Bug#509732: closed by Don Armstrong d...@debian.org (Re: Bug#509732: Kalle's message #68)
From: Don Armstrong d...@debian.org To: 509732-d...@bugs.debian.org Subject: Re: Bug#509732: Kalle's message #68 Date: Sat, 27 Dec 2008 17:53:34 -0800 User-Agent: Mutt/1.5.18 (2008-05-17) On Sat, 27 Dec 2008, José Luis González wrote: And can another developer file a serious bug if an error was found in the Manual? If so, please, let the description of Severe in the canonical location be updated. If such a bug was possible, it would fall afoul of makes unrelated software on the system (or the whole system) break. The bug is about the Manual, not the policy package. The debian-policy package wouldn't [make] unrelated software on the system (or the whole system) break, only the Manual. If the erroneous Manual was not yet in the package that severity wouldn't apply to the package. If it was, the severity would apply to the Document in the package. The former isn't falling afoul of, the latter could. My point was that if nobody able to set the severity to serious cared about it the error could remain and so the RC bugs it could have caused. Anyone can set the severity to an RC level; only certain people can make the final adjudication of severity. And according to the description of severities none of the RC levels are appliable to the hypothetical case described above. Yes, it is. I understand the description of serious only legitimates the mantainer to set this severity to the policy package when a bug about an error in the Manual that can lead to RC bugs is filed. I'm not following you at all here. It sounds like you're concerned about some kind of hypothetical bug in policy which could cause breakage in other packages. If such a bug were to happen, and it were filed against the debian-policy package, it will have its severity properly set It can't have it's severity set to an RC level if the current description of severities is applied, for the reasons stated before (please, see Message #80 and #111.) It is not whether it is possible per-se to set the severity, but that the current descriptions of severities do not make that hypothetical bug RC. Since they don't make it RC, whomever wishes to set the severity to RC could see somebody else set it back to non-RC and if there was a disagreement between both, it would be necessary that a Release Manager considered it RC and intervened since the description would make it non-RC. , as there are no less than 10 people who read this list who are also policy maintainers and/or RMs and/or responsible for the BTS. May I know if bugs under debian-policy are sent to the list? If they are not automatically it is still possible that it won't get into the list and won't have its severity properly set. Regardless, I'd much rather restrict myself to the non-hypothetical problems that we need to deal with in Debian, so I'll stop here. Closing this report with this message. Please, do not close bugs just because you don't want to deal with them. If the bug is still present it must remain open. If you can argument that the bug is not present anymore I will accept that it is closed. Since this hasn't happened yet I am going to reopen. If you are still uncomfortable with this bug getting too much attention please remember that its severity is minor. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#509732: closed by Don Armstrong d...@debian.org (Re: Bug#509732: Kalle's message #68)
On Sun, 28 Dec 2008, José Luis González wrote: The bug is about the Manual, not the policy package. The debian-policy package wouldn't [make] unrelated software on the system (or the whole system) break, only the Manual. If the erroneous Manual was not yet in the package that severity wouldn't apply to the package. If it was, the severity would apply to the Document in the package. The former isn't falling afoul of, the latter could. I have no clue what you're talking about here, then. This hypothetical case makes no sense. If it breaks other packages, it's an RC bug. If the bug is in debian-policy, but doesn't actually break other packages, then it's (probably) not an RC bug. People are watching -policy and can make the determination. May I know if bugs under debian-policy are sent to the list? If they are not automatically it is still possible that it won't get into the list and won't have its severity properly set. They are automatically sent. Please, do not close bugs just because you don't want to deal with them. I closed this bug with that message because: 1) The problem that I could understand was entirely hypothetical 2) The failure class that you presented is already covered Thus, the problem has already been dealt with. Don Armstrong -- Sentenced to two years hard labor (for sodomy), Oscar Wilde stood handcuffed in driving rain waiting for transport to prison. If this is the way Queen Victoria treats her prisoners, he remarked, she doesn't deserve to have any. http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#509732: Kalle's message #68
On Fri, 26 Dec 2008 20:11:03 -0800 Don Armstrong d...@debian.org wrote: On Fri, 26 Dec 2008, José Luis González wrote: If you still don't understand, imagine that a text is mistakenly introduced in the Policy and it causes a RC bug. How should this bug about the Policy be reported? According to the Policy Manual: It should be filed against debian-policy with the appropriate severity. What is the appropriate severity? The current description in bug-maint-info.txt and the lack of a directive in the Policy Manual that would allow filing it as serious would make it important, but the bug is causing a RC bug. If the error is reported to the list it can remain and the package with the erroneous policy be released if it is included in the package and the Policy isn't amended before. That assumes that the people reading the list won't file the appropriate bug on the appropriate package. I never known that not to be the case. Can another developer file a serious bug on the debian-policy package if the mantainer doesn't? According to bug-maint-info.txt a severe bug can be filed when it violates the Policy or the *mantainer* considers the *package* unsuitable for release. If another developer can, please, update the description. If the error is reported to the Bug Tracking System we are in a similar case as with the mailing list. Only if a mantainer raises severity of the debian-policy *package* to serious would the bug be considered as RC. If the bug is actually RC, someone has to recognize it as such and raise the severity. This isn't a serious problem for packages such as -policy which are watched by loads of people, Again, the same problem. If the mantainer is the only person who can consider the package unsuitable for release then only him can raise the severity to RC. and if a package isn't watched by people who know, then the bug probably isn't going to seriously affect the release anyway. This goes against the social contract: 4. Our priorities are our users and free software We will be guided by the needs of our users and the free software community. We will place their interests first in our priorities. If it is watched by a user he can't file a RC bug. The present report is only about Debian Policy bugs, but it could happen with any other Debian document that affects the release, i.e. the social contract or the constitution. I know those are foundational documents and can only be changed by developers, but if the bug is found by a user he couldn't file a RC bug. If nobody who knows watches doc-base it could affect the release. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#509732: Kalle's message #68
On Sat, 27 Dec 2008, José Luis González wrote: On Fri, 26 Dec 2008 20:11:03 -0800 Don Armstrong d...@debian.org wrote: It should be filed against debian-policy with the appropriate severity. What is the appropriate severity? Depends on the bug. I can't think of a non-packaging mistake in debian-policy that would by itself be a RC bug, but I suppose such a pathological case could be invented. That assumes that the people reading the list won't file the appropriate bug on the appropriate package. I never known that not to be the case. Can another developer file a serious bug on the debian-policy package if the mantainer doesn't? Of course. According to bug-maint-info.txt a severe bug can be filed when it violates the Policy or the *mantainer* considers the *package* unsuitable for release. Anyone can file an RC bug. This sentence is talking about the fact that a maintainer can decide that a package is unsuitable for release in addition to all of the other things that can make a package unsuitable for release. I should note too that the canonical location for this documentation is http://bugs.debian.org, *not* doc-debian. [doc-debian is a convenience copy.] and if a package isn't watched by people who know, then the bug probably isn't going to seriously affect the release anyway. This goes against the social contract: [snip SC §4] If it is watched by a user he can't file a RC bug. Users *can* and *do* file RC bugs.[1] My point was that if a bug was filed at the wrong severity, and no one noticed, then presumably the package isn't popular enough for enough people to care about it to set the severities appropriately. Here, the fundamental problem appears to be a misunderstanding of who can alter severities (anyone) versus who makes the final decision as to what the severities shall be (maintainers + RMs). Don Armstrong 1: As an aside, any time one might think that SC §4 is being violated, before trotting it out on the list, think long and hard about whether it actually is being violated. Odds are it's not. -- 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://rzlab.ucr.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#509732: Kalle's message #68
On Sat, 27 Dec 2008 02:57:25 -0800 Don Armstrong d...@debian.org wrote: On Sat, 27 Dec 2008, José Luis González wrote: On Fri, 26 Dec 2008 20:11:03 -0800 Don Armstrong d...@debian.org wrote: It should be filed against debian-policy with the appropriate severity. What is the appropriate severity? Depends on the bug. I can't think of a non-packaging mistake in debian-policy that would by itself be a RC bug, but I suppose such a pathological case could be invented. An error in the Manual can lead to an RC bug. That assumes that the people reading the list won't file the appropriate bug on the appropriate package. I never known that not to be the case. Can another developer file a serious bug on the debian-policy package if the mantainer doesn't? Of course. And can another developer file a serious bug if an error was found in the Manual? If so, please, let the description of Severe in the canonical location be updated. According to bug-maint-info.txt a severe bug can be filed when it violates the Policy or the *mantainer* considers the *package* unsuitable for release. Anyone can file an RC bug. This sentence is talking about the fact that a maintainer can decide that a package is unsuitable for release in addition to all of the other things that can make a package unsuitable for release. I should note too that the canonical location for this documentation is http://bugs.debian.org, *not* doc-debian. [doc-debian is a convenience copy.] serious is a severe violation of Debian policy (roughly, it violates a must or required directive), or, in the package maintainer's or release manager's opinion, makes the package unsuitable for release. If an error was found in the Manual none of the other RC severities could be applied, and serious would only be appliable if in the package mantainer's or release manager's opinion, makes the package unsuitable for release. So not anyone can file a RC bug if an error was found in the Policy Manual that would lead to RC bugs. If the description is wrong and, indeed, anyone can file it, then I would appreciate that the description was updated. and if a package isn't watched by people who know, then the bug probably isn't going to seriously affect the release anyway. This goes against the social contract: [snip SC §4] If it is watched by a user he can't file a RC bug. Users *can* and *do* file RC bugs.[1] My point was that if a bug was filed at the wrong severity, and no one noticed, then presumably the package isn't popular enough for enough people to care about it to set the severities appropriately. My point was that if nobody able to set the severity to serious cared about it the error could remain and so the RC bugs it could have caused. Here, the fundamental problem appears to be a misunderstanding of who can alter severities (anyone) versus who makes the final decision as to what the severities shall be (maintainers + RMs). Yes, it is. I understand the description of serious only legitimates the mantainer to set this severity to the policy package when a bug about an error in the Manual that can lead to RC bugs is filed. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#509732: Kalle's message #68
On Fri, 26 Dec 2008 12:05:15 +0200 Kalle wrote: I think what José is reporting is that in his opinion the BTS (and actually all pseudopackages available in the BTS) should be considered a part of the release and there should be Policy instructions as to how the BTS should work Only with regard to the Policy. From the note in the retitling message: Please see the severities in bug-maint-info.txt. From bug-maint-info.txt (unrelated to #227941): BEGIN QUOTE: Severity levels The bug system records a severity level with each bug report. This is set to normal by default, but can be overridden either by supplying a Severity line in the pseudo-header when the bug is submitted (see the instructions for reporting bugs), or by using the severity command with the control request server. The severity levels are: critical makes unrelated software on the system (or the whole system) break, or causes serious data loss, or introduces a security hole on systems where you install the package. grave makes the package in question unusable or mostly so, or causes data loss, or introduces a security hole allowing access to the accounts of users who use the package. serious is a severe violation of Debian policy (roughly, it violates a must or required directive), or, in the package maintainer's opinion, makes the package unsuitable for release. important a bug which has a major effect on the usability of a package, without rendering it completely unusable to everyone. normal the default value, applicable to most bugs. minor a problem which doesn't affect the package's usefulness, and is presumably trivial to fix. wishlist for any feature request, and also for any bugs that are very difficult to fix due to major design considerations. Certain severities are considered release-critical, meaning the bug will have an impact on releasing the package with the stable release of Debian. Currently, these are critical, grave and serious. For complete and canonical rules on what issues merit these severities, see the list of Release-Critical Issues for Etch. END QUOTE. Since the Policy isn't a package, the critical and grave severities don't apply to a bug caused by the policy that would be Release Critical. Since a bug is serious when it «is a severe violation of Debian policy (roughly, it violates a must or required directive), or, in the package maintainer's opinion, makes the package unsuitable for release» it is impossible to file a serious bug about it (the bug caused by the Policy which is Release Critical) since there's nothing in the Policy that mandates what is required to happen when there is a RC bug caused by the Policy. The only solution would be to mail the mantainer of the debian-policy package asking that he raises the severity of the bug against the debian-policy *package* to serious since the Policy is causing a release-critical bug so the *package* is unsuitable for release. If you still don't understand, imagine that a text is mistakenly introduced in the Policy and it causes a RC bug. How should this bug about the Policy be reported? According to the Policy Manual: If you discover an error in this manual or if you want to give any comments, suggestions, or criticisms please send an email to the Debian Policy List, debian-pol...@lists.debian.org, or submit a bug report against the debian-policy package. If the error is reported to the list it can remain and the package with the erroneous policy be released if it is included in the package and the Policy isn't amended before. If it is amended before Debian is released and the package includes the unamended text a bug with RC severity could only be filed by the mantainer (if an NMU can fix it then the description in bug-maint-info.txt should be updated to reflect this.) If the error is reported to the Bug Tracking System we are in a similar case as with the mailing list. Only if a mantainer raises severity of the debian-policy *package* to serious would the bug be considered as RC. If the mantainer doesn't and no other developer does the error in the Policy that is causing a RC bug would not be considered RC. If the erroneous Policy Manual gets into the package the bug against the package wouldn't be considered RC until a mantainer raises severity to serious. If no mantainer raises severity to serious the package could be released with the erroneous Policy Manual (that is causing the RC bug) in the next Debian release. The only solution would be to include a directive in the Policy requiring that errors in the manual that cause RC bugs are required not to get into the debian-policy package. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact
Bug#509732: Kalle's message #68
On Fri, 26 Dec 2008, José Luis González wrote: If you still don't understand, imagine that a text is mistakenly introduced in the Policy and it causes a RC bug. How should this bug about the Policy be reported? According to the Policy Manual: It should be filed against debian-policy with the appropriate severity. If the error is reported to the list it can remain and the package with the erroneous policy be released if it is included in the package and the Policy isn't amended before. That assumes that the people reading the list won't file the appropriate bug on the appropriate package. I never known that not to be the case. If the error is reported to the Bug Tracking System we are in a similar case as with the mailing list. Only if a mantainer raises severity of the debian-policy *package* to serious would the bug be considered as RC. If the bug is actually RC, someone has to recognize it as such and raise the severity. This isn't a serious problem for packages such as -policy which are watched by loads of people, and if a package isn't watched by people who know, then the bug probably isn't going to seriously affect the release anyway. Don Armstrong -- Show me your flowcharts and conceal your tables, and I shall continue to be mystified. Show me your tables, and I won't usually need your flowcharts; they'll be obvious. -- Fredrick P. Brooks Jr., The Mythical Man Month http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
