Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 02:51:10PM -0800, Don Armstrong wrote: On Mon, 03 Mar 2008, Colin Watson wrote: Yes, I know - I meant conceptually rather than in detail. notfixed should not leave the bug in a fixed/done-type state. The problem is that giving it side effects (reopening the bug) means that it's no longer associative. (commutative is the word you're looking for -- associative is (a+b)+c = a+(b+c)) That seems to violate (at least for me) the principle of least surprise. Having reopen NNN, found NNN VER and notfixed NNN VER be different probably already violates the principle of least surprise. Cheers, aj signature.asc Description: Digital signature
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 01:55:29PM -0800, Don Armstrong wrote: On Mon, 03 Mar 2008, Colin Watson wrote: On Mon, Mar 03, 2008 at 10:01:00PM +0100, Steinar H. Gunderson wrote: On Mon, Mar 03, 2008 at 08:57:25PM +, Colin Watson wrote: Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. I guess what he was trying to say, was that a versioned close + notfixed is the same as an unversioned close. (notfixed removes the version, but does not change the status of the bug.) Ah. For the record I am inclined to think that is at best unintuitive and probably a bug. We need a command that does what notfixed does. Well, ultimately, the problem's that done doesn't really mean all that much anymore -- we've got: real bugs that applied in version between X and Y real bugs that applied in versions = X real bugs that can't be associated with a version because they apply to pseudopackages, or to packages not in the archive things that should never have been reported as bugs We could introduce a pseudoversion for pseudopackages, eg make it the day in UTC formatted as .MM.DD -- then we can require a Version to always be present, and deal with things as: filed/closed bug against a real package, versions should be as expected filed against a pseudopackage -- version should be the date the bug first appeared closed bug against a pseudopackage -- version should be the date the bug was fixed version of a pseudopackage in stable/testing/unstable == today's date; pseudochangelog for pseudopackages that gives it a daily history closing a bug that should never have existed: just mark it notfound in any version and not fixed in any version (or mark it fixedin every version that it's foundin) The done field could then either be dropped entirely, or left as a comment that's only displayed for bugs that're determined to be resolved purely by looking at versions. For a version-free view (ie without dist=unstable, etc), you might be able to determine open/resolved based on based on whether max(found-in) max(fixed-in). You'd need to have something, to avoid relying on the done field. Cheers, aj signature.asc Description: Digital signature
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Tue, 04 Mar 2008, Anthony Towns wrote: On Mon, Mar 03, 2008 at 02:51:10PM -0800, Don Armstrong wrote: On Mon, 03 Mar 2008, Colin Watson wrote: Yes, I know - I meant conceptually rather than in detail. notfixed should not leave the bug in a fixed/done-type state. The problem is that giving it side effects (reopening the bug) means that it's no longer associative. (commutative is the word you're looking for -- associative is (a+b)+c = a+(b+c)) Right. That seems to violate (at least for me) the principle of least surprise. Having reopen NNN, found NNN VER and notfixed NNN VER be different probably already violates the principle of least surprise. That may be, however, we need some command that does what notfixed does, and notfixed is the best name that I've been able to come up with. [If they all did the same thing, I'd deprecate all but one of them.] Don Armstrong -- Quite the contrary; they *love* collateral damage. If they can make you miserable enough, maybe you'll stop using email entirely. Once enough people do that, then there'll be no legitimate reason left for anyone to run an SMTP server, and the spam problem will be solved. -- Craig Dickson in [EMAIL PROTECTED] http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 01:55:29PM -0800, Don Armstrong wrote: On Mon, 03 Mar 2008, Colin Watson wrote: On Mon, Mar 03, 2008 at 10:01:00PM +0100, Steinar H. Gunderson wrote: I guess what he was trying to say, was that a versioned close + notfixed is the same as an unversioned close. (notfixed removes the version, but does not change the status of the bug.) Ah. For the record I am inclined to think that is at best unintuitive and probably a bug. We need a command that does what notfixed does. I'm not arguing that notfixed shouldn't do what it does now. I'm arguing that it should *also* mark the bug as open if it's currently marked as done and there are no fixed versions remaining - straightforward bookkeeping, IMO. -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Tue, 04 Mar 2008, Colin Watson wrote: On Mon, Mar 03, 2008 at 01:55:29PM -0800, Don Armstrong wrote: We need a command that does what notfixed does. I'm not arguing that notfixed shouldn't do what it does now. I'm arguing that it should *also* mark the bug as open if it's currently marked as done and there are no fixed versions remaining - straightforward bookkeeping, IMO. That's exactly the side effect that I don't want it to have. If I wanted it to have that side effect, I wouldn't have implemented it at all and instead pointed people at found. Don Armstrong -- Grimble left his mother in the food store and went to the launderette and watched the clothes go round. It was a bit like colour television only with less plot. -- Clement Freud _Grimble_ http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Tue, 04 Mar 2008, Don Armstrong wrote: On Tue, 04 Mar 2008, Colin Watson wrote: On Mon, Mar 03, 2008 at 01:55:29PM -0800, Don Armstrong wrote: We need a command that does what notfixed does. I'm not arguing that notfixed shouldn't do what it does now. I'm arguing that it should *also* mark the bug as open if it's currently marked as done and there are no fixed versions remaining - straightforward bookkeeping, IMO. That's exactly the side effect that I don't want it to have. If I wanted it to have that side effect, I wouldn't have implemented it at all and instead pointed people at found. Just so it's clear, the reason why it's important that it not reopen the bug is because it enables me to fix entries in the fixed version list without having to worry about whether or not my changes have caused the bug to be open.[1] At the end of the set of control commands, the bug is almost invariably going to have fixed entries, so should be considered done, but if notfixed reopens it when the fixed set is zero (or to be consistent with found, when the largest found version exceeds the fixed version) I've got to issue two extra nonsensical versioning commands to keep everything sane. Thinking more about it, it may be better if we change the notfound/notfixed to be removefound/removefixed which is literally what they do. [As notfound removing all versions likewise doesn't cause the bug to be marked done.] Don Armstrong 1: Usually I'm fixing bugs which have been resolved to the wrong source/version pair. -- Il semble que la perfection soit atteinte non quand il n'y a plus rien a ajouter, mais quand il n'y a plus rien a retrancher. (Perfection is apparently not achieved when nothing more can be added, but when nothing else can be removed.) -- Antoine de Saint-Exupe'ry, Terres des Hommes http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 08:15:16PM +, Debian Bug Tracking System wrote: On Mon, 03 Mar 2008, Lionel Elie Mamane wrote: Bug #461617 is marked as resolved on http://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=wordpress;dist=unstable, but it is not solved in _any_ version, it was only transiently marked so, but that was cancelled with notfixed and found commands. The version graph on the bug page gets it correctly (it doesn't show any version as green), but the summary page for wordpress gets it wrong for some weird reason. Because you've marked it as done without specifying a version, the BTS assumes that it is fixed for all versions. 1) If this (marked done without specifying a version) were true, then we still have a bug: the version graph gets it wrong. 2) I can't find the message that marks it done without specifying a version; could you please point me to it? Thanks. I see only two messages to -close, both with a Source-Version pseudo-header, as automatically generated by dinstall, and no message to control@ that ever closes it / marks it as done. -- Lionel -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 08:57:25PM +, Colin Watson wrote: By using notfixed you've effectively sent an unversioned -done. Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. I guess what he was trying to say, was that a versioned close + notfixed is the same as an unversioned close. (notfixed removes the version, but does not change the status of the bug.) /* Steinar */ -- Homepage: http://www.sesse.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 12:47:16PM -0800, Don Armstrong wrote: On Mon, 03 Mar 2008, Lionel Elie Mamane wrote: 2) I can't find the message that marks it done without specifying a version; could you please point me to it? Thanks. I see only two messages to -close, both with a Source-Version pseudo-header, as automatically generated by dinstall, and no message to control@ that ever closes it / marks it as done. By using notfixed you've effectively sent an unversioned -done. Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, 03 Mar 2008, Lionel Elie Mamane wrote: On Mon, Mar 03, 2008 at 08:15:16PM +, Debian Bug Tracking System wrote: On Mon, 03 Mar 2008, Lionel Elie Mamane wrote: Bug #461617 is marked as resolved on http://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=wordpress;dist=unstable, but it is not solved in _any_ version, it was only transiently marked so, but that was cancelled with notfixed and found commands. The version graph on the bug page gets it correctly (it doesn't show any version as green), but the summary page for wordpress gets it wrong for some weird reason. Because you've marked it as done without specifying a version, the BTS assumes that it is fixed for all versions. 1) If this (marked done without specifying a version) were true, then we still have a bug: the version graph gets it wrong. The version graph only shows the known fixed/unfixed state for versions. In the cases where there is an unversioned done the bug is assumed to be resolved everywhere, but the version graph doesn't show this because a graph of all green would be rather pointless. 2) I can't find the message that marks it done without specifying a version; could you please point me to it? Thanks. I see only two messages to -close, both with a Source-Version pseudo-header, as automatically generated by dinstall, and no message to control@ that ever closes it / marks it as done. By using notfixed you've effectively sent an unversioned -done. If you wanted to say that it's still buggy in that version and cause it to be reopened as a side effect, use 'found' or just 'reopen'. See http://www.debian.org/Bugs/server-control#found vs http://www.debian.org/Bugs/server-control#notfixed Don Armstrong -- I may not have gone where I intended to go, but I think I have ended up where I needed to be. -- Douglas Adams _The Long Dark Tea-Time of the Soul_ http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 10:01:00PM +0100, Steinar H. Gunderson wrote: On Mon, Mar 03, 2008 at 08:57:25PM +, Colin Watson wrote: Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. I guess what he was trying to say, was that a versioned close + notfixed is the same as an unversioned close. (notfixed removes the version, but does not change the status of the bug.) Ah. For the record I am inclined to think that is at best unintuitive and probably a bug. -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, 03 Mar 2008, Colin Watson wrote: Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. No, notfixed should be the opposite of fixed, which it is. reopen (or found) is the opposite of -done. All notfixed does is removed a version from the fixed list. Its primary purpose is fixing typos and other mistakes in the fixed version list when that's *all* you want to do. If you're expecting other side effects you're either looking for found or reopen. notfound is its pair which performs precisely the same operation but on the found list. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. I added the notfixed and notfound commands myself, so I'm responsible for them. [I'm not entirely sure why they were being used in this case, as it seems clear from the bug log that what was meant was that the bug was actually still present in 2.3.3+fr-1, in which case 'found' or 'reopen' are the appropriate things.] Don Armstrong -- If you find it impossible to believe that the universe didn't have a creator, why don't you find it impossible that your creator didn't have one either? -- Anonymous Coward http://slashdot.org/comments.pl?sid=167556cid=13970629 http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, 03 Mar 2008, Colin Watson wrote: On Mon, Mar 03, 2008 at 10:01:00PM +0100, Steinar H. Gunderson wrote: On Mon, Mar 03, 2008 at 08:57:25PM +, Colin Watson wrote: Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. I guess what he was trying to say, was that a versioned close + notfixed is the same as an unversioned close. (notfixed removes the version, but does not change the status of the bug.) Ah. For the record I am inclined to think that is at best unintuitive and probably a bug. We need a command that does what notfixed does. I don't particularly care what we call it, but notfixed is the most symmetric thing that I could come up with, and it's in a namespace that wasn't already being used by any of the other commands. If we decide to rename it to something more clear, then notfound also needs to be renamed as well. Personally, unless there's some stunningly obvious name that I've missed, I'd rather just keep it the way it is. Don Armstrong -- PowerPoint is symptomatic of a certain type of bureaucratic environment: one typified by interminable presentations with lots of fussy little bullet-points and flashy dissolves and soundtracks masked into the background, to try to convince the audience that the goon behind the computer has something significant to say. -- Charles Stross _The Jennifer Morgue_ p33 http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, Mar 03, 2008 at 01:42:21PM -0800, Don Armstrong wrote: On Mon, 03 Mar 2008, Colin Watson wrote: Blink. Is that a typo? notfixed should be, if anything, the opposite of -done. No, notfixed should be the opposite of fixed, which it is. reopen (or found) is the opposite of -done. All notfixed does is removed a version from the fixed list. Yes, I know - I meant conceptually rather than in detail. notfixed should not leave the bug in a fixed/done-type state. If what you say is the case then I think we (or indeed probably I) buggered up the semantics rather badly. I added the notfixed and notfound commands myself, As it happens, I added notfound and you added fixed/notfixed. ;-) -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#469148: closed by Don Armstrong [EMAIL PROTECTED] (Re: Bug#469148: bugs.debian.org: bug marked as resolved when it is not in any version)
On Mon, 03 Mar 2008, Colin Watson wrote: Yes, I know - I meant conceptually rather than in detail. notfixed should not leave the bug in a fixed/done-type state. The problem is that giving it side effects (reopening the bug) means that it's no longer associative. That is, fixed NNN VER1 notfixed NNN VER would be (potentially) very different from notfixed NNN VER fixed NNN VER1 however, found NNN VER1 notfound NNN VER would still be the same as notfound NNN VER found NNN VER1 That seems to violate (at least for me) the principle of least surprise. Perhaps the right way is to make the documentation for notfixed slightly more clear than it already is and explain that you normally actually want found, not notfixed. [It is already the last of the version-modifying methods mentioned because it's only useful in specific cases.[1]] Any time that it was improperly marked as fixed in a normal situtation, you'd want to use found, because you know that the version you've just tested is buggy. As it happens, I added notfound and you added fixed/notfixed. ;-) Heh. I've touched that code so much now that it just seems like I wrote it originally. ;-) Don Armstrong 1: The main reason why I need a command like it is that it makes it really easy to fix source/version problems where the source package is wrong. I just notfixed version fixed version; without worying about whether it's going to screw with the -done-ness of the bug. -- The state must declare the child to be the most precious treasure of the people. As long as the government is perceived as working for the benefit of the children, the people will happily endure almost any curtailment of liberty and almost any deprivation. -- Adolf Hitler _Mein Kampf_ p403 http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]