Re: Dangers of "stable" in sources.list
On Fri, May 04, 2007 at 03:49:08AM -0400, Kevin Mark wrote: > why I'd never use it. So I'd be for one of these two: > -removing the public link to 'stable' > -putting a strong warning in the Debian reference about the hazards of > using it. > So if someone uses 'stable', do tell. And if so, would you want a newbie > to use it? > -k If I'm building machines for work or anywhere that needs machines to stay static for a long time: Change to release name _at the time of release_ to make sure that I actually get the "new" stable. Immediately I'm installed - switch it to stable. If I then come back to the machine six months later - cat /etc/apt/sources.list - think "OK, this one was installed as stable and will just run". Code names were/are a convenience for developers and discussion and widespread use came about following Debian 1.0 :) As my fairly regular message goes: stable/testing/unstable don't necessarily refer to stability of the apps. They do relate to how fast/far the archive changes over time: unstable may mean large package churn daily/weekly and testing will periodically receive very large changes. Stable, however, once released is designed to be unconditionally stable - new features/versions are NOT introduced. Security fixes are back ported to avoid unpleasant surprises mid-release cycle. Stable serves a very useful purpose: it reminds me that the machine is meant to be stable - "Etch" may not mean anything to anyone else but "stable" does, especially to the ears of management :) Two years from now, you may be struggling to recall codenames immediately but you will know the difference between Debian 3.0, 3.1 and 4.0 It might be an idea to change policy for stable on just one package: with the final release of say, Sarge (which came a few hours before the release of Etch) change the base configuration files so that they and the /etc/apt/sources.list flip to oldstable. If the change is also publicised on the mailing lists etc. - those who don't want to get hit by the upgrade don't have to follow it immediately. If need be, add an optional little package "staystable" which does this change which conflicts/replaces "trackrelease" (this might be handled with a debconf question as an alternative). Andy > -- > | .''`. == Debian GNU/Linux == | my web site: | > | : :' : The Universal |mysite.verizon.net/kevin.mark/| > | `. `' Operating System| go to counter.li.org and | > | `-http://www.debian.org/ |be counted! #238656 | > | my keyserver: subkeys.pgp.net | my NPO: cfsg.org | > |join the new debian-community.org to help Debian! | > |___ Unless I ask to be CCd, assume I am subscribed ___| > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Fri, May 04, 2007 at 12:18:37AM -0400, Greg Folkert wrote: > On Thu, 2007-05-03 at 21:44 -0400, Max Hyre wrote: > > Gentlefolk: > > > >The discussion of `stable' vs. `etch' vs. `lenny' > > vs. ... got me to thinking. Is there any reason to offer > > `stable' as an entry in sources.list? Its drawback seems to > > be: > > > > o Every so often `stable' whacks you with about > > seventeen million updates, with the chance that you'll > > be left dead in the water. > > > > Using the name (`sarge', e.g.) has the drawback that: > > > > o Eventually a named distro will drop off the end of the > > world, and get no more security updates. > > > > OTOH, `unstable' is a necessary warning sign: Here be > > dragons. Someone starting with Debian needs to know that > > unstable has more surprises. (Though, in my experience, > > they're mostly like the ones you find in a box of Cracker > > Jacks.) > > > >So, my modest suggestion is that `stable' as a name > > should be eradicated. Roughly no downside, only closer > > adherence to the principle of least astonishment. > > Okay, so let me get this straight. > > You propose to eliminate "stable" as a release. To keep people from > hurting themselves. Especially unwitting "auto-updating" ID10Ts. Ok, let > me get this straight... How is this a good thing? I find the unstable/sid, $RELEASE_NAME. and testing tags useful. But I dont recall any use for using 'stable' in a source list. If you use unstable/sid, it need constant monitoring, and this is what all expect. Using $RELEASE_NAME, at or after a release, gives a useful result. Using a $RELEASE_NAME, before a release, gives different results depending upon when you use them in the release cycle, but if carefully used, can give reasonable results. 'testing' is similar to $RELEASE_NAME when used before a release but can lead to some more issues at release time, and thus is less useful. And similarly, stable, is similar to $RELEASE_NAME, but has far worse results if you dont watch the release cycle, which is why I'd never use it. So I'd be for one of these two: - -removing the public link to 'stable' - -putting a strong warning in the Debian reference about the hazards of using it. So if someone uses 'stable', do tell. And if so, would you want a newbie to use it? - -k - -- | .''`. == Debian GNU/Linux == | my web site: | | : :' : The Universal |mysite.verizon.net/kevin.mark/| | `. `' Operating System| go to counter.li.org and | | `-http://www.debian.org/ |be counted! #238656 | | my keyserver: subkeys.pgp.net | my NPO: cfsg.org | |join the new debian-community.org to help Debian! | |___ Unless I ask to be CCd, assume I am subscribed ___| -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFGOuV0v8UcC1qRZVMRAugCAJ9c8KTQ0fTHlz7ibWka36YSgRyUgACfdAbI AlXCW/D5bWVghDeXYqU3HvA= =0X7D -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On Fri, May 04, 2007 at 06:12:48PM -0400, Roberto C. Sánchez wrote: > On Fri, May 04, 2007 at 06:06:01PM -0400, Max Hyre wrote: > > > >I'm pointing out that the `stable' distro becomes > > massively unstable periodically. Admitted, that period is > > on the order of multiple years, but it _is_ being shortened. > > Additionally, the people least likely to be able to handle a > > badly- or non-working system are the ones most likely to be > > blind-sided by it. > > > Acutally, the stable distro does not become massively unstable. We have > things called releases. It turns out that these releases are announced > well in advance. There is lots of list traffic about them. They even > make it to the debian-announce list. > this is a side-effect of growing adoption of linux. People who would not be bothered to harden their windows installation and properly maintain it are likely to be the same way with their linux installation. The result is that people may very likely learn just enough to be dangerous to themselves through things like an ill-timed dist-upgrade. By no means does that absolve them of their responsibility to deal with the mess they may make by not properly researching what they do. I am not accusing anyone here of this behavior, btw. A signature.asc Description: Digital signature
Re: Dangers of "stable" in sources.list
Max Hyre <[EMAIL PROTECTED]> writes: >I'm pointing out that the `stable' distro becomes > massively unstable periodically. Admitted, that period is > on the order of multiple years, but it _is_ being shortened. > Additionally, the people least likely to be able to handle a > badly- or non-working system are the ones most likely to be > blind-sided by it. > > John L Fjellstad wrote: >> If you remove 'stable', then you kind of have to remove >> 'testing' too. > >I don't think that follows. If someone, even a newbie, > signs up for `testing', they've got to know there's going to > be a bumpy ride. Just after a distro goes stable, testing will get an influx of packages that might break the system. And the way it's set up, it might take some time for the fixes to come down from unstable. testing is pretty calm after a freeze, and since freezes are pretty long in Debian, some newbies might not realize this if they started with testing during the frozen period. And the people most likely to to not being able to handle a non-working system is probably not updateing willy-nilly anyways. -- John L. Fjellstad web: http://www.fjellstad.org/ Quis custodiet ipsos custodes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On Fri, May 04, 2007 at 06:06:01PM -0400, Max Hyre wrote: > >I'm pointing out that the `stable' distro becomes > massively unstable periodically. Admitted, that period is > on the order of multiple years, but it _is_ being shortened. > Additionally, the people least likely to be able to handle a > badly- or non-working system are the ones most likely to be > blind-sided by it. > Acutally, the stable distro does not become massively unstable. We have things called releases. It turns out that these releases are announced well in advance. There is lots of list traffic about them. They even make it to the debian-announce list. Regards, -Roberto -- Roberto C. Sánchez http://people.connexer.com/~roberto http://www.connexer.com signature.asc Description: Digital signature
Re: Dangers of "stable" in sources.list
On Fri, May 04, 2007 at 12:18:37AM -0400, Greg Folkert wrote: > > You propose to eliminate "stable" as a release. To keep > > people from hurting themselves. Especially unwitting > > "auto-updating" ID10Ts. Ok, let me get this > > straight... How is this a good thing? Paul Condon opined: > Greg, > In fairness to Max, he surely did not have in mind > possible confusion by an experienced Admin who is new to > Debian, but rather a simple user who is working in an > environment where he must act as his own Admin. Yeah, what he said. :-) I'm pointing out that the `stable' distro becomes massively unstable periodically. Admitted, that period is on the order of multiple years, but it _is_ being shortened. Additionally, the people least likely to be able to handle a badly- or non-working system are the ones most likely to be blind-sided by it. Paul E Condon wrote: > To reorganize the internals would involve a lot of work. Touche'. John L Fjellstad wrote: > If you remove 'stable', then you kind of have to remove > 'testing' too. I don't think that follows. If someone, even a newbie, signs up for `testing', they've got to know there's going to be a bumpy ride. -- Best wishes, Max Hyre signature.asc Description: OpenPGP digital signature
Re: Dangers of "stable" in sources.list
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Jan Sneep wrote: [snip] >> just a guess, but maybe so that no matter when you install, that >> install disk will get you moving into stable. so you could use a >> really old installer and automatically move right up to stable with >> the next dist-upgrade. > > I had a NetInstall CD of Sarge that I made in January and when I did the > update last week I lost everything. I found I couldn't use that CD it get > Etch installed. It would crap-out because it was trying to replace the > kernel from the CD with the one from the mirror and kaboom! No go. Funny, I recently had a similar situation and I could use the sarge CD to install stable etch just fine. It would install the packages (including the kernel) from the CD and just replace them once it updated the package list over the net. >> also, ISTM, that if you are paying attention at all, you'd notice when >> the change happened. If you do a regular dist-upgrade, there will >> suddenly be a pile of upgrades instead of the usual trickle. THat >> should be enough to cause one to review what is happening and hold off >> on an upgrade if its necessary. >> > Not really a, IMHO, very good way to make sure someone doesn't make the > mistake I made ... REALLY should have some check built in to ALL the various > methods one can use to do an update to FLAG that you are about to perform > and UPGRADE not just an UPDATE. I hadn't done an update since I installed > Sarge back in January, because I didn't know how. Read some of the posts to > this list that in effect said that Aptitude was the way to go, so read what > I though I needed from the user guide and it's "marketing" message confirmed > that Aptitude was the next-best-thing-to-sliced-bread as it would manage all > the dependancies automatically that with other methods you would have to do > manually. So I with great confidence told it to go and update everything for > me! Well as you know I lost everything. Had to get a new Etch NetInst CD and Well I'm sure you didn't loose 'everything'. Maybe aptitude 'accidentially' removed more packages than you would have liked, but instead of reinstalling from scratch you'd probably just have installed the packages you want manually. aptitude install xserver-xorg kde gnome whatever > start over again. The ONLY good side to this is that I'm getting pretty good > at running the NetInst CD and have almost all the prompts memorized ... :O) To be fair: as far as the proprietary OS that you seem to be using for sending your mails to this list is concerned, it certainly does not even have the capabilities to upgrade your system to the next release and it won't keep track of all the software you have installed and it won't check for security updates for all the software you have installed on your system and it won't etc. etc. etc. , *but* it has the capabilities to break your installation. And if it's broke there is *no* way to fix it, except for a fresh install from scratch. > I guess if we're taking votes ... cast my newbie vote for using the name > rather than "stable" ... As pointed out by Joey Hess it's been changed already. Cheers, Johannes -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFGOwWqC1NzPRl9qEURAsmVAJ4i8DzpPrgkwKDUCXlVvs3x2lxDcACfVF3q MCguGgOVpT1eYO/Vo9FtGks= =Sxpy -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On Thu, May 03, 2007 at 09:44:10PM -0400, Max Hyre wrote: >The discussion of `stable' vs. `etch' vs. `lenny' > vs. ... got me to thinking. Is there any reason to offer > `stable' as an entry in sources.list? Its drawback seems to > be: > > o Every so often `stable' whacks you with about > seventeen million updates, with the chance that you'll > be left dead in the water. As others have mentioned: This only happens if your admin blindly installs all the updates available. > Using the name (`sarge', e.g.) has the drawback that: > > o Eventually a named distro will drop off the end of the > world, and get no more security updates. [snip unstable] >So, my modest suggestion is that `stable' as a name > should be eradicated. Roughly no downside, only closer > adherence to the principle of least astonishment. I believe that this "astonishment" comes from a basic misunderstanding. - Referencing any of stable/testing/unstable/sid means "follow the debian releases" - Referencing any of hamm/woody/sarge/etch means "Stick with $codename and hope for security updates" There is a need to be able to express both - different sysadmins, different skills, different preferences etc etc. Whether the default install should say "stable" or "etch" is a different matter. Don't be surprised if the current lenny (or any future RC) says "testing" though... > [Runs for blast shelter...] You may be assimilated. Resistance will provide greater knowledge. -- Karl E. Jorgensen [EMAIL PROTECTED] http://www.jorgensen.org.uk/ [EMAIL PROTECTED] http://karl.jorgensen.com Today's fortune: Nullum magnum ingenium sine mixtura dementiae fuit. [There is no great genius without some touch of madness.] -- Seneca signature.asc Description: Digital signature
Re: Dangers of "stable" in sources.list
On Fri, May 04, 2007 at 12:18:37AM -0400, Greg Folkert wrote: > On Thu, 2007-05-03 at 21:44 -0400, Max Hyre wrote: > > Gentlefolk: > > > >The discussion of `stable' vs. `etch' vs. `lenny' > > vs. ... got me to thinking. Is there any reason to offer > > `stable' as an entry in sources.list? Its drawback seems to > > be: > > > > o Every so often `stable' whacks you with about > > seventeen million updates, with the chance that you'll > > be left dead in the water. > > > > Using the name (`sarge', e.g.) has the drawback that: > > > > o Eventually a named distro will drop off the end of the > > world, and get no more security updates. > > > > OTOH, `unstable' is a necessary warning sign: Here be > > dragons. Someone starting with Debian needs to know that > > unstable has more surprises. (Though, in my experience, > > they're mostly like the ones you find in a box of Cracker > > Jacks.) > > > >So, my modest suggestion is that `stable' as a name > > should be eradicated. Roughly no downside, only closer > > adherence to the principle of least astonishment. > > Okay, so let me get this straight. > > You propose to eliminate "stable" as a release. To keep people from > hurting themselves. Especially unwitting "auto-updating" ID10Ts. Ok, let > me get this straight... How is this a good thing? > > Anyone that is an Admin worth the salt they have in their body, will not > have ANYTHING auto-updating, but "auto-staging". > Greg, In fairness to Max, he surely did not have in mind possible confusion by an experienced Admin who is new to Debian, but rather a simple user who is working in an environment where he must act as his own Admin. For many new users of Debian, Debian is their first opportunity to see 'under the hood' of a real OS. For such users, Debian stable is good enough to be miles ahead of what they had before. The issue is the use of the word, 'stable', in a special jargon sense that is peculiar to Debian. The idea of suppressing 'stable' seems reasonable, except that it will be very difficult to do. So, these new users will soon see references to 'stable' all over the place, and will be put off trying to actually learn how the Debian system works. (Think of project Ministry of Truth, rewriting all the old emails to the debian-user list that mention 'stable'.) > My definition of "Auto-Staging" means: > > 1. Check for "updates" > 2. Update for "updates" > 3. Download all of those updated packages > 4. Checksum verify those packages, just downloaded re-get the ones > not verified properly > 5. Send off an e-mail to me every hour or two telling me I need to > pay some attention to it. > 6. goto #5 every one or two hours > 7. goto #1 every six or twelve hours > > I then login to said machine and "do the right thing" > > I've done this for years with HP-UX, AIX, OSF/Tru64/WTFitis when each > company finally put patches and stuff in a internet reachable > repository. I really haven't done Solaris, so I dunno about it, but I > suspect it could be done similarly. > > Now as far as Windows... blahahaha. Yeah whatever. > -- Paul E Condon [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
Max Hyre <[EMAIL PROTECTED]> writes: >So, my modest suggestion is that `stable' as a name > should be eradicated. Roughly no downside, only closer > adherence to the principle of least astonishment. If you remove 'stable', then you kind of have to remove 'testing' too. Otherwise, people who follow testing will get a blast of new software once 'whatever' becomes stable. And how do you know which code name is the current stable? -- John L. Fjellstad web: http://www.fjellstad.org/ Quis custodiet ipsos custodes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On Thu, 2007-05-03 at 21:44 -0400, Max Hyre wrote: > Gentlefolk: > >The discussion of `stable' vs. `etch' vs. `lenny' > vs. ... got me to thinking. Is there any reason to offer > `stable' as an entry in sources.list? Its drawback seems to > be: > > o Every so often `stable' whacks you with about > seventeen million updates, with the chance that you'll > be left dead in the water. > > Using the name (`sarge', e.g.) has the drawback that: > > o Eventually a named distro will drop off the end of the > world, and get no more security updates. > > OTOH, `unstable' is a necessary warning sign: Here be > dragons. Someone starting with Debian needs to know that > unstable has more surprises. (Though, in my experience, > they're mostly like the ones you find in a box of Cracker > Jacks.) > >So, my modest suggestion is that `stable' as a name > should be eradicated. Roughly no downside, only closer > adherence to the principle of least astonishment. Okay, so let me get this straight. You propose to eliminate "stable" as a release. To keep people from hurting themselves. Especially unwitting "auto-updating" ID10Ts. Ok, let me get this straight... How is this a good thing? Anyone that is an Admin worth the salt they have in their body, will not have ANYTHING auto-updating, but "auto-staging". My definition of "Auto-Staging" means: 1. Check for "updates" 2. Update for "updates" 3. Download all of those updated packages 4. Checksum verify those packages, just downloaded re-get the ones not verified properly 5. Send off an e-mail to me every hour or two telling me I need to pay some attention to it. 6. goto #5 every one or two hours 7. goto #1 every six or twelve hours I then login to said machine and "do the right thing" I've done this for years with HP-UX, AIX, OSF/Tru64/WTFitis when each company finally put patches and stuff in a internet reachable repository. I really haven't done Solaris, so I dunno about it, but I suspect it could be done similarly. Now as far as Windows... blahahaha. Yeah whatever. -- greg, [EMAIL PROTECTED] The technology that is Stronger, better, faster: Linux signature.asc Description: This is a digitally signed message part
Re: Dangers of "stable" in sources.list
On Thu, May 03, 2007 at 09:44:10PM -0400, Max Hyre wrote: >Gentlefolk: > >The discussion of `stable' vs. `etch' vs. `lenny' > vs. ... got me to thinking. Is there any reason to offer > `stable' as an entry in sources.list? Its drawback seems to > be: > > o Every so often `stable' whacks you with about > seventeen million updates, with the chance that you'll > be left dead in the water. > > Using the name (`sarge', e.g.) has the drawback that: > > o Eventually a named distro will drop off the end of the > world, and get no more security updates. > > OTOH, `unstable' is a necessary warning sign: Here be > dragons. Someone starting with Debian needs to know that > unstable has more surprises. (Though, in my experience, > they're mostly like the ones you find in a box of Cracker > Jacks.) > >So, my modest suggestion is that `stable' as a name > should be eradicated. Roughly no downside, only closer > adherence to the principle of least astonishment. > > [Runs for blast shelter...] > I'm not a Debian developer or maintainer, but I have looked at the structure of the file system of Debian repositories and found a lot of evidence that it grew rather than having been designed top down and according to some well structured method. Internally, there are places where a code name, such as 'etch' is treated as an alias for a status name such as 'stable', and other places where the reverse (status name is alias for code name) is true. To reorganize the internals would involve a lot of work. There are a lot of functioning programs that have knowledge of the existing structure hard coded in. But, it works. And the effort involved in training new users in the meaning of 'stable' and how it relates to the meaning of 'etch' is really not great. JMTCW -- Paul E Condon [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
Gentlefolk: The discussion of `stable' vs. `etch' vs. `lenny' vs. ... got me to thinking. Is there any reason to offer `stable' as an entry in sources.list? Its drawback seems to be: o Every so often `stable' whacks you with about seventeen million updates, with the chance that you'll be left dead in the water. Using the name (`sarge', e.g.) has the drawback that: o Eventually a named distro will drop off the end of the world, and get no more security updates. OTOH, `unstable' is a necessary warning sign: Here be dragons. Someone starting with Debian needs to know that unstable has more surprises. (Though, in my experience, they're mostly like the ones you find in a box of Cracker Jacks.) So, my modest suggestion is that `stable' as a name should be eradicated. Roughly no downside, only closer adherence to the principle of least astonishment. [Runs for blast shelter...] -- Best wishes, Max Hyre signature.asc Description: OpenPGP digital signature
Re: Dangers of "stable" in sources.list
On Thu, 3 May 2007 11:24:56 -0400 "Jan Sneep" <[EMAIL PROTECTED]> wrote: > > > > -Original Message- > > From: Andrew Sackville-West [mailto:[EMAIL PROTECTED] > > Sent: May 3, 2007 10:44 AM > > To: debian-user@lists.debian.org > > Subject: Re: Dangers of "stable" in sources.list > > > > mildly humorous to think someone could be *surprised* by a debian > > release ;-O. > > Actually I was ... not sure if it was mentioned here on the list or not ? > > The list produces hundreds of postings a day and while I try to read most of > them ... there are days when work gets so busy and then there are six or > seven hundred messages to scan ... and while I remember reading something > about Etch being released in April, I totally missed the actual event ... > and I assume the correspondig instructions on how to properly do an Upgrade [snip] Subscribe to debian-announce [0], and pay special attention to its messages (there aren't a whole lot). [0] http://lists.debian.org/debian-announce/ Celejar -- mailmin.sourceforge.net - remote access via secure (OpenPGP) email ssuds.sourceforge.net - A Simple Sudoku Solver and Generator -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On 5/3/07, Joey Hess <[EMAIL PROTECTED]> wrote: martin f krafft wrote: > also sprach Martin Marcher <[EMAIL PROTECTED]> [2007.05.03.1434 +0200]: > > Exactly my opinion too, i was more interested in hearing why I would > > wan't stable instead of the hardcoded name. I just can't think of any > > reason to do that and practically have really use for it (except for > > the testing or unstable distribution as I pointed out initially) > > I agree. I suggest you file a bug against debian-installer. I suggest that people try a current version of debian-installer before fuling bugs on it. well it seems I broke something off here :) I should have specified a little clearer. I'm mainly refering to the servers I maintain. On my notebook i run osx for browsing and multimedia stuff and debian testing when i need to get things done. I tend to set up my servers so that they are somewhat "self maintaining" e.g. informing me about packages that could need updating strange events yadda yadda the usual stuff. And personally do use the codenames everywhere (except for testing on my notebook as I never had problems and it was a smooth transition at all times) But since I saw quite a few posts where the problems where exactly having stable in sources.list and thus relying on the long release cycles of debian caused these problems i'm trying to make something up so that stable in my sources.list could be an advantage. Maybe something like when /etc/debian_version changes from 3.1 to 4.0 let all the bells ring and alert a few people that there is some work to be done. Then again I guess the file won't change without actually updating so this is an chicken egg problem. As for the installer, I can't really remember wether by default stable or the codename is put into sources.list - but putting stable in there is very a) very well a bug b) absolutely not a but a) because with all the hype about linux that comes up with new releases of major distros newbies will be disappointed for sure if they update after a week and break their system despite it's stable and b) yes people should read documentation but we aren't in a perfect world, which brings us back to a) martin -- Martin Marcher [EMAIL PROTECTED] http://www.mycorners.com https://www.xing.com/profile/Martin_Marcher http://www.linkedin.com/in/martinmarcher http://www.studivz.net/profile.php?ids=9f83ea8c5996b8ec http://www.amazon.de/gp/registry/wishlist/3KDAGCL2NKOIM/ref=reg_hu-wl_goto-registry/302-4432803-5146435?ie=UTF8&sort=date-added -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
martin f krafft wrote: > also sprach Martin Marcher <[EMAIL PROTECTED]> [2007.05.03.1434 +0200]: > > Exactly my opinion too, i was more interested in hearing why I would > > wan't stable instead of the hardcoded name. I just can't think of any > > reason to do that and practically have really use for it (except for > > the testing or unstable distribution as I pointed out initially) > > I agree. I suggest you file a bug against debian-installer. I suggest that people try a current version of debian-installer before fuling bugs on it. -- see shy jo signature.asc Description: Digital signature
Re: Dangers of "stable" in sources.list
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Jan Sneep wrote: [snip] > > I just check my sources.list file and interestingly the default when doing a > clean NetInst is to point to the Etch folders on the miror site, not > "stable". > > No virus found in this outgoing message. > Checked by AVG Free Edition. > Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 2007.05.02 > 2:16 PM > > You just checked sources,list yet you use: X-Mailer: Microsoft Outlook CWS, Build 9.0.6604 (9.0.2911.0) Hmm, have to wonder. At least you're bright enough to have a semi-decent anti-virus program checking your mail. As for the topic at hand, it makes no difference whether one uses "stable" or "etch" until that is, Lenny goes stable. Of course, people who had Sarge and used the word "stable" would now have Etch, and without doing apt-get (or aptitude) dist-upgrade, could find their systems in a mess. A dist-upgrade should sort out most problems. Joe Joe - -- Registerd Linux user #443289 at http://counter.li.org/ -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGOhx0iXBCVWpc5J4RAoDsAKCInW1RRNuY/ITvPBw8hrKUaJxi2ACfa0ro yk93//y/n2quvletx4QOZcw= =D9UN -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: Dangers of "stable" in sources.list
> -Original Message- > From: Andrew Sackville-West [mailto:[EMAIL PROTECTED] > Sent: May 3, 2007 11:54 AM > To: debian-user@lists.debian.org > Subject: Re: Dangers of "stable" in sources.list > > > On Thu, May 03, 2007 at 11:24:56AM -0400, Jan Sneep wrote: > > > > > > > -Original Message- > > > From: Andrew Sackville-West [mailto:[EMAIL PROTECTED] > > > Sent: May 3, 2007 10:44 AM > > > To: debian-user@lists.debian.org > > > Subject: Re: Dangers of "stable" in sources.list > > > > > > mildly humorous to think someone could be *surprised* by a debian > > > release ;-O. > > > > Actually I was ... not sure if it was mentioned here on the > list or not ? > > well, the list traffic certainly went through the roof when it > happened including lots of "congratulations on the release" posts, but > really, I was only making a lame joke about the snail's pace of debian > releases. Something I fully get and appreciate, BTW. > > [...] > > > > > > > just a guess, but maybe so that no matter when you install, that > > > install disk will get you moving into stable. so you could use a > > > really old installer and automatically move right up to > stable with > > > the next dist-upgrade. > > > > > > > I had a NetInstall CD of Sarge that I made in January and > when I did the > > update last week I lost everything. I found I couldn't use > that CD it get > > Etch installed. It would crap-out because it was trying to > replace the > > kernel from the CD with the one from the mirror and kaboom! No go. > > > hmm.. yeah, I guess that makes sense, now that I actually apply my > brain to it. There are ways to work around it, but from the newbie > perspective, they are less than ideal. > > > > > > > > > also, ISTM, that if you are paying attention at all, > you'd notice when > > > the change happened. If you do a regular dist-upgrade, there will > > > suddenly be a pile of upgrades instead of the usual trickle. THat > > > should be enough to cause one to review what is happening > and hold off > > > on an upgrade if its necessary. > > > > > > > Not really a, IMHO, very good way to make sure someone > doesn't make the > > mistake I made ... REALLY should have some check built in > to ALL the various > > methods one can use to do an update to FLAG that you are > about to perform > > and UPGRADE not just an UPDATE. > > with all due respect, and only the best of intentions, the Apt system > always asks if you really want to do what you are about to do. (unless > you're installing 1 and only 1 package). I will agree though > that if one > isn't up to speed yet, they'd not recognise that upgrading a couple > hundred packages is not a normal situation for the stable > distribution. > > I hadn't done an update since I installed > > Sarge back in January, because I didn't know how. Read some > of the posts to > > this list that in effect said that Aptitude was the way to > go, so read what > > I though I needed from the user guide and it's "marketing" > message confirmed > > that Aptitude was the next-best-thing-to-sliced-bread as it > would manage all > > the dependancies automatically that with other methods you > would have to do > > manually. So I with great confidence told it to go and > update everything for > > me! Well as you know I lost everything. > > are you refering here to your netowrk problem and your reinstalling? > if so, its really not fair to blame that on the apt system :). > > > Had to get a new Etch NetInst CD and > > start over again. The ONLY good side to this is that I'm > getting pretty good > > at running the NetInst CD and have almost all the prompts > memorized ... :O) > > wait til you learn how to have it install everything automatically, > including answering all the questions :) > > > > > I guess if we're taking votes ... cast my newbie vote for > using the name > > rather than "stable" ... > > Don't get me wrong, I'm not suggesting one method over the other, just > trying to point out how I think it works. you are free to use > whichever method you choose. I just check my sources.list file and interestingly the default when doing a clean NetInst is to point to the Etch folders on the miror site, not "stable". No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 2007.05.02 2:16 PM -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
"Martin Marcher" <[EMAIL PROTECTED]> writes: > Do you generally use stable in your sources.list or do you actually > use sarge/etch/whatever. I usually use the code names, mostly because lately when I have installed a new system, the testing distro had gotten far enought hat I feel comfortable running it. Also, the stable has gotten so stale, I can't run it. That said, I try to keep servers that was installed long ago, on stable instead of codenames. By the time a new distro is ready, given Debian looong cycle, I'm not receiving much of updates of any kind anymore. That means when the new distro goes stable, it's never much of a surprise. Then again, I don't feel like running "aptitude update && aptitude dist-upgrade" every day, especially on my servers. -- John L. Fjellstad web: http://www.fjellstad.org/ Quis custodiet ipsos custodes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On Thu, May 03, 2007 at 11:24:56AM -0400, Jan Sneep wrote: > > > > -Original Message- > > From: Andrew Sackville-West [mailto:[EMAIL PROTECTED] > > Sent: May 3, 2007 10:44 AM > > To: debian-user@lists.debian.org > > Subject: Re: Dangers of "stable" in sources.list > > > > mildly humorous to think someone could be *surprised* by a debian > > release ;-O. > > Actually I was ... not sure if it was mentioned here on the list or not ? well, the list traffic certainly went through the roof when it happened including lots of "congratulations on the release" posts, but really, I was only making a lame joke about the snail's pace of debian releases. Something I fully get and appreciate, BTW. [...] > > > > just a guess, but maybe so that no matter when you install, that > > install disk will get you moving into stable. so you could use a > > really old installer and automatically move right up to stable with > > the next dist-upgrade. > > > > I had a NetInstall CD of Sarge that I made in January and when I did the > update last week I lost everything. I found I couldn't use that CD it get > Etch installed. It would crap-out because it was trying to replace the > kernel from the CD with the one from the mirror and kaboom! No go. hmm.. yeah, I guess that makes sense, now that I actually apply my brain to it. There are ways to work around it, but from the newbie perspective, they are less than ideal. > > > > also, ISTM, that if you are paying attention at all, you'd notice when > > the change happened. If you do a regular dist-upgrade, there will > > suddenly be a pile of upgrades instead of the usual trickle. THat > > should be enough to cause one to review what is happening and hold off > > on an upgrade if its necessary. > > > > Not really a, IMHO, very good way to make sure someone doesn't make the > mistake I made ... REALLY should have some check built in to ALL the various > methods one can use to do an update to FLAG that you are about to perform > and UPGRADE not just an UPDATE. with all due respect, and only the best of intentions, the Apt system always asks if you really want to do what you are about to do. (unless you're installing 1 and only 1 package). I will agree though that if one isn't up to speed yet, they'd not recognise that upgrading a couple hundred packages is not a normal situation for the stable distribution. I hadn't done an update since I installed > Sarge back in January, because I didn't know how. Read some of the posts to > this list that in effect said that Aptitude was the way to go, so read what > I though I needed from the user guide and it's "marketing" message confirmed > that Aptitude was the next-best-thing-to-sliced-bread as it would manage all > the dependancies automatically that with other methods you would have to do > manually. So I with great confidence told it to go and update everything for > me! Well as you know I lost everything. are you refering here to your netowrk problem and your reinstalling? if so, its really not fair to blame that on the apt system :). Had to get a new Etch NetInst CD and > start over again. The ONLY good side to this is that I'm getting pretty good > at running the NetInst CD and have almost all the prompts memorized ... :O) wait til you learn how to have it install everything automatically, including answering all the questions :) > > I guess if we're taking votes ... cast my newbie vote for using the name > rather than "stable" ... Don't get me wrong, I'm not suggesting one method over the other, just trying to point out how I think it works. you are free to use whichever method you choose. A signature.asc Description: Digital signature
Re: Dangers of "stable" in sources.list
On Thu, 3 May 2007 11:24:56 -0400 "Jan Sneep" <[EMAIL PROTECTED]> wrote: > I had a NetInstall CD of Sarge that I made in January and when I did > the update last week I lost everything. I found I couldn't use that > CD it get Etch installed. It would crap-out because it was trying to > replace the kernel from the CD with the one from the mirror and > kaboom! No go. I had a similar problem. I used the last RC netinstall disc to install post-release Etch. Or so I thought: the installer had put 'testing' rather that 'etch' in the apt sources, so it had in fact fetched the base packages belonging to what is now lenny. Luckily I spotted the problem immediately, and downloaded a new netinstall disc. -- Liam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: Dangers of "stable" in sources.list
> -Original Message- > From: Andrew Sackville-West [mailto:[EMAIL PROTECTED] > Sent: May 3, 2007 10:44 AM > To: debian-user@lists.debian.org > Subject: Re: Dangers of "stable" in sources.list > > mildly humorous to think someone could be *surprised* by a debian > release ;-O. Actually I was ... not sure if it was mentioned here on the list or not ? The list produces hundreds of postings a day and while I try to read most of them ... there are days when work gets so busy and then there are six or seven hundred messages to scan ... and while I remember reading something about Etch being released in April, I totally missed the actual event ... and I assume the correspondig instructions on how to properly do an Upgrade ... :O( > > > > > And ever since, I've wondered why the default sources.list > specifies > > "stable" instead of a specific distribution. It seems like > a recipe for > > disaster for an awful lot of people. > > just a guess, but maybe so that no matter when you install, that > install disk will get you moving into stable. so you could use a > really old installer and automatically move right up to stable with > the next dist-upgrade. > I had a NetInstall CD of Sarge that I made in January and when I did the update last week I lost everything. I found I couldn't use that CD it get Etch installed. It would crap-out because it was trying to replace the kernel from the CD with the one from the mirror and kaboom! No go. > also, ISTM, that if you are paying attention at all, you'd notice when > the change happened. If you do a regular dist-upgrade, there will > suddenly be a pile of upgrades instead of the usual trickle. THat > should be enough to cause one to review what is happening and hold off > on an upgrade if its necessary. > Not really a, IMHO, very good way to make sure someone doesn't make the mistake I made ... REALLY should have some check built in to ALL the various methods one can use to do an update to FLAG that you are about to perform and UPGRADE not just an UPDATE. I hadn't done an update since I installed Sarge back in January, because I didn't know how. Read some of the posts to this list that in effect said that Aptitude was the way to go, so read what I though I needed from the user guide and it's "marketing" message confirmed that Aptitude was the next-best-thing-to-sliced-bread as it would manage all the dependancies automatically that with other methods you would have to do manually. So I with great confidence told it to go and update everything for me! Well as you know I lost everything. Had to get a new Etch NetInst CD and start over again. The ONLY good side to this is that I'm getting pretty good at running the NetInst CD and have almost all the prompts memorized ... :O) I guess if we're taking votes ... cast my newbie vote for using the name rather than "stable" ... Cheers, Jan No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 2007.05.02 2:16 PM -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
On Thu, May 03, 2007 at 08:49:24AM -0400, Miles Fidelman wrote: > Martin Marcher wrote: > >Hi, > > > >On 5/3/07, martin f krafft <[EMAIL PROTECTED]> wrote: > >>also sprach Martin Marcher <[EMAIL PROTECTED]> > >>[2007.05.03.1217 +0200]: > >>> So what are the hints wether to use stable the actual name or not? > >> > >>From my book: > >> > > > >Exactly my opinion too, i was more interested in hearing why I would > >wan't stable instead of the hardcoded name. I just can't think of any > >reason to do that and practically have really use for it (except for > >the testing or unstable distribution as I pointed out initially) > > > >Am I solving problems here that wouldn't exist in the first place? > >Am I ignorant of why there is a "stable" tree (apart from the > >psychological impact that it now indeed is stable)? > > > As soon as I read the recommendation in Martin Krafft's book, I > immediately changed all the lines in my sources.list from stable to > sarge - if I hadn't, I expect I would have been very unpleasantly > surprised upon doing my first apt-get upgrade following Etch becoming > stable. Thanks Martin! mildly humorous to think someone could be *surprised* by a debian release ;-O. > > And ever since, I've wondered why the default sources.list specifies > "stable" instead of a specific distribution. It seems like a recipe for > disaster for an awful lot of people. just a guess, but maybe so that no matter when you install, that install disk will get you moving into stable. so you could use a really old installer and automatically move right up to stable with the next dist-upgrade. also, ISTM, that if you are paying attention at all, you'd notice when the change happened. If you do a regular dist-upgrade, there will suddenly be a pile of upgrades instead of the usual trickle. THat should be enough to cause one to review what is happening and hold off on an upgrade if its necessary. .02 A signature.asc Description: Digital signature
Re: Dangers of "stable" in sources.list
also sprach Martin Marcher <[EMAIL PROTECTED]> [2007.05.03.1434 +0200]: > Exactly my opinion too, i was more interested in hearing why I would > wan't stable instead of the hardcoded name. I just can't think of any > reason to do that and practically have really use for it (except for > the testing or unstable distribution as I pointed out initially) I agree. I suggest you file a bug against debian-installer. -- Please do not send copies of list mail to me; I read the list! .''`. martin f. krafft <[EMAIL PROTECTED]> : :' : proud Debian developer, author, administrator, and user `. `'` http://people.debian.org/~madduck - http://debiansystem.info `- Debian - when you have better things to do than fixing systems "we are like shop windows in which we are continually arranging, concealing or illuminating the supposed qualities other ascribe to us -- in order to deceive ourselves." - friedrich nietzsche signature.asc Description: Digital signature (GPG/PGP)
Re: Dangers of "stable" in sources.list
Martin Marcher wrote: Hi, On 5/3/07, martin f krafft <[EMAIL PROTECTED]> wrote: also sprach Martin Marcher <[EMAIL PROTECTED]> [2007.05.03.1217 +0200]: > So what are the hints wether to use stable the actual name or not? From my book: Exactly my opinion too, i was more interested in hearing why I would wan't stable instead of the hardcoded name. I just can't think of any reason to do that and practically have really use for it (except for the testing or unstable distribution as I pointed out initially) Am I solving problems here that wouldn't exist in the first place? Am I ignorant of why there is a "stable" tree (apart from the psychological impact that it now indeed is stable)? As soon as I read the recommendation in Martin Krafft's book, I immediately changed all the lines in my sources.list from stable to sarge - if I hadn't, I expect I would have been very unpleasantly surprised upon doing my first apt-get upgrade following Etch becoming stable. Thanks Martin! And ever since, I've wondered why the default sources.list specifies "stable" instead of a specific distribution. It seems like a recipe for disaster for an awful lot of people. Miles -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
Hi, On 5/3/07, martin f krafft <[EMAIL PROTECTED]> wrote: also sprach Martin Marcher <[EMAIL PROTECTED]> [2007.05.03.1217 +0200]: > So what are the hints wether to use stable the actual name or not? From my book: Exactly my opinion too, i was more interested in hearing why I would wan't stable instead of the hardcoded name. I just can't think of any reason to do that and practically have really use for it (except for the testing or unstable distribution as I pointed out initially) Am I solving problems here that wouldn't exist in the first place? Am I ignorant of why there is a "stable" tree (apart from the psychological impact that it now indeed is stable)? martin -- Martin Marcher [EMAIL PROTECTED] http://www.mycorners.com https://www.xing.com/profile/Martin_Marcher http://www.linkedin.com/in/martinmarcher http://www.studivz.net/profile.php?ids=9f83ea8c5996b8ec http://www.amazon.de/gp/registry/wishlist/3KDAGCL2NKOIM/ref=reg_hu-wl_goto-registry/302-4432803-5146435?ie=UTF8&sort=date-added -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Dangers of "stable" in sources.list
also sprach Martin Marcher <[EMAIL PROTECTED]> [2007.05.03.1217 +0200]:
> So what are the hints wether to use stable the actual name or not?
From my book:
… and the \release{stable} and \release{testing} symlinks changed
to point to the next release generation. For this reason, it is
advisable to hardcode the release codename in
\file{/etc/apt/sources.list}, rather than its canonical name.
Specifically, for a \codename{sarge} system, I recommend changing
all occurrences of ``stable'' with ``sarge.'' While Debian release
is unlikely to catch you off-guard, using the code names for the
\programme{\ac{APT}} archive allows an upgrade to the next
official release on your own schedule, and not when the symlinks
in the archive change. When the next release follows, all you need
to do is replace ``sarge'' with ``etch'' and then
\command{dist-upgrade} as usual (see \cref{apt-upgrades}).
--
Please do not send copies of list mail to me; I read the list!
.''`. martin f. krafft <[EMAIL PROTECTED]>
: :' : proud Debian developer, author, administrator, and user
`. `'` http://people.debian.org/~madduck - http://debiansystem.info
`- Debian - when you have better things to do than fixing systems
NP: Thomas Truax - Swappin' Spit
signature.asc
Description: Digital signature (GPG/PGP)
