Re: Allow comments and discussion even though an update was pushed to stable
Fabio Valentini writes: > On Sat, Dec 7, 2019, 17:30 Mattia Verga via devel < > devel@lists.fedoraproject.org> wrote: > >> Il 06/12/19 21:51, Johannes Lips ha scritto: >> >> "Johannes Lips" > >> >> >> >> >> We already have a tool for reporting issues and problems, and it's not >> >> bodhi. If there's a problem with an already pushed update, it needs to >> >> be in bugzilla - where it's actually discoverable - not in bodhi, where >> >> it will go nowhere. >> > I am not intending to use bodhi as a bug tracker, but I would like to be >> able to reference issues, which were introduced with an update and I don't >> really see a lot of negative effects of this. >> > I think it's a way to make issues in bugzilla more easily traceable, >> since if an update is closely related with a new bug it is easier to just >> look into bodhi first and see if the issue affected others as well. If by >> chance the first reporter of a bug is too late in bodhi this connection is >> simply lost. >> > I don't see it as bodhi replacing bugzilla, but rather as an additional >> entry point, when looking for known issues in close relation with a bug. >> > >> > Johannes >> > >> As a user, if I find a bug in a package I would look into Bugzilla. As a >> packager, I would expect to receive bug against my packages in Bugzilla. >> >> In my opinion, allowing comments on already pushed updates would mean >> that some users think that reporting bugs into Bodhi equals to reporting >> bugs into Bugzilla... so as a user I would have to look into Bodhi AND >> Bugzilla when I find regressions and as a packager I would have to deal >> with BZ tickets AND Bodhi comments without BZ tickets. >> >> If you **really** want to re-open comments on stable updates, please add >> a very big, red warning which states that the package maintainer may not >> look at new comments and to open a new BZ ticket! >> >> Mattia >> > > I'm curious about this. I do get email notifications for all changes > and comments for all my own bodhi updates plus for every update that I > commented on. Have you turned this off, or do you mean to say you're > going to ignore bodhi emails? It's really hard to justify reading bodhi emails at this point. The vast majority of bodhi emails that I receive are for automatic process for rawhide - which I neither care about nor wanted in any way. Thanks, --Robbie signature.asc Description: PGP signature ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Sun, 2019-12-08 at 08:51 +, Mattia Verga via devel wrote: > Il 07/12/19 19:32, Adam Williamson ha scritto: > > But we already explained this. Comments on stable updates are not > > primarily for the maintainer, they are for *users*. Users tend to refer > > to Bodhi notes, if anything, more than maintainers do. > > > ...and that's exactly the behavior that the change tries to avoid. From > the original RFE https://github.com/fedora-infra/bodhi/issues/2050 : > > > It appears that some users think that adding comments to an update > > which has been push to stable for weeks is a reasonable way to report > > an issue. And while I don't want to ignore them, I also don't have a > > good way to answer them besides adding further spam to a long-closed > > update. Still no. That's talking about the behaviour of *the person posting the comment*. That's not who I'm talking about. I'm talking about *people who run Fedora*, who notice an issue with their system after installing a particular update. A perfectly normal thing to do in that case is to go and look at the update in question in Bodhi. A comment added to that update will be seen by such a user. It is much harder for our poor user to go and look up a bug in Bugzilla, because it's a much more cumbersome system to use and there may well be several hundred other bug reports on the same package. Finding one caused by the specific update in question may well require finicky search work the user may not have the ability/inclination to carry out. > Closing comments on a stable update will force users to refer to BZ, so > that maintainers can focus on that channel. Instead, enabling users to > comment on closed updates it probably means (50%?) that the same user > will not open any bug in BZ, because we're all lazy, and "hey, I already > posted a comment, I don't want to bother to write the same things on BZ"... Again, not the 'user' I'm talking about. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
Il 07/12/19 19:32, Adam Williamson ha scritto: > But we already explained this. Comments on stable updates are not > primarily for the maintainer, they are for *users*. Users tend to refer > to Bodhi notes, if anything, more than maintainers do. > ...and that's exactly the behavior that the change tries to avoid. From the original RFE https://github.com/fedora-infra/bodhi/issues/2050 : > It appears that some users think that adding comments to an update > which has been push to stable for weeks is a reasonable way to report > an issue. And while I don't want to ignore them, I also don't have a > good way to answer them besides adding further spam to a long-closed > update. Closing comments on a stable update will force users to refer to BZ, so that maintainers can focus on that channel. Instead, enabling users to comment on closed updates it probably means (50%?) that the same user will not open any bug in BZ, because we're all lazy, and "hey, I already posted a comment, I don't want to bother to write the same things on BZ"... In https://github.com/fedora-infra/bodhi/issues/3748#issuecomment-562867117 I proposed to give users a way to flag a problematic update with a link to a (previously opened) BZ ticket, so that 1) the first user that notice a problem with the update will be forced to open a ticket in BZ and 2) following users will be noticed that the update has been reported to have some problems and can CC to the discussion on the BZ ticket. Mattia ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Sat, 2019-12-07 at 10:46 -0700, Jerry James wrote: > On Sat, Dec 7, 2019 at 9:50 AM Fabio Valentini wrote: > > > I'm curious about this. I do get email notifications for all changes and > > comments for all my own bodhi updates plus for every update that I > > commented on. Have you turned this off, or do you mean to say you're going > > to ignore bodhi emails? > > > > Speaking only for myself, yes I ignore bodhi emails. Why? I get huge > amounts of email, so I've developed some survival strategies. One of them > is to ignore sources of noise. Bodhi emails are mostly noise. Almost > always, when bodhi sends me email, the email tells me something I already > know. This part I agree with, actually - I don't monitor my Bodhi mail as closely as I should because of this issue. For example, take this update: https://bodhi.fedoraproject.org/updates/FEDORA-2019-81bcdfa3d5 Basically nothing of interest happened in its lifecycle. I created it, it got pushed to updates-testing, it reached the time threshold for being pushed stable, it was pushed stable. For that I get *five* emails notifying me of comments from bodhi itself: 1. "This update has been submitted for testing by adamwill." (yes, I know, that's *me*, I just *did* it) 2. "This update has been pushed to testing." (this is hardly news: it happens every time. Do I need an email to tell me about it?) 3. "This update can be pushed to stable now if the maintainer wishes" (maybe useful if I *don't* have autopush-for-time turned on, but is it any use if I *do*?) 4. "This update has been submitted for stable by bodhi." (obviously this arrived at exactly the same time as #3; there's really no need for both) 5. "This update has been pushed to stable." (again, this is hardly unexpected, do I really need yet another email to tell me?) Now, considering a fairly common workflow is to submit updates for multiple releases at a time - up to 6, if we have two stable releases, Branched, and you're submitted to three EPEL branches - this gets pretty silly: for 6 branches, I'm guaranteed to get *30* emails. Thus my Bodhi mail folder is a disaster area and I generally don't notice *interesting* things sent there, instead I spot them when I happen to go to the Bodhi web UI and look at updates there. Now, I could I guess do some fancy filtering for this client side; I could just try and filter out every mail that's only telling me about a comment from "bodhi" (although this isn't that easy because there is no convenient header to tell me this, it would have to be done by parsing the message content). But it might be nicer/more efficient to have some simple server-side choices here, like "don't send me mails for bodhi telling me about perfectly normal things happening", or something like that... > A build finished. Yes, I know. I saw the "fedpkg build" command > complete. Nit: Bodhi would not send you an email for this, because Bodhi has nothing to do with builds, only updates. At this point no update exists and there's nothing Bodhi could possibly email you about. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Sat, 2019-12-07 at 10:46 -0700, Jerry James wrote: > On Sat, Dec 7, 2019 at 9:50 AM Fabio Valentini wrote: > > > I'm curious about this. I do get email notifications for all changes and > > comments for all my own bodhi updates plus for every update that I > > commented on. Have you turned this off, or do you mean to say you're going > > to ignore bodhi emails? > > > > Speaking only for myself, yes I ignore bodhi emails. Why? I get huge > amounts of email, so I've developed some survival strategies. One of them > is to ignore sources of noise. Bodhi emails are mostly noise. Almost > always, when bodhi sends me email, the email tells me something I already > know. A build finished. Yes, I know. I saw the "fedpkg build" command > complete. I didn't need even one email about it, much less multiple > emails. If you want me to pay attention to bodhi emails, then the > signal-to-noise ratio has to improve dramatically. > > It would be enough to let users comment on an update for, let's say, a week > > after it's pushed to stable. That way, immediate feedback isn't lost, but > > bugs that are found after that time period need to go to bugzilla. > > > > I agree with Mattia on this. Once an update has gone stable, there is no > possibility of modifying it, so commenting on it further is not useful. > File a bug please. Then I can create a new update to fix the problem, and > a link to the update will be added to the bug. But we already explained this. Comments on stable updates are not primarily for the maintainer, they are for *users*. Users tend to refer to Bodhi notes, if anything, more than maintainers do. Multiple people have already said this is not an either/or question. It's not a case of *either* flagging an issue up via Bugzilla *or* doing it via Bodhi. It's both: you file a bug for the maintainer, and you add a note on Bodhi pointing to the bug for users who monitor Bodhi. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Sat, Dec 7, 2019 at 9:50 AM Fabio Valentini wrote: > I'm curious about this. I do get email notifications for all changes and > comments for all my own bodhi updates plus for every update that I > commented on. Have you turned this off, or do you mean to say you're going > to ignore bodhi emails? > Speaking only for myself, yes I ignore bodhi emails. Why? I get huge amounts of email, so I've developed some survival strategies. One of them is to ignore sources of noise. Bodhi emails are mostly noise. Almost always, when bodhi sends me email, the email tells me something I already know. A build finished. Yes, I know. I saw the "fedpkg build" command complete. I didn't need even one email about it, much less multiple emails. If you want me to pay attention to bodhi emails, then the signal-to-noise ratio has to improve dramatically. It would be enough to let users comment on an update for, let's say, a week > after it's pushed to stable. That way, immediate feedback isn't lost, but > bugs that are found after that time period need to go to bugzilla. > I agree with Mattia on this. Once an update has gone stable, there is no possibility of modifying it, so commenting on it further is not useful. File a bug please. Then I can create a new update to fix the problem, and a link to the update will be added to the bug. Here's the problem with allowing comments on stable updates: they're going to get lost in the shuffle. Those of us who maintain hundreds of packages need a fast, easy way to answer the question, "What outstanding problems do my packages have?" That way we can select what to work on next. I can go to bugzilla and see a list of open bugs. I can go to bodhi and see a list of updates in testing. Where, exactly, do I go to see a list of stable updates that have comments indicating more work needs to be done? Nowhere, which means I'm going to forget about those updates and the problems won't get fixed. Regards, -- Jerry James http://www.jamezone.org/ ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
Il 07/12/19 17:49, Fabio Valentini ha scritto: > I'm curious about this. I do get email notifications for all changes and > comments for all my own bodhi updates plus for every update that I commented > on. Have you turned this off, or do you mean to say you're going to ignore > bodhi emails? > > It would be enough to let users comment on an update for, let's say, a week > after it's pushed to stable. That way, immediate feedback isn't lost, but > bugs that are found after that time period need to go to bugzilla. > > Fabio > >> I usually ignore Bodhi emails and I have a filter which marks Bugzilla emails as high priority. I only look at Bodhi comments and karma when I have to manually push an update into stable. Mattia___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Sat, Dec 7, 2019, 17:30 Mattia Verga via devel < devel@lists.fedoraproject.org> wrote: > Il 06/12/19 21:51, Johannes Lips ha scritto: > >> "Johannes Lips" >> > >> > >> We already have a tool for reporting issues and problems, and it's not > >> bodhi. If there's a problem with an already pushed update, it needs to > >> be in bugzilla - where it's actually discoverable - not in bodhi, where > >> it will go nowhere. > > I am not intending to use bodhi as a bug tracker, but I would like to be > able to reference issues, which were introduced with an update and I don't > really see a lot of negative effects of this. > > I think it's a way to make issues in bugzilla more easily traceable, > since if an update is closely related with a new bug it is easier to just > look into bodhi first and see if the issue affected others as well. If by > chance the first reporter of a bug is too late in bodhi this connection is > simply lost. > > I don't see it as bodhi replacing bugzilla, but rather as an additional > entry point, when looking for known issues in close relation with a bug. > > > > Johannes > > > As a user, if I find a bug in a package I would look into Bugzilla. As a > packager, I would expect to receive bug against my packages in Bugzilla. > > In my opinion, allowing comments on already pushed updates would mean > that some users think that reporting bugs into Bodhi equals to reporting > bugs into Bugzilla... so as a user I would have to look into Bodhi AND > Bugzilla when I find regressions and as a packager I would have to deal > with BZ tickets AND Bodhi comments without BZ tickets. > > If you **really** want to re-open comments on stable updates, please add > a very big, red warning which states that the package maintainer may not > look at new comments and to open a new BZ ticket! > > Mattia > I'm curious about this. I do get email notifications for all changes and comments for all my own bodhi updates plus for every update that I commented on. Have you turned this off, or do you mean to say you're going to ignore bodhi emails? It would be enough to let users comment on an update for, let's say, a week after it's pushed to stable. That way, immediate feedback isn't lost, but bugs that are found after that time period need to go to bugzilla. Fabio > ___ > devel mailing list -- devel@lists.fedoraproject.org > To unsubscribe send an email to devel-le...@lists.fedoraproject.org > Fedora Code of Conduct: > https://docs.fedoraproject.org/en-US/project/code-of-conduct/ > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: > https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org > ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
Il 06/12/19 21:51, Johannes Lips ha scritto: >> "Johannes Lips" > >> >> We already have a tool for reporting issues and problems, and it's not >> bodhi. If there's a problem with an already pushed update, it needs to >> be in bugzilla - where it's actually discoverable - not in bodhi, where >> it will go nowhere. > I am not intending to use bodhi as a bug tracker, but I would like to be able > to reference issues, which were introduced with an update and I don't really > see a lot of negative effects of this. > I think it's a way to make issues in bugzilla more easily traceable, since if > an update is closely related with a new bug it is easier to just look into > bodhi first and see if the issue affected others as well. If by chance the > first reporter of a bug is too late in bodhi this connection is simply lost. > I don't see it as bodhi replacing bugzilla, but rather as an additional entry > point, when looking for known issues in close relation with a bug. > > Johannes > As a user, if I find a bug in a package I would look into Bugzilla. As a packager, I would expect to receive bug against my packages in Bugzilla. In my opinion, allowing comments on already pushed updates would mean that some users think that reporting bugs into Bodhi equals to reporting bugs into Bugzilla... so as a user I would have to look into Bodhi AND Bugzilla when I find regressions and as a packager I would have to deal with BZ tickets AND Bodhi comments without BZ tickets. If you **really** want to re-open comments on stable updates, please add a very big, red warning which states that the package maintainer may not look at new comments and to open a new BZ ticket! Mattia ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
Fabio Valentini wrote: > Right. So should the default of -3/+3 be changed for "critpath" packages? > They already need 14 days in testing, but they also get orders of > magnitude more feedback, so raising the karma limits to -2/+6 or something > like that sounds reasonable to me. > > On the other hand, I wouldn't even have any objections to changing the > defaults to something like -2/+6 for all packages, since it wouldn't make > any difference at all for the majority of packages that reach 7 days in > testing without any feedback whatsoever. IMHO, packages should just not be autopushed at all, no matter what threshold. An update can have +2 karma and be perfectly good to push, or it can have +10 karma and still have issues. It depends on what the users actually tested. So the maintainer should always read the freeform feedback text and decide based on that. (In fact, I would abolish the numeric karma entirely, provide only the freeform text field for feedback, and let the maintainer decide after reading it all. Humans will always make better decisions than naïve algorithms. But IMHO, even if the minimum karma requirement of +1 or 7 days wait time for non-critpath and +2 or 14 days wait time for critpath remains unchanged, it would still be an improvement to not allow automatic pushes of any kind.) Kevin Kofler ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Fri, 2019-12-06 at 20:51 +, Johannes Lips wrote: > > "Johannes Lips" > > > > > We already have a tool for reporting issues and problems, and it's not > > bodhi. If there's a problem with an already pushed update, it needs to > > be in bugzilla - where it's actually discoverable - not in bodhi, where > > it will go nowhere. > I am not intending to use bodhi as a bug tracker, but I would like to be able > to reference issues, which were introduced with an update and I don't really > see a lot of negative effects of this. > I think it's a way to make issues in bugzilla more easily traceable, since if > an update is closely related with a new bug it is easier to just look into > bodhi first and see if the issue affected others as well. If by chance the > first reporter of a bug is too late in bodhi this connection is simply lost. > I don't see it as bodhi replacing bugzilla, but rather as an additional entry > point, when looking for known issues in close relation with a bug. FWIW, there is in fact an open Bodhi ticket for this: https://github.com/fedora-infra/bodhi/issues/3748 which references the earlier ticket that requested comments be disallowed on stable updates: https://github.com/fedora-infra/bodhi/issues/2050 -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
> "Johannes Lips" > > We already have a tool for reporting issues and problems, and it's not > bodhi. If there's a problem with an already pushed update, it needs to > be in bugzilla - where it's actually discoverable - not in bodhi, where > it will go nowhere. I am not intending to use bodhi as a bug tracker, but I would like to be able to reference issues, which were introduced with an update and I don't really see a lot of negative effects of this. I think it's a way to make issues in bugzilla more easily traceable, since if an update is closely related with a new bug it is easier to just look into bodhi first and see if the issue affected others as well. If by chance the first reporter of a bug is too late in bodhi this connection is simply lost. I don't see it as bodhi replacing bugzilla, but rather as an additional entry point, when looking for known issues in close relation with a bug. Johannes > > From the other side, as a maintainer, it is important to have as few > channels for reporting issues as possible. Keeping everything in one > place is important so nothing gets lost, and it all can be queried and > triaged appropriately. > > Please do not use bodhi as a bug reporting tool. > > Thanks, > --Robbie ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
"Johannes Lips" writes: > What I found weird is that you can't comment on an update, which is > already pushed to stable. A lot of users are only hit by a bug, once > it reaches stable and then you don't have any possibility to highlight > a bug report or an issue with this update. I would like to have the > possibility to add such an information to an update, which introduced > the issue. We already have a tool for reporting issues and problems, and it's not bodhi. If there's a problem with an already pushed update, it needs to be in bugzilla - where it's actually discoverable - not in bodhi, where it will go nowhere. >From the other side, as a maintainer, it is important to have as few channels for reporting issues as possible. Keeping everything in one place is important so nothing gets lost, and it all can be queried and triaged appropriately. Please do not use bodhi as a bug reporting tool. Thanks, --Robbie signature.asc Description: PGP signature ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Fri, 2019-12-06 at 06:34 +, Johannes Lips wrote: > Hi all, > > I was recently bit by a bug, which was caused by a mismatch between > texlive-biblatex and biber. The technical side is not so important, > only so much that they need each other in a pretty specific version, > which is not reflected on the rpm level. > What I found weird is that you can't comment on an update, which is > already pushed to stable. A lot of users are only hit by a bug, once > it reaches stable and then you don't have any possibility to > highlight a bug report or an issue with this update. I would like to > have the possibility to add such an information to an update, which > introduced the issue. Yeah, I noticed this too. I think it's new, and it doesn't make a lot of sense to me either, it's an arbitrary restriction. I can only assume it's intended to avoid spam comments on long-dead updates or something, but there certainly are legit reasons to comment on a pushed update. This should probably be filed as an issue against upstream Bodhi, though. > Also I would like to ask if it is possible for important updates, > like the texlive one to increase the stable karma. It really depends > which mirrors you are using and if you are unlucky the updates get > pushed to stable, before it reaches updates-testing for you and then > again there's nothing to add, once it's pushed. This is set by the person who creates the update. It defaults to 3, but the submitter can set it to any value (well, not less than 2 for a critpath update I think). Any provenpackager can edit it too. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Fedora mirror selection (was: Allow comments and discussion even though an update was pushed) to stable
On Friday, 06 December 2019 at 10:57, Petr Pisar wrote: [...] > Maybe DNF could support setting a prefered mirror while still checking > for the latest metadata because in my experience the automatic mirror > selection does not always provide the best performance. (E.g. when > I connected an IPv6 only host in Germany, it resorted to a USA mirror.) I got hit by that as well when I deployed IPv6. GeoIP/geolite2 simply has wrong information about where your IP is located. Regards, Dominik -- Fedora https://getfedora.org | RPM Fusion http://rpmfusion.org There should be a science of discontent. People need hard times and oppression to develop psychic muscles. -- from "Collected Sayings of Muad'Dib" by the Princess Irulan ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On 2019-12-06, Johannes Lips wrote: > It really depends which mirrors you are using and if you are unlucky > the updates get pushed to stable, before it reaches updates-testing > for you and then again there's nothing to add, once it's pushed. > If you use metalink in your repository configuration (a default configuration), DNF always asks Fedora servers for the metadata. Thus DNF always gets the latest metadata. Then when downloading packages, outdated mirros are automatically excluded. If you manually changed the configuration to baseurl, then you deliberately bypass Fedora infrastructure and you are at the mercy of the mirror administrators. Maybe DNF could support setting a prefered mirror while still checking for the latest metadata because in my experience the automatic mirror selection does not always provide the best performance. (E.g. when I connected an IPv6 only host in Germany, it resorted to a USA mirror.) -- Petr ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Fri, Dec 06, 2019 at 07:51:40AM +0100, Fabio Valentini wrote: >On Fri, Dec 6, 2019, 07:35 Johannes Lips <[1]johannes.l...@gmail.com> >wrote: > > Hi all, > > I was recently bit by a bug, which was caused by a mismatch between > texlive-biblatex and biber. The technical side is not so important, only > so much that they need each other in a pretty specific version, which is > not reflected on the rpm level. > What I found weird is that you can't comment on an update, which is > already pushed to stable. A lot of users are only hit by a bug, once it > reaches stable and then you don't have any possibility to highlight a > bug report or an issue with this update. > >I agree, this is definitely a regression with newer bodhi versions (5.0+ I >think). I also often get hit by bugs in updates that are either queued for >stable, or already pushed. The best place to discuss this is likely the upstream issue tracker and more precisely, likely this ticket: https://github.com/fedora-infra/bodhi/issues/3748 Pierre ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Fri, Dec 6, 2019, 08:46 Mattia Verga via devel < devel@lists.fedoraproject.org> wrote: > Il 06/12/19 07:34, Johannes Lips ha scritto: > > Hi all, > > > > I was recently bit by a bug, which was caused by a mismatch between > texlive-biblatex and biber. The technical side is not so important, only so > much that they need each other in a pretty specific version, which is not > reflected on the rpm level. > > What I found weird is that you can't comment on an update, which is > already pushed to stable. A lot of users are only hit by a bug, once it > reaches stable and then you don't have any possibility to highlight a bug > report or an issue with this update. I would like to have the possibility > to add such an information to an update, which introduced the issue. > That was requested and discussed long time ago in > https://github.com/fedora-infra/bodhi/issues/2050 > > The rationale behind this was about the nonsense to have users > commenting on an already pushed update, since this cannot be undone. > > My opinion is that once an update is pushed to stable, new bugs should > be reported in Bugzilla, not as comments in Bodhi. However there's an > open discussion about restoring the previous behavior: > https://github.com/fedora-infra/bodhi/issues/3748 > > > Also I would like to ask if it is possible for important updates, like > the texlive one to increase the stable karma. It really depends which > mirrors you are using and if you are unlucky the updates get pushed to > stable, before it reaches updates-testing for you and then again there's > nothing to add, once it's pushed. > > > The stable-karma and stable-days parameters can be set by the user in > the web form or by CLI. By default stable-karma is set to 3, but it can > be changed when creating the update. > Right. So should the default of -3/+3 be changed for "critpath" packages? They already need 14 days in testing, but they also get orders of magnitude more feedback, so raising the karma limits to -2/+6 or something like that sounds reasonable to me. On the other hand, I wouldn't even have any objections to changing the defaults to something like -2/+6 for all packages, since it wouldn't make any difference at all for the majority of packages that reach 7 days in testing without any feedback whatsoever. Fabio > Mattia > > > ___ > devel mailing list -- devel@lists.fedoraproject.org > To unsubscribe send an email to devel-le...@lists.fedoraproject.org > Fedora Code of Conduct: > https://docs.fedoraproject.org/en-US/project/code-of-conduct/ > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: > https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org > ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
Il 06/12/19 07:34, Johannes Lips ha scritto: > Hi all, > > I was recently bit by a bug, which was caused by a mismatch between > texlive-biblatex and biber. The technical side is not so important, only so > much that they need each other in a pretty specific version, which is not > reflected on the rpm level. > What I found weird is that you can't comment on an update, which is already > pushed to stable. A lot of users are only hit by a bug, once it reaches > stable and then you don't have any possibility to highlight a bug report or > an issue with this update. I would like to have the possibility to add such > an information to an update, which introduced the issue. That was requested and discussed long time ago in https://github.com/fedora-infra/bodhi/issues/2050 The rationale behind this was about the nonsense to have users commenting on an already pushed update, since this cannot be undone. My opinion is that once an update is pushed to stable, new bugs should be reported in Bugzilla, not as comments in Bodhi. However there's an open discussion about restoring the previous behavior: https://github.com/fedora-infra/bodhi/issues/3748 > Also I would like to ask if it is possible for important updates, like the > texlive one to increase the stable karma. It really depends which mirrors you > are using and if you are unlucky the updates get pushed to stable, before it > reaches updates-testing for you and then again there's nothing to add, once > it's pushed. > The stable-karma and stable-days parameters can be set by the user in the web form or by CLI. By default stable-karma is set to 3, but it can be changed when creating the update. Mattia ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: Allow comments and discussion even though an update was pushed to stable
On Fri, Dec 6, 2019, 07:35 Johannes Lips wrote: > Hi all, > > I was recently bit by a bug, which was caused by a mismatch between > texlive-biblatex and biber. The technical side is not so important, only so > much that they need each other in a pretty specific version, which is not > reflected on the rpm level. > What I found weird is that you can't comment on an update, which is > already pushed to stable. A lot of users are only hit by a bug, once it > reaches stable and then you don't have any possibility to highlight a bug > report or an issue with this update. I agree, this is definitely a regression with newer bodhi versions (5.0+ I think). I also often get hit by bugs in updates that are either queued for stable, or already pushed. I would like to have the possibility to add such an information to an > update, which introduced the issue. > Also I would like to ask if it is possible for important updates, like the > texlive one to increase the stable karma. It really depends which mirrors > you are using and if you are unlucky the updates get pushed to stable, > before it reaches updates-testing for you and then again there's nothing to > add, once it's pushed. > What's even worse, some updates skip updates-testing altogether when enough people give positive feedback on the koji builds fast enough. Then there's literally *no* opportunity for "normal" users of updates-testing to provide feedback in bodhi. Fabio > Thanks > johannes > ___ > devel mailing list -- devel@lists.fedoraproject.org > To unsubscribe send an email to devel-le...@lists.fedoraproject.org > Fedora Code of Conduct: > https://docs.fedoraproject.org/en-US/project/code-of-conduct/ > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: > https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org > ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Allow comments and discussion even though an update was pushed to stable
Hi all, I was recently bit by a bug, which was caused by a mismatch between texlive-biblatex and biber. The technical side is not so important, only so much that they need each other in a pretty specific version, which is not reflected on the rpm level. What I found weird is that you can't comment on an update, which is already pushed to stable. A lot of users are only hit by a bug, once it reaches stable and then you don't have any possibility to highlight a bug report or an issue with this update. I would like to have the possibility to add such an information to an update, which introduced the issue. Also I would like to ask if it is possible for important updates, like the texlive one to increase the stable karma. It really depends which mirrors you are using and if you are unlucky the updates get pushed to stable, before it reaches updates-testing for you and then again there's nothing to add, once it's pushed. Thanks johannes ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org