Re: New bodhi bugfix release in production
On Tue, Oct 25, 2011 at 06:26:13PM -0400, Luke Macken wrote: On Tue, Oct 25, 2011 at 02:59:51PM -0700, Adam Williamson wrote: On Tue, 2011-10-25 at 17:18 -0400, Luke Macken wrote: bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually *was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. I'd be happy to revert this if the majority of people prefer the other format. Bodhi will still use the n-v-r style URLs for the updates-testing digests, but will default to the static IDs otherwise. The biggest problem with using the builds in the URL is that the URLs break if they are edited to add/remove/update them. I guess we could add some additional logic to try and be clever and find the update even if one of the builds is missing or modified. If an update has the n-e-v-r changed, the n-e-v-r being replaced is not likely to ever be used in a different future update. So each update could maintain a list of all n-e-v-r's that have ever been associated with it. Then make bodhi support URLs for all n-e-v-r's associated with the update rather than only the latest. When sending URLs in email just use the latest n-e-v-r. Thus changing an existing update will never break any URL Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
Adam Williamson wrote: In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually *was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. And I've resorted to creating the ID-style URLs manually previously, and in fact had found the change from IDs to packagename-version URLs in some places (which has apparently been undone now) to be a serious regression. The ID is a permalink, the packagename-version URLs are not, which means the ID is a lot more useful. E.g. when Bodhi posts a link in Bugzilla using packagename-version format, that link stops working when the update is edited. Kevin Kofler -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On 26/10/11 12:45, Kevin Kofler wrote: Maybe we could do what some sites like kde-apps.org do and default to URLs which include BOTH the ID and the packagename-version list, but have Bodhi only actually use the ID and ignore the packagename-version entirely. See e.g.: http://kde-apps.org/content/show.php/Apper?content=84745 http://kde-apps.org/content/show.php/KPackageKit?content=84745 http://kde-apps.org/content/show.php/Pink_Pony?content=84745 http://kde-apps.org/content/show.php?content=84745 (The first link is what the site produces by default, the second is what it used to produce before the package got renamed, the third is to prove you can write in anything and it will be ignored and the fourth is the equivalent of Bodhi's current ID-based URLs, which also works.) But I'd put the ID first so it's easier to rip off the ignored stuff from the link if a short URL is needed. Many CMS systems and the like work in that way. It's also what things like stackoverflow do, for example: http://stackoverflow.com/questions/7901782/war-does-not-start-on-tomcat5-on-redhat-enterprise-server where only the question number really matters - you can change the text which follows to anything. Tom -- Tom Hughes (t...@compton.nu) http://compton.nu/ -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
Daniel P. Berrange wrote: If an update has the n-e-v-r changed, the n-e-v-r being replaced is not likely to ever be used in a different future update. So each update could maintain a list of all n-e-v-r's that have ever been associated with it. Then make bodhi support URLs for all n-e-v-r's associated with the update rather than only the latest. When sending URLs in email just use the latest n-e-v-r. Thus changing an existing update will never break any URL There have been cases where update groups have been split, i.e. builds removed from the group and filed separately, or added to another group. Kevin Kofler -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On 2011-10-25 15:17, Adam Williamson wrote: It's not just the updates-testing list, though. When I go to the web interface, search for updates to, say, grub2, get a list, and click on one of the results, I get an ID-based URL, not a package name-based one. I then paste that into an email, IRC conversation, or trac compose request ticket, and no-one can see what the update *is* unless they click on the link. And after a few hours that link may have false information and stop working altogether. It doesn't even have to wait until the next push happens. Multi-package updates are especially fragile, as a change in any constituent can break all existing links, invalidating browser histories and links in bugzilla and e-mail messages. They also lead to links of incredible length. Perhaps the permanence problem could be solved for the majority of cases if bodhi were to remember the last update with which each n-v-r was associated rather than only the n-v-rs that are currently associated with updates. If the change to links outside of mailing lists will also be reverted, then instances where length matters (e.g. IRC) could be improved by making update IDs in search results and individual update pages into ID-based links so people at least don't have to construct them on their own. Thoughts? -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Wed, 2011-10-26 at 13:45 +0200, Kevin Kofler wrote: Adam Williamson wrote: Really this is just a tooling question of whether it's overall more convenient to have Bodhi use IDs and then implement convenience scripts/tools wherever we refer to the updates which can identify them - like a bot in IRC, and a modification to the updates-testing email to make it use the old format, etc - or whether it's better to have Bodhi use NEVRs and then have to somehow deal with the problem of updates with dozens of packages, and the problem of updates which are edited to include different NEVRs. Hmmm, a suggestion: Maybe we could do what some sites like kde-apps.org do and default to URLs which include BOTH the ID and the packagename-version list, but have Bodhi only actually use the ID and ignore the packagename-version entirely. See e.g.: http://kde-apps.org/content/show.php/Apper?content=84745 http://kde-apps.org/content/show.php/KPackageKit?content=84745 http://kde-apps.org/content/show.php/Pink_Pony?content=84745 http://kde-apps.org/content/show.php?content=84745 (The first link is what the site produces by default, the second is what it used to produce before the package got renamed, the third is to prove you can write in anything and it will be ignored and the fourth is the equivalent of Bodhi's current ID-based URLs, which also works.) But I'd put the ID first so it's easier to rip off the ignored stuff from the link if a short URL is needed. That should bring us the best of both worlds (and people who are bothered by the redundant stuff could simply rip out everything after the ID from the URL, just as we're doing now for that CSRF junk anyway). That sure sounds good to me, if we hit on a format that's easily readable. nice idea. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
Or perhaps even: https://admin.fedoraproject.org/updates/FEDORA--N/package1-1.1.fc16,package2-1.1.fc16 where anything after the FEDORA--N doesn't matter, but could contain all the current packages in the update. kevin signature.asc Description: PGP signature -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Wed, Oct 26, 2011 at 03:04:12PM -0700, Garrett Holmstrom wrote: On Wed, Oct 26, 2011 at 12:07 PM, Kevin Fenzi ke...@scrye.com wrote: Or perhaps even: https://admin.fedoraproject.org/updates/FEDORA--N/package1-1.1.fc16,package2-1.1.fc16 where anything after the FEDORA--N doesn't matter, but could contain all the current packages in the update. This sounds reasonable to me. How feasible is teaching bodhi to parse that sort of URI that way? Very feasible :) https://fedorahosted.org/bodhi/changeset/86ec2fb28d15c2fc76866924a84f1380221948d6 luke -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Wed, Oct 26, 2011 at 07:17:10PM -0400, Luke Macken wrote: On Wed, Oct 26, 2011 at 03:04:12PM -0700, Garrett Holmstrom wrote: On Wed, Oct 26, 2011 at 12:07 PM, Kevin Fenzi ke...@scrye.com wrote: Or perhaps even: https://admin.fedoraproject.org/updates/FEDORA--N/package1-1.1.fc16,package2-1.1.fc16 where anything after the FEDORA--N doesn't matter, but could contain all the current packages in the update. This sounds reasonable to me. How feasible is teaching bodhi to parse that sort of URI that way? Very feasible :) https://fedorahosted.org/bodhi/changeset/86ec2fb28d15c2fc76866924a84f1380221948d6 Of course, I pulled the git-push trigger too early, and the above commit has a couple of issues, which have since been resolved. Kevin Fenzi's suggestion for using /updates/ID/builds as a default URL structure has been implemented. Since it only looks for the update by the ID the builds can change and it will still take you to the same update. The update IDs are assigned when they are first pushed to testing, so pending updates will still have the same /updates/builds URL that they always have. Thanks to everyone who contributed their ideas in this thread! luke -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On 10/26/2011 05:27 PM, Tom Hughes wrote: Many CMS systems and the like work in that way. It's also what things like stackoverflow do, for example: http://stackoverflow.com/questions/7901782/war-does-not-start-on-tomcat5-on-redhat-enterprise-server where only the question number really matters - you can change the text which follows to anything. Yep. So does Ask Fedora. http://ask.fedoraproject.org Rahul -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
New bodhi bugfix release in production
bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. https://admin.fedoraproject.org/updates I raced to get this out before the infrastructure freeze today, and since then there have already been many more bugfixes in git, so expect another release shortly after F16 is released. Please file bugs here: https://fedorahosted.org/bodhi/newticket Client fixes - bodhi -L dies with out-of-range exception after branching f16 https://fedorahosted.org/bodhi/ticket/625 https://bugzilla.redhat.com/show_bug.cgi?id=746780 - bodhi -r dist-f14 -b 676195 don't respect -r option https://bugzilla.redhat.com/show_bug.cgi?id=747939 Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 - fedora-easy-karma submits too many comments to bodhi when bodhi has a server problem (edit) https://bugzilla.redhat.com/show_bug.cgi?id=698441 - Bodhi no longer adds comments to Security Response bugs https://fedorahosted.org/bodhi/ticket/485 Buildroot override fixes - Buildroot overrides require commit access to devel branch rather than branch override applies to https://fedorahosted.org/bodhi/ticket/620 - Cannot request build root override https://bugzilla.redhat.com/show_bug.cgi?id=729722 - buildroot overrides stay after expiration date https://bugzilla.redhat.com/show_bug.cgi?id=723071 Masher fixes - Updates-testing report emails should use package names not update number https://fedorahosted.org/bodhi/ticket/644 - Current updateinfo data is broken (epoch=None) https://bugzilla.redhat.com/show_bug.cgi?id=652296 - Fedora Update System suggests to reboot when not asked to do so https://bugzilla.redhat.com/show_bug.cgi?id=681850 Package fixes - - bodhi-server should require python-fedora-turbogears https://bugzilla.redhat.com/show_bug.cgi?id=743975 pgpcXyy8mbom9.pgp Description: PGP signature ___ devel-announce mailing list devel-annou...@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel-announce-- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tue, 2011-10-25 at 17:18 -0400, Luke Macken wrote: bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually *was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tue, 25 Oct 2011 14:59:51 -0700 Adam Williamson awill...@redhat.com wrote: On Tue, 2011-10-25 at 17:18 -0400, Luke Macken wrote: bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually *was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. Please read down... Masher fixes - Updates-testing report emails should use package names not update number https://fedorahosted.org/bodhi/ticket/644 But see the test list... there's an issue with the new package name based links. ;) kevin signature.asc Description: PGP signature -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tue, 2011-10-25 at 16:10 -0600, Kevin Fenzi wrote: On Tue, 25 Oct 2011 14:59:51 -0700 Adam Williamson awill...@redhat.com wrote: On Tue, 2011-10-25 at 17:18 -0400, Luke Macken wrote: bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually *was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. Please read down... Masher fixes - Updates-testing report emails should use package names not update number https://fedorahosted.org/bodhi/ticket/644 But see the test list... there's an issue with the new package name based links. ;) It's not just the updates-testing list, though. When I go to the web interface, search for updates to, say, grub2, get a list, and click on one of the results, I get an ID-based URL, not a package name-based one. I then paste that into an email, IRC conversation, or trac compose request ticket, and no-one can see what the update *is* unless they click on the link. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tue, Oct 25, 2011 at 02:59:51PM -0700, Adam Williamson wrote: On Tue, 2011-10-25 at 17:18 -0400, Luke Macken wrote: bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually *was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. I'd be happy to revert this if the majority of people prefer the other format. Bodhi will still use the n-v-r style URLs for the updates-testing digests, but will default to the static IDs otherwise. The biggest problem with using the builds in the URL is that the URLs break if they are edited to add/remove/update them. I guess we could add some additional logic to try and be clever and find the update even if one of the builds is missing or modified. luke -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
Luke Macken wrote: In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually*was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. I'd be happy to revert this if the majority of people prefer the other format. Bodhi will still use the n-v-r style URLs for the updates-testing digests, but will default to the static IDs otherwise. The biggest problem with using the builds in the URL is that the URLs break if they are edited to add/remove/update them. I guess we could add some additional logic to try and be clever and find the update even if one of the builds is missing or modified. Think about how bugzilla bugs are handled in IRC. Bugs all have ID numbers. Why should updates be different? I vote for static IDs because I have run into the case of modified updates and broken URLs. Adam, can you not pursue an enhancement to the IRC bot that translates bug URLs into descriptions to also handle bodhi IDs? -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tuesday, October 25, 2011, 6:32:26 PM, Michael wrote: Luke Macken wrote: In case you hadn't noticed, response to this has so far been pretty negative. It seems people liked being able to tell from the URL what the update actually*was*. I must admit I do to. I've resorted to creating the 'old-style' URLs manually when I do lists of updates on test@ or in trac, now. I'd be happy to revert this if the majority of people prefer the other format. Bodhi will still use the n-v-r style URLs for the updates-testing digests, but will default to the static IDs otherwise. The biggest problem with using the builds in the URL is that the URLs break if they are edited to add/remove/update them. I guess we could add some additional logic to try and be clever and find the update even if one of the builds is missing or modified. Think about how bugzilla bugs are handled in IRC. Bugs all have ID numbers. Why should updates be different? I vote for static IDs because I have run into the case of modified updates and broken URLs. Adam, can you not pursue an enhancement to the IRC bot that translates bug URLs into descriptions to also handle bodhi IDs? This is surreal. Are you trying to single handely kill what little real user testing is being done on the various Fedora releases? Now you want to make users bring up yet another tool - an IRC client? Why not just be done with it, and bury the reports in a locked filing cabinet in a barred sub-basement room labeled Ignore me - do not open? Perhaps there are simpler alternatives. The whole point of the updates testing reports is to provide information that _quickly_ makes folks aware of what new packages are available in updates-testing for a given release. Real users know the names of the packages that they use. That information is now gone - hidden behind a VERY SLOW process of is following links. I tried the first day the report changed. I gave up, as it was taking a significant time to bring up each link. The first reaction in the proven testers meeting was that the new reports were not at all useful, and should be immediately reverted. It has been a number of weeks since then, but it appears we now have something else instead. The report generated by the latest iteration is broken. This has already been noted by others, and they have made suggestions as to how to fix this (listing the package names below the URL). I had an idea a number of weeks ago to increase the visibility of those packages sitting for long periods of time in updates-testing, in faint hope that someone would care enough to test and give karma. Kevin Fenzi encouraged me to open a TRAC request. It was to simply show the number of days that each package has been in updates-testing in the report, something that bodhi should have readily at hand. Having URLs that are not brokwn is important. Showing the package name and the number of days it has been in updates testing (to the left of the name) is equally important. Please consider doing all of these in your next revision. Al -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tue, 2011-10-25 at 19:03 -0400, Al Dunsmuir wrote: Think about how bugzilla bugs are handled in IRC. Bugs all have ID numbers. Why should updates be different? I vote for static IDs because I have run into the case of modified updates and broken URLs. Adam, can you not pursue an enhancement to the IRC bot that translates bug URLs into descriptions to also handle bodhi IDs? This is surreal. Are you trying to single handely kill what little real user testing is being done on the various Fedora releases? Now you want to make users bring up yet another tool - an IRC client? Why not just be done with it, and bury the reports in a locked filing cabinet in a barred sub-basement room labeled Ignore me - do not open? Erm, put down the flamethrower. IRC was just an example. Really this is just a tooling question of whether it's overall more convenient to have Bodhi use IDs and then implement convenience scripts/tools wherever we refer to the updates which can identify them - like a bot in IRC, and a modification to the updates-testing email to make it use the old format, etc - or whether it's better to have Bodhi use NEVRs and then have to somehow deal with the problem of updates with dozens of packages, and the problem of updates which are edited to include *different* NEVRs. Whichever one turns out to be the most efficient solution is what we'll go with, I'm sure. No-one's out to get you. Deep breaths. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
Adam Williamson wrote: On Tue, 2011-10-25 at 17:18 -0400, Luke Macken wrote: bodhi v0.8.3 Yesterday I pushed out a new bugfix release of bodhi into production. The bodhi-client is currently on it's way to updates-testing for all releases. Server fixes - Default to update ID-based URLs https://fedorahosted.org/bodhi/ticket/632 In case you hadn't noticed, response to this has so far been pretty negative. Coming from someone (me) who often does updates involving 1 pkg that sometimes require removing/adding components after initial submission, I very much welcome the new ID-based (default) URL's -- rex -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: New bodhi bugfix release in production
On Tue, 2011-10-25 at 20:30 -0500, Rex Dieter wrote: In case you hadn't noticed, response to this has so far been pretty negative. Coming from someone (me) who often does updates involving 1 pkg that sometimes require removing/adding components after initial submission, I very much welcome the new ID-based (default) URL's Indeed, packagename-based urls are not very practical for 60+ package updates... -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel