Le blog de pingou - Tag - Fedora - CommentsLe blog de pingou, ses actualités sur Fedora, ses RPMs, ses tests, son Linux... :-)
Pingou's weblog, his fedora's news, his RPMs, his tests, his Linux... :-)2022-02-17T10:46:15+01:00pingouurn:md5:66db5ce1ed1a80cb2f424695b4bb7780DotclearSome stats about our dist-git usage - Paul W. Frieldsurn:md5:d24c8f7d3ca4e546111da9c3dac967582017-02-23T21:29:34+01:002017-02-23T21:29:34+01:00Paul W. Frields<p>This is interesting data. So on average, we'd be looking at 17-23 commits/hour to trigger potential CI work. This assumes every package change is a trigger, of course. Thanks for finding these stats!</p>FESCo vote history - Kevin Koflerurn:md5:b44da96f4ac1e4d34cb1e7bdc01c79eb2015-06-30T11:13:36+02:002015-06-30T10:13:36+02:00Kevin Kofler<p>You should maybe ask for the reasons for this. I think there are several factors:<br />
* insufficient publicization of the vote. For example, in past elections, announcements and "I voted" badges were spammed all over Planet Fedora. This time, there was only the interview series from Fedora Magazine.<br />
* bad timing for the election. In most of Europe, it was the last or second-to-last week before the summer break. In the USA, it was IN the summer break.<br />
* unhappiness with what FESCo is doing and/or a feeling that a new FESCo won't change anything anyway (also because important decisions now get imposed by the Council/Board or even Red Hat). (See "Fedora.Next", a big failure that turned Fedora on the desktop into a GNOME monocracy.)<br />
* unhappiness with democracy in Fedora. For example, the new "Council" STILL has appointed seats. (It is also neither an actual council nor a directly elected board, but a mix of both and of appointed seats. With that composition, it fails both at the "council" idea of representing the subcommunities AND at allowing the community to elect its board.)</p>PyCon 2015 - Montreal - pingouurn:md5:ee7d9638d459403d681333c57f8ab4d42015-04-22T14:09:27+02:002015-04-22T13:09:27+02:00pingou<p>I would expect it to be on the youtube channel, I don't know where to look for it otherwise.</p>
<p>Maybe there is just a small delay on posting it? Otherwise, might be worth asking the pycon crew if they have more info.</p>PyCon 2015 - Montreal - Jamesurn:md5:c5753ea628cc8760eed4e633bb9e67862015-04-22T09:50:20+02:002015-04-22T08:50:20+02:00James<p>Any idea where Gary Bernhardt's keynote video is? I didn't get to see it in person. Can't find it on the pycon youtube channel.</p>
<p>Thanks!</p>About SourceForge and anitya - pingouurn:md5:fec4e0582633f70c1be51711e1115e542015-02-25T22:24:30+01:002015-02-25T22:24:30+01:00pingou<p>Option #2 would only affect the sourceforge backend yes, it's the only backend (iirc) that uses the version_url field to store info about the name of the tarball (currently).</p>
<p>It's the easiest option and clearly it would work for most things, but that does give us some ugly project names :-/</p>
<p>Saw the proposal from scop at: <a href="https://github.com/fedora-infra/anitya/issues/80#issuecomment-76054615" title="https://github.com/fedora-infra/anitya/issues/80#issuecomment-76054615" rel="ugc nofollow">https://github.com/fedora-infra/ani...</a> ?</p>Check your packages in pkgdb and anitya - pingouurn:md5:d42e0d5f440dc982cbcad5a2d8c7f1bf2015-02-25T22:22:18+01:002015-02-25T22:22:18+01:00pingou<p>Ahah, sorry about that (ok not entirely) ;-)</p>
<p>Cool stuff on the lightsaber indeed.</p>About SourceForge and anitya - Ralph Beanurn:md5:f9e23824c36444174e61b0e239e0c22b2015-02-25T19:22:13+01:002015-02-25T19:22:13+01:00Ralph Bean<p>I think I like option #2 but I don't think I fully understand the scope of it yet. Would making that change *only* affect sourceforge project listings? Or would it affect all project listings?</p>
<p>#1 seems like it would introduce confusion in the code since (some number of) things have been written expecting the unique constraint to combine only the project name and the homepage.</p>Check your packages in pkgdb and anitya - Ralph Beanurn:md5:8e9b5d2920a854a4d2a12aa83486ba182015-02-25T19:15:29+01:002015-02-25T19:15:29+01:00Ralph Bean<p>Nice work!</p>
<p>You blew up the spot on ship-it on the list, so I had to add some dev instructions... ;P <a href="https://github.com/ralphbean/ship-it#warning" title="https://github.com/ralphbean/ship-it#warning" rel="ugc nofollow">https://github.com/ralphbean/ship-i...</a></p>
<p>FWIW, I added a script to my personal battery that forces a new check of a package on Anitya. This might be useful to others: <a href="https://github.com/ralphbean/lightsaber/blob/develop/roles/development/files/bin/check-anitya.py" title="https://github.com/ralphbean/lightsaber/blob/develop/roles/development/files/bin/check-anitya.py" rel="ugc nofollow">https://github.com/ralphbean/lights...</a></p>New branch request process - pingouurn:md5:54f38bbccbf9c6d756976384c85e6b9a2015-02-25T12:15:26+01:002015-02-26T08:04:47+01:00pingou<p>I used http://asciiflow.com/ for the diagram and did a few manual adjustment</p>New branch request process - Leourn:md5:9eead1305faf3d533032fe08363d4c972015-01-26T16:05:54+01:002015-01-26T16:05:54+01:00Leo<p>Helllo,</p>
<p>What did you use to make the ascii diagram btw?</p>Firefox private browsing directly - pingouurn:md5:d97c90611762754c9db0329288dd32602014-12-30T13:19:36+01:002014-12-30T13:19:36+01:00pingou<p>I find this to be easier than switching between profiles.</p>
<p>I think the private mode is similar or very close to a profile set to never remember anything with the addition that cookies are deleted when the window is closed, there is no history, no password, no forms pre-filled & so on.</p>
<p>You could probably achieve the same with 2 profils, but it would be more work :)</p>Firefox private browsing directly - Sudhir Khangerurn:md5:682ddc8b112890de478a71619d459f962014-12-30T12:32:25+01:002014-12-30T13:17:29+01:00Sudhir Khanger<p>You can simply create another profile and set it to "never remember anything." What's the difference between firefox -private-window vs set History to "Never remember anything"?</p>Infra FAD 2014 - Part 2: Ansible - pingouurn:md5:e7f4615762013da7b3162e2a7ce357862014-12-13T16:42:04+01:002014-12-13T16:42:04+01:00pingou<p>Ok, I won't :)</p>Infra FAD 2014 - Part 2: Ansible - Kushal Dasurn:md5:1dacec62976c17bd3749e0b26c8227232014-12-13T11:41:49+01:002014-12-13T11:41:49+01:00Kushal Das<p>Don't tell me that person in right side is threebean :)</p>New package, new branch, new workflow? - Paul W. Frieldsurn:md5:bf336100ac8c084dea0b10ac13bc6f322014-07-23T19:12:35+02:002014-07-23T22:16:24+02:00Paul W. Frields<p>Overall it looks like the new procedure actually saves the manual step of cvsadmin setting fedora-cvs+. The other "added" steps don't look like manual steps, as you pointed out. And new branches look pretty easy too.</p>Evolution of our contributors in Fedora - pingouurn:md5:12d4fb57ecb0531bc47faa4e5821c4d02014-02-04T21:55:49+01:002014-02-04T21:56:01+01:00pingou<p>@ Neville, This will be quite tricky as we mostly do not register when people leave. Maybe they'll deactivate their account but mostly not and I'm not sure (I'm in fact pretty sure in most cases we are not).</p>
<p>@Mirek, you're just being modest there ;-)<br />
But then either we have much more contributors (and F20 is much more popular than the previous ones) or it's related to the bump in planet related fedmsg messages we can see in thisweekinfedora.org</p>Evolution of our contributors in Fedora - Neville A. Crossurn:md5:3c25e4e5c5043cb8bcdc923e031ee9a72014-02-04T20:08:18+01:002014-02-04T20:08:18+01:00Neville A. Cross<p>As Fedora ambassador some times I struggle to make presentations without any statistics. It is a reasonable question to ask: "How many people use Fedora".</p>
<p>But recently I have think about other concern. What is the burn out rate of Fedora collaborators? What is the average time that a person is an active collaborator? This may lead us to think about when will be most useful to sponsor people to events. In the other hand, we can find out a threshold, that people with more than "x" years, will keep going. Another issue to take into account is about their employer, obviously people hired by Red Hat to do stuff in Fedora should have a different metrics.</p>Evolution of our contributors in Fedora - Mirek Suchýurn:md5:9b637af7086e16f5ecd808c9f85c76292014-02-04T20:05:25+01:002014-02-04T21:48:16+01:00Mirek Suchý<p>While I would like to see Copr as cause, the true is that "only" 117 developers submited at least one task to Copr so far. So it could not justify jump of 1500 heads.</p>
<p>And do not forget that in December was released Fedora 20.</p>Dynamic point of contact assignment - pingouurn:md5:2c051f4daec142f32dfa26913fd7b2512014-02-04T07:24:22+01:002014-02-04T07:24:22+01:00pingou<p>Hey Kalev,</p>
<p>According to these stats you are already taking care of these additional 127 packages, so that shouldn't change much from now ;-)</p>Dynamic point of contact assignment - Kalevurn:md5:f53e1223e0601e021d304fcf5e96ae8e2014-02-04T00:20:47+01:002014-02-04T07:16:44+01:00Kalev<p>"kalev gains 127 packages : 43 -> 170" -- please no, I would go crazy if I have to take responsibility for 127 additional packages.</p>
<p>Otherwise it's really cool that someone is working on pkgdb again, keep up the good work!</p>