Re: Reporting bugs in Stable
On Mon 20 Apr 2020 at 00:11:08 (-0700), Ihor Antonov wrote: > On Sunday, 19 April 2020 23:30:43 PDT Andrei POPESCU wrote: > > On Du, 19 apr 20, 13:28:57, Ihor Antonov wrote: > > > Reporting from Debian Sid, everything is quite stable. I do run ZFS on > > > root > > > and make snapshots prior to big upgrades as a pre-caution, but so far > > > I did not have a reason to revert anything. > > > > It's just a matter of time. Even if Debian does much more automated > > testing now than in the past some serious issues could still slip > > through. > > I know, for me this is exactly the point: unstable becomes stable only if > someone uses it and finds out issues, reports/fixes them. > > > > I was using Archlinux for a long time, and I can say that Sid feels > > > more stable than Archlinux, although software is less fresh. But > > > overall quite usable as a daily driver on my Lenovo X1 Extreme > > > > As far as I know Archlinux is also not a beginners distro (like Mint or > > Ubuntu), so issues that may appear trivial to you can be major > > showstoppers for others. > > Absolutely, no disputing that. > I was trying to make a point that "unstable", despite scary name is quite > usable. Also as someone mentioned - backports should be the first option to > try > if you run stable. I run a few servers stable + backports and everything is > rock-solid. > > But I am afraid that we have deviated from the original topic. If I > understood Carl correctly - he was expressing his pain because of > bureaucratic scrutiny of filing bugs to stable that brings absolutely no > results. "Told not to bother" and "bureaucratic scrutiny" don't exactly explain the problem. > I can't help much here as I am just a mere user, but IMHO if software > in stable does not work - it is a severe bug. A bit of an assumption here. There are such things as normal and minor bugs, and also the fact that many users may never happen upon the circumstances that trigger a bug's effect. That's one of the reasons they need reporting. > It has to be either fixed or > software should be removed from stable. Steady on: we want some software to remain available. Even some serious and important bugs may have no effect on users' workflow, or be easily worked around if people are able to find out that they exist (by being reported). Cheers, David.
Re: Reporting bugs in Stable
On Sunday, 19 April 2020 23:30:43 PDT Andrei POPESCU wrote: > On Du, 19 apr 20, 13:28:57, Ihor Antonov wrote: > > Reporting from Debian Sid, everything is quite stable. I do run ZFS on > > root > > and make snapshots prior to big upgrades as a pre-caution, but so far > > I did not have a reason to revert anything. > > It's just a matter of time. Even if Debian does much more automated > testing now than in the past some serious issues could still slip > through. I know, for me this is exactly the point: unstable becomes stable only if someone uses it and finds out issues, reports/fixes them. > > I was using Archlinux for a long time, and I can say that Sid feels > > more stable than Archlinux, although software is less fresh. But > > overall quite usable as a daily driver on my Lenovo X1 Extreme > > As far as I know Archlinux is also not a beginners distro (like Mint or > Ubuntu), so issues that may appear trivial to you can be major > showstoppers for others. Absolutely, no disputing that. I was trying to make a point that "unstable", despite scary name is quite usable. Also as someone mentioned - backports should be the first option to try if you run stable. I run a few servers stable + backports and everything is rock-solid. But I am afraid that we have deviated from the original topic. If I understood Carl correctly - he was expressing his pain because of bureaucratic scrutiny of filing bugs to stable that brings absolutely no results. I can't help much here as I am just a mere user, but IMHO if software in stable does not work - it is a severe bug. It has to be either fixed or software should be removed from stable. Thanks Ihor Antonov
Re: Reporting bugs in Stable
On Du, 19 apr 20, 13:28:57, Ihor Antonov wrote: > > Reporting from Debian Sid, everything is quite stable. I do run ZFS on root > and make snapshots prior to big upgrades as a pre-caution, but so far > I did not have a reason to revert anything. It's just a matter of time. Even if Debian does much more automated testing now than in the past some serious issues could still slip through. > I was using Archlinux for a long time, and I can say that Sid feels > more stable than Archlinux, although software is less fresh. But > overall quite usable as a daily driver on my Lenovo X1 Extreme As far as I know Archlinux is also not a beginners distro (like Mint or Ubuntu), so issues that may appear trivial to you can be major showstoppers for others. Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: Reporting bugs in Stable
On 20/4/20 12:26 am, Andrei POPESCU wrote: On Du, 19 apr 20, 09:43:46, Carl Fink wrote: So this has bugged me every time I run Debian Stable: you find a bug. You try to report it, and are told not to bother because there's a newer version. By? I'm guessing you mean the standard request from reportbug to try a newer version. Why is reportbug even in Stable? Why not just replace it with a script that says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, that's literally the Debian policy, that only security fixes are done in Stable. Actually bugs of severity "important" or higher can be fixed in stable, provided certain criteria are met. So, actual question: how usable is the current Testing? It's usable, though with less guarantees. Have a backup plan in case of breakage. Mine is typically a paralel stable install, though other methods exist (e.g. snapshots) Kind regards, Andrei Agree. I have had a couple of issues, but reverting a few days with my snapshot has saved my a couple of times - not for several months now, though. I use timeshift, which is usable from terminal or a live .iso of a linux which includes timeshift as a default. Try Mint if you don't have a better option. You will get more downloads with package updates though. But that is the real difference between stable and testing - testing finds bits that still need updating. Even SID is a stable OS, just gets even more packages updated (from memory, packages stay in SID until about 2 weeks of no major issues?). -- Keith Bainbridge ke1th3...@zoho.com +61 (0)447 667 468
Re: Reporting bugs in Stable
On Sun 19 Apr 2020 at 09:43:46 (-0400), Carl Fink wrote: > So this has bugged me every time I run Debian Stable: you find a bug. You > try to report it, and are told not to bother because there's a newer > version. There is no way to install the newer version without manually > fiddling with pointlessly arcane configuration files that are sort of > documented > if you squint. > > (Yes, the pun on "bug" is deliberate.) > > Why is reportbug even in Stable? Why not just replace it with a script that > says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, that's > literally the Debian policy, that only security fixes are done in Stable. > > Yes, technically if the version number in Stable and Experimental are the > same, the bug might get fixed, but the fix would never actually be in Stable > until the current Testing is released. Aren't you assuming that a bug fix is the sole use for the Bug Tracking System. I find it's a help when I suspect that I might be seeing the effect of a bug. It can also be useful for finding workarounds, and for comparing competing packages. You can also get advanced warning of serious bugs by apt-listbugs before you install a package. > So, actual question: how usable is the current Testing? Because Stable is > ... not so much, and decreasing. (It's fine as a server OS, it's just as a > client box that it effectively degrades over time as software upgrades don't > happen.) I'm not sure I understand: isn't it your perspective that changes, as you discover unfixed bugs. The software itself stays the same. However, as someone obviously keen to report bugs, your using Testing could be valuable for the project. Cheers, David.
Re: Reporting bugs in Stable
On 4/19/20, Andrei POPESCU wrote: > On Du, 19 apr 20, 09:43:46, Carl Fink wrote: >> So this has bugged me every time I run Debian Stable: you find a bug. You >> try to report it, and are told not to bother because there's a newer >> version. > > By? I'm guessing you mean the standard request from reportbug to try a > newer version. > >> Why is reportbug even in Stable? Why not just replace it with a script >> that >> says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, >> that's >> literally the Debian policy, that only security fixes are done in Stable. > > Actually bugs of severity "important" or higher can be fixed in stable, > provided certain criteria are met. Just in case it helps: you also have Backports. «Backports are packages taken from the next Debian release (called "testing"), adjusted and recompiled for usage on Debian stable.»[1] >> So, actual question: how usable is the current Testing? > > It's usable, though with less guarantees. Have a backup plan in case of > breakage. Mine is typically a paralel stable install, though other > methods exist (e.g. snapshots) Another option - if disk-space is a concern - is to have a pendrive at hand with a couple of your preferred (updated) LiveUSB distros in something like MultiSystem[2] (which I *really* don't know why isn't available in Debian repositories...) or anything similar that you like[3]. Best regards. [1] https://backports.debian.org/ [2] http://liveusb.info/dotclear/ [3] https://alternativeto.net/software/multisystem/?license=opensource=linux
Re: Reporting bugs in Stable
On Sunday, 19 April 2020 07:26:31 PDT Andrei POPESCU wrote: > On Du, 19 apr 20, 09:43:46, Carl Fink wrote: > > So this has bugged me every time I run Debian Stable: you find a bug. You > > try to report it, and are told not to bother because there's a newer > > version. > > By? I'm guessing you mean the standard request from reportbug to try a > newer version. > > > Why is reportbug even in Stable? Why not just replace it with a script > > that > > says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, > > that's literally the Debian policy, that only security fixes are done in > > Stable. > Actually bugs of severity "important" or higher can be fixed in stable, > provided certain criteria are met. > > > So, actual question: how usable is the current Testing? Reporting from Debian Sid, everything is quite stable. I do run ZFS on root and make snapshots prior to big upgrades as a pre-caution, but so far I did not have a reason to revert anything. I was using Archlinux for a long time, and I can say that Sid feels more stable than Archlinux, although software is less fresh. But overall quite usable as a daily driver on my Lenovo X1 Extreme --- Ihor Antonov
Re: Reporting bugs in Stable
On Sun, 2020-04-19 at 10:27 -0400, Carl Fink wrote: > On Sun, Apr 19, 2020 at 09:51:02AM -0400, Jim Popovitch wrote: > > > What applications do you feel aren't up-to-date enough for your liking? > > I'm genuinely curious. > > Mr. Heskett's comments made me want to tell him how to lower the CPU usage > of BOINC. However, boinc-manager in Stable, at least on my system, has a bug > resulting in a blank Computing Preferences dialog. > > (Options >> Computing preferences) > > > > So, actual question: how usable is the current Testing? Because Stable is > > > ... not so much, and decreasing. (It's fine as a server OS, it's just as a > > > client box that it effectively degrades over time as software upgrades > > > don't > > > happen.) > > > > I run stable on a work laptop, it's quite stable (which is what I want > > out of it) > > This is, of course, not actually an answer to my question. It wasn't meant to be. It was a comment on how stable, for me, is certainly not degrading over time. Best wishes, -Jim P.
Re: Reporting bugs in Stable
On Sun, 19 Apr 2020 09:51:02 -0400 Jim Popovitch wrote: > On Sun, 2020-04-19 at 09:43 -0400, Carl Fink wrote: > > Why is reportbug even in Stable? Why not just replace it with a script that > > says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, that's > > literally the Debian policy, that only security fixes are done in Stable. > > I agree with your sentiments, but need to point out that some > applications are updated regularly in stable (Firefox-ESR is one that > comes to mind), and there are regular point-releases that contain > updates. To clarify: the official absolute assertion that: "Once a Debian version is released and tagged `stable' it will only get security updates. That is, only packages for which a security vulnerability has been found after the release will be upgraded." [1] is not quite accurate - point releases also fix "important bugs in the current release." [2] Additionally, it's not always clear what constitutes a security vulnerability. In any event, a quick browsing of (for example) the most recent (10.3) point release notice shows that numerous bugs that aren't necessarily security vulnerabilities have indeed been fixed. [3] [1] https://www.debian.org/doc/manuals/debian-faq/getting-debian.en.html#updatestable [2] https://wiki.debian.org/DebianReleases/PointReleases [3] https://www.debian.org/News/2020/20200208 Celejar
Re: Reporting bugs in Stable
On Sun, Apr 19, 2020 at 09:51:02AM -0400, Jim Popovitch wrote: > What applications do you feel aren't up-to-date enough for your liking? > I'm genuinely curious. Mr. Heskett's comments made me want to tell him how to lower the CPU usage of BOINC. However, boinc-manager in Stable, at least on my system, has a bug resulting in a blank Computing Preferences dialog. (Options >> Computing preferences) > > So, actual question: how usable is the current Testing? Because Stable is > > ... not so much, and decreasing. (It's fine as a server OS, it's just as a > > client box that it effectively degrades over time as software upgrades don't > > happen.) > > I run stable on a work laptop, it's quite stable (which is what I want > out of it) This is, of course, not actually an answer to my question. -- Carl Fink c...@finknetwork.com https://reasonablyliterate.com https://nitpicking.com If you want to make a point, somebody will take the point and stab you with it. -Kenne Estes
Re: Reporting bugs in Stable
On Du, 19 apr 20, 09:43:46, Carl Fink wrote: > So this has bugged me every time I run Debian Stable: you find a bug. You > try to report it, and are told not to bother because there's a newer > version. By? I'm guessing you mean the standard request from reportbug to try a newer version. > Why is reportbug even in Stable? Why not just replace it with a script that > says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, that's > literally the Debian policy, that only security fixes are done in Stable. Actually bugs of severity "important" or higher can be fixed in stable, provided certain criteria are met. > So, actual question: how usable is the current Testing? It's usable, though with less guarantees. Have a backup plan in case of breakage. Mine is typically a paralel stable install, though other methods exist (e.g. snapshots) Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: Reporting bugs in Stable
On Sun, 2020-04-19 at 09:43 -0400, Carl Fink wrote: > Why is reportbug even in Stable? Why not just replace it with a script that > says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, that's > literally the Debian policy, that only security fixes are done in Stable. I agree with your sentiments, but need to point out that some applications are updated regularly in stable (Firefox-ESR is one that comes to mind), and there are regular point-releases that contain updates. What applications do you feel aren't up-to-date enough for your liking? I'm genuinely curious. > So, actual question: how usable is the current Testing? Because Stable is > ... not so much, and decreasing. (It's fine as a server OS, it's just as a > client box that it effectively degrades over time as software upgrades don't > happen.) I run stable on a work laptop, it's quite stable (which is what I want out of it) -Jim P.
Reporting bugs in Stable
So this has bugged me every time I run Debian Stable: you find a bug. You try to report it, and are told not to bother because there's a newer version. There is no way to install the newer version without manually fiddling with pointlessly arcane configuration files that are sort of documented if you squint. (Yes, the pun on "bug" is deliberate.) Why is reportbug even in Stable? Why not just replace it with a script that says "Sorry, bugs in Stable are never fixed. Try Testing." Seriously, that's literally the Debian policy, that only security fixes are done in Stable. Yes, technically if the version number in Stable and Experimental are the same, the bug might get fixed, but the fix would never actually be in Stable until the current Testing is released. So, actual question: how usable is the current Testing? Because Stable is ... not so much, and decreasing. (It's fine as a server OS, it's just as a client box that it effectively degrades over time as software upgrades don't happen.) -- Carl Fink c...@finknetwork.com https://reasonablyliterate.com https://nitpicking.com If you want to make a point, somebody will take the point and stab you with it. -Kenne Estes