on the box. We should not be larting people who use other operating systems, I run debian across several different archs (Sparc, Alpha, HPPA, soon: MIPS, m68k, powerpc), but on x86 and PPC I also run the native OS' (Mac OS and Windows (windows stays on my laptop because I can kernel-panic linux with a keystroke...)). For example packages like samba would be ideally packaged by someone who had at least one windows box to test with (both as a client and server), many other packages are similar. On Mon, Apr 21, 2003 at 02:32:30PM +0300, Richard Braakman arranged a set of bits into the following: On Mon, Apr 21, 2003 at 12:07:24PM +0100, Matt Ryan wrote: Trouble is I need to know what the sender of the email, I'm replying to, wanted in regards to getting copies of the response to both list and direct. One could manually parse the email headers and set the reply appropriately but this is a rather onerous task when replying to a large number of emails. If you're not sure, you should just follow list policy and not Cc. It's also nice to attribute your quotes, btw. Richard Braakman pgpKWNhJPI7gY.pgp Description: PGP signature
Re: stop abusing debconf already
Matt Ryan (2003-04-21 11:03:49 +0100) : Again, if anyone knows of another client that supports both requirements I'll give it a go. A good mail client that works on Windows, provides IMAP and obeys standard headers? I suggest Gnus. It does all that, and more. Roland. -- Roland Mas Sauvez les castors, plantez des arbres.
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 11:03:03AM -0400, Joey Hess wrote: Denis Barbier wrote: On Fri, Apr 18, 2003 at 07:14:19PM -0400, Joey Hess wrote: Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Stop slapping incorrect uses of debconf in everywhere. Feel free to run any package using debconf by me before you upload it, or take the time to understand yourself how things should work. I do not understand exactly what is good and bad use of debconf. For instance all questions asked by the debconf package have good default values, so there is no reason to prompt user, a configuration file is enough. So what am I missing? The only reasons debconf itself asks those questions still are: - example of how debconf works - they _are_ overridable with the config file, in an appropriate way - it's not asked at new installs anyway - people are used to that I did have some bad ideas about how debconf could be used back years ago when I wrote it. Those debconf questions (and some really ill-advised stuff slrn used to do) were the result. I hope my understanding of the big picture has been seriously improved. Honestly you should not be so upset by these debconf abuses about configuration files overwriting, this is a difficult issue and AFAICT documentation does not help. IMO debconf itself could be a better example. Currently configuration items can indeed be overridden in debconf.conf, but its parsing is performed by debconf itself, so other packages cannot reproduce the same logic. As you explained, they have to parse their configuration files and set updated question values, in order not to lose user changes. Debconf has a different behavior when configuring itself, it does not update question values with those found in debconf.conf. To be a truly useful example, it could also overwrite its own configuration file in order to show how to do it right. Denis
Re: stop abusing debconf already
* Roland Mas [EMAIL PROTECTED]: Matt Ryan (2003-04-21 11:03:49 +0100) : Again, if anyone knows of another client that supports both requirements I'll give it a go. A good mail client that works on Windows, provides IMAP and obeys standard headers? I suggest Gnus. It does all that, and more. Mulberry, but it's not free -- Ralf Hildebrandt (Im Auftrag des Referat V a) [EMAIL PROTECTED] Charite Campus MitteTel. +49 (0)30-450 570-155 Referat V a - Kommunikationsnetze - Fax. +49 (0)30-450 570-916 AIM: ralfpostfix
Re: stop abusing debconf already
On Mon, 21 Apr 2003 12:07:24 +0100, Matt Ryan [EMAIL PROTECTED] said: I would like to point out that support for the Mail-Followup-To header is not required. It is sufficient that the mail client lets you edit the headers before sending the mail. This works in all mail clients I'm aware of, even if some of them make things a bit awkward. Trouble is I need to know what the sender of the email, I'm replying to, wanted in regards to getting copies of the response to both list and direct. One could manually parse the email headers and set the reply appropriately but this is a rather onerous task when replying to a large number of emails. *Sigh*. The rule, for Debian lists, is *NO* to send CC's to anyone, and is spelled out on the mailing list policy/rules of conduct web page. So if you do not know better, the default is _not_ to send copies. manoj -- Going to church does not make a person religious, nor does going to school make a person educated, any more than going to a garage makes a person a car. Manoj Srivastava [EMAIL PROTECTED] http://www.debian.org/%7Esrivasta/ 1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Re: stop abusing debconf already
On 19-Apr-03, 11:44 (CDT), David B Harris [EMAIL PROTECTED] wrote: From debconf-devel(8): low: Very trivial items that have defaults that will work in the vast majority of cases; oinly control freaks see these. From Debian policy, 11.7.3, regarding how to achieve the requirement of not overwriting maintainer changes: The easy way to achieve this behavior is to make the configuration file a conffile. This is appropriate only if it is possible to distribute a default version that will work for most installations, although some system administrators may choose to modify it. This implies that the default version will be part of the package distribution, and must not be modified by the maintainer scripts during installation (or at any other time). When I wrote this, and it was approved, I believe the intent was that the conffile mechanism be used unless a package configuration did not meet the requirement regarding acceptable defaults. The unfortunate phrase This is appropriate *only* if... is there because at the time, the problem was with people modifying conffiles, rather than not using conffiles. Now that the pendulum has swung the other way, perhaps we need to rephrase it, to make it clear that conffiles are the preferred choice. If you have a package that is asking only medium and lower priority debconf questions, then debconf should not be used at all. Those priorities *exist* because there are packages that have a high-priority, non-defaultable question, and once you've broken the conffile system, you might as well include those questions. Perhaps it was a bad idea. Another use for those lower priorities is for notes to the admin. I contend that this second use *is* a bad idea, because the common implementation is to NOT include the same information under /usr/share/doc/pkg, and thus those of us who have low and medium priority turned off lose that info. Steve -- Steve Greenland The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net
Re: stop abusing debconf already
On 20-Apr-03, 21:14 (CDT), Steve Langasek [EMAIL PROTECTED] wrote: Windows-centric? FFS, where do you think the term registry /comes/ from? While the term registry comes from Windows, it's worth noting that AIX had the equivalent ('object manager') (which has nothing to do with OOP) long before. And it's just as much a pain in the ass on AIX, as one never knows where a configuration item comes from, and whether changing the usual configuration file will do something useful, or whether the change will be preserved. Well, it's not quite as bad as the Windows version, because it doesn't corrupt itself when looked at sideways, since it was written by competent people. Steve -- Steve Greenland The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net
Re: stop abusing debconf already
On Mon Apr 21, 10:05am -0500, Steve Greenland wrote: On 19-Apr-03, 11:44 (CDT), David B Harris [EMAIL PROTECTED] wrote: From debconf-devel(8): low: Very trivial items that have defaults that will work in the vast majority of cases; oinly control freaks see these. If you have a package that is asking only medium and lower priority debconf questions, then debconf should not be used at all. Those priorities *exist* because there are packages that have a high-priority, non-defaultable question, and once you've broken the conffile system, you might as well include those questions. Perhaps it was a bad idea. Another use for those lower priorities is for notes to the admin. I contend that this second use *is* a bad idea, because the common implementation is to NOT include the same information under /usr/share/doc/pkg, and thus those of us who have low and medium priority turned off lose that info. I never said otherwise, and in fact I have followed that practice myself. pgpK4f2EaPA2M.pgp Description: PGP signature
Re: stop abusing debconf already
On Mon, Apr 21, 2003 at 03:49:43PM +0200, Denis Barbier wrote: [...] Honestly you should not be so upset by these debconf abuses about configuration files overwriting, this is a difficult issue and AFAICT documentation does not help. Correction, debconf-devel(7) explains how to do it right, it contains a detailed example. Denis
Re: stop abusing debconf already
On Mon, 21 Apr 2003, Steve Greenland wrote: If you have a package that is asking only medium and lower priority debconf questions, then debconf should not be used at all. I (apparently incorrectly) presumed that debconf was also intended to allow for the eventual automation of replicated Debian installations. As such, I had viewed management of medium and lower priority configuration options through debconf as a good thing, provided of course, that local admin modications remained intact. Is there currently a better machanism of providing this functionality than conffiles? [Or are admins who want automated installs stuck with diffing packages whose configurations they wish to change?] Don Armstrong -- People selling drug paraphernalia ... are as much a part of drug trafficking as silencers are a part of criminal homicide. -- John Brown, DEA Chief http://www.donarmstrong.com http://www.anylevel.com http://rzlab.ucr.edu pgpUG8ClzbIHs.pgp Description: PGP signature
Re: stop abusing debconf already
Don Armstrong writes: I (apparently incorrectly) presumed that debconf was also intended to allow for the eventual automation of replicated Debian installations. I distinctly remember reading exactly that. -- John Hasler [EMAIL PROTECTED] (John Hasler) Dancing Horse Hill Elmwood, WI
Re: stop abusing debconf already
On Sat, 19 Apr 2003 16:47:26 -0500, Steve Langasek [EMAIL PROTECTED] wrote: Um, no. *Policy* says that it may not be used as a registry. This has always prompted me to ask myself _why_ debconf entries are persistent then. If I _really_ have to parse config files in my config script to properly seed debconf to ask the right questions, then why does debconf have a database in the first place? Greetings Marc -- -- !! No courtesy copies, please !! - Marc Haber |Questions are the | Mailadresse im Header Karlsruhe, Germany | Beginning of Wisdom | Fon: *49 721 966 32 15 Nordisch by Nature | Lt. Worf, TNG Rightful Heir | Fax: *49 721 966 31 29
Re: stop abusing debconf already
On Sun, 20 Apr 2003 08:58:14 +0200, Marc Haber [EMAIL PROTECTED] said: On Sat, 19 Apr 2003 16:47:26 -0500, Steve Langasek [EMAIL PROTECTED] wrote: Um, no. *Policy* says that it may not be used as a registry. This has always prompted me to ask myself _why_ debconf entries are persistent then. So one does not have to answer the same question over and over on each install. If I _really_ have to parse config files in my config script to properly seed debconf to ask the right questions, then why does debconf have a database in the first place? Because not all debconf responses are stored in configuration files? Some debconf answers influence installation scripts behaviour, and never impact the installed program behaviour. manoj -- No problem is so formidable that you can't just walk away from it. Schulz Manoj Srivastava [EMAIL PROTECTED] http://www.debian.org/%7Esrivasta/ 1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 08:58:14AM +0200, Marc Haber wrote: | | This has always prompted me to ask myself _why_ debconf entries are | persistent then. If I _really_ have to parse config files in my config | script to properly seed debconf to ask the right questions, then why | does debconf have a database in the first place? | So that most of the time, you don't have to hit enter mindlessly to confirm the answers to questions that you've already answered when upgrading? Cameron.
Re: stop abusing debconf already
On Sat, 19 Apr 2003 20:17:16 +0100, Matt Ryan [EMAIL PROTECTED] said: Policy is what matters not the opinion of some jumped up developers! Conside rthis: when considering input from a ``jumped up developer'' who has demonstrated competence and has put in the effort like Joey Hess, and has intituted a couple of major changes in how Debian works, and an unknown twit, guess who am I going to listen to? manoj -- My dear People. My dear Bagginses and Boffins, and my dear Tooks and Brandybucks, and Grubbs, and Chubbs, and Burrowses, and Hornblowers, and Bolgers, Bracegirdles, Goodbodies, Brockhouses and Proudfoots. Also my good Sackville Bagginses that I welcome back at last to Bag End. Today is my one hundred and eleventh birthday: I am eleventy-one today! Tolkien Manoj Srivastava [EMAIL PROTECTED] http://www.debian.org/%7Esrivasta/ 1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Re: debconf review of cvsd (was Re: stop abusing debconf already)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Arthur de Jong wrote: Ok, could you review my cvsd package for me for correct debconf usage and tell me what you do and don't like? Thanks for taking advantage of that offer. (So far you're the only one.) I am ccing this to -devel just because. All of the debconf questions are pretty well worded and clear, but there's always room for improvement: Thanks for the review. I've changed most of the wording according to your suggestions (English isn't my native language). s/zero (0)/0/ # Apparently writing it out has the possibility to make # someone enter the number the wrong way so why not just # not write it out? I spelled out zero because some (most) fonts don't represent 0 differently from O. Same goes for 1 and l. - It's sort of odd that you use spaces to separate multiple items in one question, and then colons to separate them in the next. It would be nice if debconf would have some unified way of entering lists. I'm also not very happy with the way it is currently done, but I've modeled it after the cvs questions and a colon is the standard seporator for directories, while using : in addresses may provide problems with IPv6 addresses (improving this is on my TODO list). - I was fairly confused by cvsd/limit_memorylocked, because I thought that only programs run as root could lock memory, and why would cvs want to? Was this just added for (over)completeness? Probably. I provided questions for all the limits I could find. The code also covers limits for things that are not settable on linux (there is even a debconf question about it that is never asked). - You have no translations for the debconf templates in the package. The DDTP project has a complete German translation at http://ddtp.debian.org/debconf/source/cvsd. Of course you may want to make the above changes first, and wait for an updated translation.. I have received a Brazillian translation of the debconf questions that I'm merging into cvsd (bug #187795). I saw the German translation at http://ddtp.debian.org/cgi-bin/ddtp.cgi?part=debconfpackage=cvsd before but I never saw the page you linked (very useful page but I can't find it linked from the site). And since the site says (Now) we don't need any action from the debian package maintainer. I didn't take any action regarding the translation. If I reconfigure cvsd, pick all the limits, and take the default value for everything else, it exits with code 20: debconf (developer): -- GET cvsd/limits debconf (developer): -- 0 coredumpsize, cputime, datasize, filesize, memorylocked, openfiles, maxproc, memoryuse, stacksize, virtmem debconf (developer): -- GET cvsd/limit_coredumpsize cputime datasize filesize memorylocked openfiles maxproc memoryuse stacksize virtmem debconf (developer): -- 20 Incorrect number of arguments zsh: exit 20DEBCONF_DEBUG=. dpkg-reconfigure cvsd Looks like a bug.. After doing this I also did not see all the limits stuff in cvsd.conf, it had only these config items: Uid cvsd Gid cvsd PidFile /var/run/cvsd.pid RootJail /var/lib/cvsd MaxConnections 10 Nice 1 Listen * 2401 This looks like it may be due to a bug (or incompatibility) in zsh. Do you have /bin/sh set to zsh? I have some strange results if I use zsh to process the postinst. I'll do some more testing. Somehow the result of the 'GET cvsd/limits' is passed to the next command. Looking at your priorities, I'm not sure why the maximum connection limit is at priority medium. I already changed it to low for the next release (couldn't remember either why it was set to medium). I've set it up to ask first whether to use debconf to manage configuration of /etc/cvsd/cvsd.conf (not a conffile). If the user chooses to not use debconf, he's on his own (an example cvsd.conf is provided). Of course this is where in my opinion you lose. First of all, as Colin pointed out recently, any question that asks use debconf and defaults to true results in the package violating debian policy on any system where the debconf priority is higher than the question, or noninteractive mode is used. If the question is not shown to the user at all, and a default config is put into place, and the user edits that config and loses changes on upgrade, then you've violated policy. Secondly of course, I hate all use debconf questions. It's the wrong way to use debconf; you should be intelligently parsing answers out of the file on reconfigure and upgrade, and intelligently editing them back into the file, so you do not lose an admin's modifications or comments. No config file in debian should have warnings at the top about not editing it. I wince every time I see another do not edit this file; I consider every one at least one user lost to gentoo or some other distribution. I also don't like the question and would rather see it removed. I was working
Re: debconf review of cvsd (was Re: stop abusing debconf already)
Arthur de Jong wrote: I have received a Brazillian translation of the debconf questions that I'm merging into cvsd (bug #187795). I saw the German translation at http://ddtp.debian.org/cgi-bin/ddtp.cgi?part=debconfpackage=cvsd before but I never saw the page you linked (very useful page but I can't find it linked from the site). And since the site says (Now) we don't need any action from the debian package maintainer. I didn't take any action regarding the translation. I think they want thier translations to be used, but I don't really know. Don't expect the DDTP to make things very useful or organized though. :-/ I have a script that can automatically download translations from the DDTP, it's only a prototype however. If I reconfigure cvsd, pick all the limits, and take the default value for everything else, it exits with code 20: debconf (developer): -- GET cvsd/limits debconf (developer): -- 0 coredumpsize, cputime, datasize, filesize, memorylocked, openfiles, maxproc, memoryuse, stacksize, virtmem debconf (developer): -- GET cvsd/limit_coredumpsize cputime datasize filesize memorylocked openfiles maxproc memoryuse stacksize virtmem debconf (developer): -- 20 Incorrect number of arguments zsh: exit 20DEBCONF_DEBUG=. dpkg-reconfigure cvsd Looks like a bug.. After doing this I also did not see all the limits stuff in cvsd.conf, it had only these config items: Uid cvsd Gid cvsd PidFile /var/run/cvsd.pid RootJail /var/lib/cvsd MaxConnections 10 Nice 1 Listen * 2401 This looks like it may be due to a bug (or incompatibility) in zsh. Do you have /bin/sh set to zsh? I have some strange results if I use zsh to process the postinst. I'll do some more testing. Somehow the result of the 'GET cvsd/limits' is passed to the next command. My /bin/sh is dash. Looks like there must be some problem in the for loop that is not splitting the RET on words. Reading the current configuration is almost finished and will go into the next release. I'll look into editing the existing configuration file rather than overwriting it. I ran into some problems earlier though (especially things like where do you put new configuration options that were not there before). Also Limit, Repos and Listen may be specified multiple times so that may be tricky. Remember you can use perl for either the config or the postinst script. Might make the more complicated stuff more maneagable. I'm glad you're making progress on this. -- see shy jo pgpojKvSKBZdx.pgp Description: PGP signature
Re: debconf review of cvsd (was Re: stop abusing debconf already)
On Sun, Apr 20, 2003 at 12:03:40PM +0200, Arthur de Jong wrote: s/zero (0)/0/ # Apparently writing it out has the possibility to make # someone enter the number the wrong way so why not just # not write it out? I spelled out zero because some (most) fonts don't represent 0 differently from O. Same goes for 1 and l. In this case, I would suggest putting the word, not the number, in parentheses: 0 (zero). Regards, -- Steve Langasek postmodern programmer pgpQhQ4mvmpq0.pgp Description: PGP signature
Re: stop abusing debconf already
Conside rthis: when considering input from a ``jumped up developer'' who has demonstrated competence and has put in the effort like Joey Hess, and has intituted a couple of major changes in how Debian works, and an unknown twit, guess who am I going to listen to? Yawn. I don't know and I doubt you would listen if I did. It doesn't really matter. Everyone has an opinion and I have voiced mine. Perhaps I should be censored to avoid my rambling encroaching into your computer? Just because they have put some effort into the project does *not* make them infallible - all it makes them is a good coder, nothing more. Emboldening them with further super-powers is ridiculous. Matt.
Re: stop abusing debconf already
No offense, but I think you joined the wrong project, then. No offence taken. I joined when Debian wasn't run by anal retentives. Sure there was the whole free software part - but not the SS Nazi version of free software that is being prompted recently. I have to say that I'm beginning to think that your assessment is right and I should find a more liberal bunch of Linux fanatics to join with. Matt. BTW: Remember at the end of the day it is only a Linux distribution, nothing more, nothing less.
Re: stop abusing debconf already
Um, no. *Policy* says that it may not be used as a registry. [SNIPPED LONG DIATRIBE THAT DOES NOT PROVE THE ABOVE STATEMENT] Sure, you delete the registry things should still work. Did I say anything different? You are making a long tenuous link to prove your point which I don't subscribe to. Unless we have a formal definition of 'registry' in policy as well we can argue this forever as to what is the right way to view it. Perhaps not taking a Windows centric view would help? I'm not sure why you think Joey's expertise doesn't qualify him to make pronouncements about the use of debconf. Unlike you, he at least gets the answer right. Well I don't agree on the right answer bit and I never said he wasn't right (in some people views). All I said was he answer was not the gospel on this issue. Matt.
Re: stop abusing debconf already
BTW the opinion of this jumped-up developer is please don't send me private copies of posts to mailing lists. Thanks. Apologies, 'reply-all' is not clever enough in Outlook Express to evaluate the sender preference on being copied on list emails. Any suggestions for a MUA that can perform this feat are appreciated. Altering top posts based on preference would be an advantage as well. Matt.
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 08:57:38PM +0100, Matt Ryan wrote: No offense, but I think you joined the wrong project, then. No offence taken. I joined when Debian wasn't run by anal retentives. Sure there was the whole free software part - but not the SS Nazi version of free software that is being prompted recently. I have to say that I'm beginning to think that your assessment is right and I should find a more liberal bunch of Linux fanatics to join with. Please do. BTW: Remember at the end of the day it is only a Linux distribution, nothing more, nothing less. Bullshit. -- .''`. ** Debian GNU/Linux ** | Andrew Suffield : :' : http://www.debian.org/ | Dept. of Computing, `. `' | Imperial College, `- -- | London, UK pgpKB0ytYttoK.pgp Description: PGP signature
Re: stop abusing debconf already
On Sun, 2003-04-20 at 15:57, Matt Ryan wrote: No offense, but I think you joined the wrong project, then. No offence taken. I joined when Debian wasn't run by anal retentives. Sure there was the whole free software part - but not the SS Nazi [...] Congratulations, you just proved (yet again) Godwin's Law.
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 08:57:38PM +0100, Matt Ryan wrote: No offense, but I think you joined the wrong project, then. No offence taken. I joined when Debian wasn't run by anal retentives. Sure there was the whole free software part - but not the SS Nazi version of free software that is being prompted recently. A great, so we can end this thread now. BTW: Remember at the end of the day it is only a Linux distribution, nothing more, nothing less. Ever had a look at http://www.debian.org/ports/#nonlinux? cheers, Michael
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 08:57:38PM +0100, Matt Ryan wrote: No offense, but I think you joined the wrong project, then. No offence taken. I joined when Debian wasn't run by anal retentives. Sure there was the whole free software part - but not the SS Nazi version of free software that is being prompted recently. I have to say that I'm beginning to think that your assessment is right and I should find a more liberal bunch of Linux fanatics to join with. Oooh, is that a promise? Anything we can do to facilitate that? manoj -- A man who turns green has eschewed protein. Manoj Srivastava [EMAIL PROTECTED] http://www.debian.org/%7Esrivasta/ 1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 09:09:41PM +0100, Matt Ryan wrote: Colin Watson wrote: BTW the opinion of this jumped-up developer is please don't send me private copies of posts to mailing lists. Thanks. Apologies, 'reply-all' is not clever enough in Outlook Express to evaluate the sender preference on being copied on list emails. Any suggestions for a MUA that can perform this feat are appreciated. Any mailer that honours the Mail-Followup-To: header that I set would do nicely. There are plenty of these in Debian, such as the one I'm using. Cheers, -- Colin Watson [EMAIL PROTECTED]
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 10:53:29PM +0100, Colin Watson wrote: On Sun, Apr 20, 2003 at 09:09:41PM +0100, Matt Ryan wrote: Colin Watson wrote: BTW the opinion of this jumped-up developer is please don't send me private copies of posts to mailing lists. Thanks. Apologies, 'reply-all' is not clever enough in Outlook Express to evaluate the sender preference on being copied on list emails. Any suggestions for a MUA that can perform this feat are appreciated. Any mailer that honours the Mail-Followup-To: header that I set would do nicely. There are plenty of these in Debian, such as the one I'm using. True, however it seems clear that he is not running Debian. (the irony is almost too much to bear) -- - mdz
Re: debconf review of cvsd (was Re: stop abusing debconf already)
Joey Hess [EMAIL PROTECTED] writes: - cvsd/listen: s/cvsd will listen on/on which cvsd will listen/ # Avoid dangling preposition This is an English usage question of the sort that will get the English and Linguistics departments at some universities to start leaving nasty notes on each others' blackboards. Suffice it to say that I, as a native speaker of American English, think that the standard written English folks get this one wrong - prepositions at the of sentences are something up with which we should put. We should acknowledge English's Germanic heritage and that it too has separable verbs. However, this an unbelievably side issue even for -devel.
Re: stop abusing debconf already
On Sun, Apr 20, 2003 at 09:02:19PM +0100, Matt Ryan wrote: Um, no. *Policy* says that it may not be used as a registry. [SNIPPED LONG DIATRIBE THAT DOES NOT PROVE THE ABOVE STATEMENT] Sure, you delete the registry things should still work. Did I say anything different? You are making a long tenuous link to prove your point which I don't subscribe to. Unless we have a formal definition of 'registry' in policy as well we can argue this forever as to what is the right way to view it. Perhaps not taking a Windows centric view would help? Windows-centric? FFS, where do you think the term registry /comes/ from? Yes, policy doesn't say anything that prevents using debconf as a registry, because I have a definition of registry that only I know and which I guarantee you is allowed by policy. Harumph. In all the stupid flame wars I've seen about debconf, you're the first person I've ever met who was unable to grasp the use of the word registry. I'm not sure why you think Joey's expertise doesn't qualify him to make pronouncements about the use of debconf. Unlike you, he at least gets the answer right. Well I don't agree on the right answer bit That, and a few other things, have been made very clear. -- Steve Langasek postmodern programmer pgp8HYslwKUzb.pgp Description: PGP signature
Re: stop abusing debconf already
Package: binutils On Fri, 2003-04-18 at 19:14, Joey Hess wrote: Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Amen. For example, we really need to kill that kernel link failure info message in binutils. Not only is it totally obsolete, it was a bad idea in the first place.
Re: stop abusing debconf already
On Fri, Apr 18, 2003 at 07:14:19PM -0400, Joey Hess wrote: Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Stop slapping incorrect uses of debconf in everywhere. Feel free to run any package using debconf by me before you upload it, or take the time to understand yourself how things should work. I do not understand exactly what is good and bad use of debconf. For instance all questions asked by the debconf package have good default values, so there is no reason to prompt user, a configuration file is enough. So what am I missing? Denis
Re: stop abusing debconf already
At 7:22 pm, Saturday, April 19 2003, Denis Barbier mumbled: I do not understand exactly what is good and bad use of debconf. For instance all questions asked by the debconf package have good default values, so there is no reason to prompt user, a configuration file is enough. So what am I missing? Well, not all use of debconf is bad. For example, libnet-perl is a terrible misuse of debconf, *but* it can be remedied by dropping the priority of the questions from medium to low. Another bad example is binutils, which spits out that silly note even during a new install. If you test that $2 is empty, don't print it. Unfortunately, I don't want to point out a good usage of debconf, since I'd only be ringing my own bell. The barrier here is common sense. Personally, I feel that debconf asking questions during install is fine, if the priority is correct, and it does not forget the answers. Cheers, -- Steve
Re: stop abusing debconf already
Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Stop slapping incorrect uses of debconf in everywhere. Feel free to run any package using debconf by me before you upload it, or take the time to understand yourself how things should work. Great, helpful input to the debate. Let me see... 1) Toys 2) Pram 3) Throw 4) Profit? Matt.
Re: stop abusing debconf already
Denis Barbier wrote: On Fri, Apr 18, 2003 at 07:14:19PM -0400, Joey Hess wrote: Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Stop slapping incorrect uses of debconf in everywhere. Feel free to run any package using debconf by me before you upload it, or take the time to understand yourself how things should work. I do not understand exactly what is good and bad use of debconf. For instance all questions asked by the debconf package have good default values, so there is no reason to prompt user, a configuration file is enough. So what am I missing? The only reasons debconf itself asks those questions still are: - example of how debconf works - they _are_ overridable with the config file, in an appropriate way - it's not asked at new installs anyway - people are used to that I did have some bad ideas about how debconf could be used back years ago when I wrote it. Those debconf questions (and some really ill-advised stuff slrn used to do) were the result. -- see shy jo pgpede47N4y44.pgp Description: PGP signature
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 02:08:27PM +0100, Matt Ryan wrote: Joey Hess wrote: Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Stop slapping incorrect uses of debconf in everywhere. Feel free to run any package using debconf by me before you upload it, or take the time to understand yourself how things should work. Great, helpful input to the debate. Let me see... 1) Toys 2) Pram 3) Throw 4) Profit? Or maybe realize that Joey might perhaps know what he's talking about with regard to debconf ... you could go find the text of his talk at the last Debian Conference if you like. -- Colin Watson [EMAIL PROTECTED]
Re: stop abusing debconf already
Steve Kowalik wrote: Well, not all use of debconf is bad. For example, libnet-perl is a terrible misuse of debconf, *but* it can be remedied by dropping the priority of the questions from medium to low. At least libnet-perl is actually asking questions and preserving some (though not all) user changes to libnet.cfg. Michael's idea to add a gateway question is probably sufficient to solve the too many questions thing. I'd be just as happy if it only asked about stuff that really needs to be configured, such as ftp passive mode though. Another bad example is binutils, which spits out that silly note even during a new install. Yes. If you test that $2 is empty, don't print it. Unfortunately, I don't want to point out a good usage of debconf, since I'd only be ringing my own bell. I'll ring it for you: xringd is one of the better users of debconf I have seen, and it gets everything right. The problem is that with debconf, even 10% of packages getting it wrong make the whole system into crap. And probably closer to 70% of packages are currently getting it wrong.. -- see shy jo pgpKTvmvVG4yb.pgp Description: PGP signature
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 03:46:32PM +0100, Colin Watson wrote: snip Or maybe realize that Joey might perhaps know what he's talking about with regard to debconf ... you could go find the text of his talk at the last Debian Conference if you like. /snip Could you (or someone else) give me a hint on where one could find Joey's talk ? I've already tried googling for it and looking at [1] but couldn't find it. [1] http://www.debconf.org -- ++--++ || André Luís Lopes [EMAIL PROTECTED] || || Debian-BR Project http://debian-br.cipsga.org.br || || Public GPG KeyID 9D1B82F6 || || Keyserver wwwkeys.eu.pgp.net || pgplhDHN8iHcp.pgp Description: PGP signature
debconf review of cvsd (was Re: stop abusing debconf already)
Arthur de Jong wrote: Ok, could you review my cvsd package for me for correct debconf usage and tell me what you do and don't like? Thanks for taking advantage of that offer. (So far you're the only one.) I am ccing this to -devel just because. All of the debconf questions are pretty well worded and clear, but there's always room for improvement: - I'm not sure why you capitalized Chroot in the second sentence of the cvsd/rootjail question (and in the short description of that question) and filehierarchy probably needs a space in it. - cvsd/rootjail: s/zero (0)/0/ # Apparently writing it out has the possibility to make # someone enter the number the wrong way so why not just # not write it out? - cvsd/nice: s/too much/too many/ - cvsd/listen: s/cvsd will listen on/on which cvsd will listen/ # Avoid dangling preposition - Though it's pretty obvious what you mean by RootJail in the cvsd/repositories question, that is in fact the first time that particular term has been used. Also, you could use a debconf substitution here to substitute in the user's choce of the chroot jail directory directly into the question, to remind them what it is. Oh you also use the inconsistently capitalized rootjail in cvsd/usedebconf. Wouldn't just chroot jail be better? Or at least some consistency there and with the Chroot jail term in cvsd/rootjail. - cvsd/limits: s/to limit of pserver process./of pserver process to limit:/ - Most of the questions have short descriptions that do not end in punctuation, and this has unfortunate results in some debconf frontends, like this: Location of Chroot jail /var/lib/cvsd It's better to put a colon or a question mark at the end. - It's sort of odd that you use spaces to separate multiple items in one question, and then colons to separate them in the next. - It was not clear to me in the cvsd/limits question what would be the result of picking any of those limits. At first I thought it would enable some kind of hard-coded limits, which made me not want to mess with it. I see reading that templates file that it in fact enables a whole set of additional questions about the limits. Noting that you'll ask these questions in cvsd/limits would be a good idea. - cvsd/limit_coredumpsize: s/filesize/file size/ # don't invent terminology s/coredump/core dump/ # I know, it sucks when one's spacebarbreaks :-) - cvsd/limit_cputime: s/prevent too much cpu time allocated/prevent too much cpu time from being allocated/ s/amount of seconds cputime/amount of cpu time/ # it is not just in seconds form, and seconds cputime is # incorrect English. - I was fairly confused by cvsd/limit_memorylocked, because I thought that only programs run as root could lock memory, and why would cvs want to? Was this just added for (over)completeness? - cvsd/limit_maxproc: s/Cvs/cvs/ # for consistency with every other mention of the program # including other starts of sentences - You have no translations for the debconf templates in the package. The DDTP project has a complete German translation at http://ddtp.debian.org/debconf/source/cvsd. Of course you may want to make the above changes first, and wait for an updated translation.. If I reconfigure cvsd, pick all the limits, and take the default value for everything else, it exits with code 20: debconf (developer): -- GET cvsd/limits debconf (developer): -- 0 coredumpsize, cputime, datasize, filesize, memorylocked, openfiles, maxproc, memoryuse, stacksize, virtmem debconf (developer): -- GET cvsd/limit_coredumpsize cputime datasize filesize memorylocked openfiles maxproc memoryuse stacksize virtmem debconf (developer): -- 20 Incorrect number of arguments zsh: exit 20DEBCONF_DEBUG=. dpkg-reconfigure cvsd Looks like a bug.. After doing this I also did not see all the limits stuff in cvsd.conf, it had only these config items: Uid cvsd Gid cvsd PidFile /var/run/cvsd.pid RootJail /var/lib/cvsd MaxConnections 10 Nice 1 Listen * 2401 You support backing up; very good job there! Looking at your priorities, I'm not sure why the maximum connection limit is at priority medium. Surely there is a sane default, and so it could be low priority? I'm iffy about asking about the jail directory at medium priority; I suppose many people will want to configure that, but /var/lib/cvsd seems reasonable. Medium is probably ok. I agree with your use of high priority for the repositories question, and all the rest look nice and low. I've set it up to ask first whether to use debconf to manage configuration of /etc/cvsd/cvsd.conf (not a conffile). If the user chooses to not use debconf, he's on his own (an example cvsd.conf is provided). Of course this is where in my opinion you lose. First of all, as Colin pointed out recently, any
Re: debconf review of cvsd (was Re: stop abusing debconf already)
One more thing that I didn't notice until purging the package. In the purge question, you refer to selecting yes and answering no. Don't do that, some debconf frontends do not use yes or no; the user might be staring at a check box when they see that text. Just ask the question, something like (changebarred): You may choose to remove the chroot jail but you will also loose all the repositories inside the chroot jail. If you have not | backed up your repositories you want to keep, do not remove it now; | manually remove it later once your repositories are safe. | If you do chose to remove the chroot directory, all directories under | it will be removed, even if they are on another filesystem. If you choose to keep the chroot jail please note that the cvsd user and group will be removed so uid and gid file information may no longer be consistent. And while it's probably ok to default it to true, the postrm should catch the db_input return code, and if the question is not displayed, do not remove anything (and say something to stderr about it maybe). Think of someone in noninteractive mode.. Good use of a debconf substitution in that last question, BTW. -- see shy jo pgpJf89k6aw5V.pgp Description: PGP signature
Re: stop abusing debconf already
On 19-Apr-03, 06:47 (CDT), Steve Kowalik [EMAIL PROTECTED] wrote: At 7:22 pm, Saturday, April 19 2003, Denis Barbier mumbled: I do not understand exactly what is good and bad use of debconf. For instance all questions asked by the debconf package have good default values, so there is no reason to prompt user, a configuration file is enough. So what am I missing? Well, not all use of debconf is bad. For example, libnet-perl is a terrible misuse of debconf, *but* it can be remedied by dropping the priority of the questions from medium to low. Huh? If all the questions it asks can be converted to priority low, then that means there are reasonable defaults, and therefore IT SHOULDN'T BE USING DEBCONF AT ALL. How hard is this? What is so unclear about Policy on this topic? Steve -- Steve Greenland The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net
Re: stop abusing debconf already
Andre Luis Lopes wrote: Could you (or someone else) give me a hint on where one could find Joey's talk ? I've already tried googling for it and looking at [1] but couldn't find it. Hmm, I don't have it online that I know of, it was mostly extemporaneous anyway. (Here, I've linked the slides online at http://kitenet.net/~joey/debconf-debconf) -- see shy jo pgpCvDzqK9Bd0.pgp Description: PGP signature
Re: stop abusing debconf already
On Sat Apr 19, 11:18am -0500, Steve Greenland wrote: On 19-Apr-03, 06:47 (CDT), Steve Kowalik [EMAIL PROTECTED] wrote: At 7:22 pm, Saturday, April 19 2003, Denis Barbier mumbled: I do not understand exactly what is good and bad use of debconf. For instance all questions asked by the debconf package have good default values, so there is no reason to prompt user, a configuration file is enough. So what am I missing? Well, not all use of debconf is bad. For example, libnet-perl is a terrible misuse of debconf, *but* it can be remedied by dropping the priority of the questions from medium to low. Huh? If all the questions it asks can be converted to priority low, then that means there are reasonable defaults, and therefore IT SHOULDN'T BE USING DEBCONF AT ALL. How hard is this? What is so unclear about Policy on this topic? From debconf-devel(8): low: Very trivial items that have defaults that will work in the vast majority of cases; oinly control freaks see these. pgpWKYL2Wz8BY.pgp Description: PGP signature
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 12:36:04PM -0400, Joey Hess wrote: Andre Luis Lopes wrote: Could you (or someone else) give me a hint on where one could find Joey's talk ? I've already tried googling for it and looking at [1] but couldn't find it. Hmm, I don't have it online that I know of, it was mostly extemporaneous anyway. (Here, I've linked the slides online at http://kitenet.net/~joey/debconf-debconf) Hmm, future plans really seems to be quite interesting. Is there a mailing list dedicated to discussing debconf ideas and implementation I could subscribe to ? I saw that there's a link to an ancient Config mailing list at kitenet, but it seems not to be active anymore since all that I can see from its recent archives are lots of spam. Better integration with dpkg, a SQL backend, debconf-2.0 and almost everything which was said in future plans seems to be too much interesting for us to be out of the fun :-) Maybe a debian-debconf mailing list could be created in order to host future debconf discussions ? -- ++--++ || André Luís Lopes [EMAIL PROTECTED] || || Debian-BR Project http://debian-br.cipsga.org.br || || Public GPG KeyID 9D1B82F6 || || Keyserver wwwkeys.eu.pgp.net || pgpYjkZRfPmmY.pgp Description: PGP signature
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 12:36:04PM -0400, Joey Hess wrote: Andre Luis Lopes wrote: Could you (or someone else) give me a hint on where one could find Joey's talk ? I've already tried googling for it and looking at [1] but couldn't find it. Hmm, I don't have it online that I know of, it was mostly extemporaneous anyway. (Here, I've linked the slides online at http://kitenet.net/~joey/debconf-debconf) I thought there was a video recording of it - that'd be Scott Dier's department, wouldn't it? -- Colin Watson [EMAIL PROTECTED]
Re: stop abusing debconf already
Or maybe realize that Joey might perhaps know what he's talking about with regard to debconf ... you could go find the text of his talk at the last Debian Conference if you like. I realise he has an opinion on how things should be done. Depending on your own viewpoint this may be more influential than others as he is the author of the tool. As far as I'm concerned I'll use debconf how I please and if that's against the 'pure' view of others (ie/ *never* call it a registry) then its just hard luck. Policy is what matters not the opinion of some jumped up developers! Matt.
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 08:17:16PM +0100, Matt Ryan wrote: Colin Watson wrote: Or maybe realize that Joey might perhaps know what he's talking about with regard to debconf ... you could go find the text of his talk at the last Debian Conference if you like. I realise he has an opinion on how things should be done. Depending on your own viewpoint this may be more influential than others as he is the author of the tool. As far as I'm concerned I'll use debconf how I please and if that's against the 'pure' view of others (ie/ *never* call it a registry) then its just hard luck. Remember that the thread which spawned this is about configuration file handling which goes against policy. I think that's the guts of the brokenness. Policy is what matters not the opinion of some jumped up developers! BTW the opinion of this jumped-up developer is please don't send me private copies of posts to mailing lists. Thanks. -- Colin Watson [EMAIL PROTECTED]
Re: stop abusing debconf already
On Sat, 2003-04-19 at 15:17, Matt Ryan wrote: Or maybe realize that Joey might perhaps know what he's talking about with regard to debconf ... you could go find the text of his talk at the last Debian Conference if you like. I realise he has an opinion on how things should be done. Depending on your own viewpoint this may be more influential than others as he is the author of the tool. As far as I'm concerned I'll use debconf how I please and if that's against the 'pure' view of others (ie/ *never* call it a registry) then its just hard luck. No offense, but I think you joined the wrong project, then.
Re: stop abusing debconf already
On Sat, Apr 19, 2003 at 08:17:16PM +0100, Matt Ryan wrote: Or maybe realize that Joey might perhaps know what he's talking about with regard to debconf ... you could go find the text of his talk at the last Debian Conference if you like. I realise he has an opinion on how things should be done. Depending on your own viewpoint this may be more influential than others as he is the author of the tool. As far as I'm concerned I'll use debconf how I please and if that's against the 'pure' view of others (ie/ *never* call it a registry) then its just hard luck. Um, no. *Policy* says that it may not be used as a registry. - Debconf answers are stored in /var/cache/debconf. Under the FHS, which is a mandatory part of Policy, the rules for /var/cache are that [the] application must be able to regenerate or restore the data, and the cached files can be deleted without data loss. *This precludes using debconf as a registry, and is by design*. Since the position of the debconf maintainer is that debconf should not be used as a registry, if you use it as a registry then *your* package, not debconf, is in violation of Policy. - Policy section 11.7.3 requires that local changes [to configuration files] must be preserved during a package upgrade. Using debconf as a registry implies giving precedence to debconf over the contents of an edited configuration file. THIS IS A VIOLATION OF POLICY. Therefore, even if you resolve the above issue, Policy only allows you to use debconf as a registry *for information that is never written to a config file*. I'm not sure why you think Joey's expertise doesn't qualify him to make pronouncements about the use of debconf. Unlike you, he at least gets the answer right. -- Steve Langasek postmodern programmer pgphbIX36Wg9R.pgp Description: PGP signature
Re: stop abusing debconf already
Andre Luis Lopes wrote: Hmm, future plans really seems to be quite interesting. Is there a mailing list dedicated to discussing debconf ideas and implementation I could subscribe to ? I saw that there's a link to an ancient Config mailing list at kitenet, but it seems not to be active anymore since all that I can see from its recent archives are lots of spam. Better integration with dpkg, a SQL backend, debconf-2.0 and almost everything which was said in future plans seems to be too much interesting for us to be out of the fun :-) Maybe a debian-debconf mailing list could be created in order to host future debconf discussions ? Yes the config list is dead. Why don't you set up a list on alioth. debconf-2.0 is only waiting on the latest debconf getting into testing before we begin that transition. All the rest is waiting on someone to do the work. -- see shy jo pgpm6V81JoUl8.pgp Description: PGP signature
stop abusing debconf already
Enough already. Folks, if you don't stop abusing debconf with useless notes that belong in README.Debian and config file overwriting, I will stop maintaining it. Stop slapping incorrect uses of debconf in everywhere. Feel free to run any package using debconf by me before you upload it, or take the time to understand yourself how things should work. -- see shy jo pgpoqSfGnHPJ8.pgp Description: PGP signature