The Fedora_28_RC_1.1 compose [1] is considered as GOLD and it is going to be shipped on 2018-May-01 as Fedora 28 Final release.
For more information please check the meeting minutes [2] from the Go/No-Go meeting.
[1] http://dl.fedoraproject.org/pub/alt/stage/28_RC-1.1/ [2] https://meetbot.fedoraproject.org/fedora-meeting-1/2018-04-26/f28-final-go-n...
Regards, Jan
On Thu, 2018-04-26 at 20:41 +0200, Jan Kurik wrote:
The Fedora_28_RC_1.1 compose [1] is considered as GOLD and it is going to be shipped on 2018-May-01 as Fedora 28 Final release.
For more information please check the meeting minutes [2] from the Go/No-Go meeting.
For the record, I have determined to at least my own satisfaction that this is the first *ever* on-time Fedora release. Go team!
For releases from 11 onwards it's easy to demonstrate that they slipped: the original dates were kept in their wiki schedule pages with a strike-through each time they slipped, so you just go to each release's page and verify it has some strikethroughs for the 'Final release" date.
For releases from 7 to 10 this wasn't done - the 'official' schedule page was just silently edited when the schedule slipped, and as the wiki at that point in history was MoinMoin not Mediawiki, we don't have the edit histories any more. However, I've found references to earlier schedules around the place (meeting logs, mailing list archives, forum posts, sometimes John Poelstra's blog) that sufficiently indicate there *was* an 'official' schedule with an earlier release date than the actual one in each case. If anyone's as sadly nerdy as me, I can provide specific references for each of these releases.
For releases from FC2 to FC6 you can find the schedules in the Wayback Machine archives for http://fedora.redhat.com/participate/schedule/ :
https://web.archive.org/web/20030701000000*/http://fedora.redhat.com/partici...
For these releases, the schedule was never claimed to be 'official', it was always referred to as a 'draft'. But I came up with a pretty conservative definition of 'delayed': I looked at the page approx. 3 weeks before the *actual* release date for each of these releases. In each case, the Final release date that was scheduled 3 weeks before the *actual* release date didn't match, it was earlier. I think it's reasonable to consider this as a 'slip' in each case - if we didn't even meet the schedule we had planned less than a month before release, it's pretty hard to argue that's not a 'slip'.
FC1 is the trickiest. I don't think any FC1 development schedule was ever really made public. So for that one I got creative. There's an article on LWN - written by Joe Brockmeier no less! - around the time of the release:
https://lwn.net/Articles/56036/
It was written on Wednesday 2018-10-29, and states in part:
"With the first stable release of the Fedora Core scheduled for early next week..."
Now, the release actually happened on 2018-11-05. Which *is* 'next week' from 2018-10-29, but it's also Wednesday of the next week. I am going to hold that no-one can reasonably claim Wednesday is "early" in a given week. Surely only Monday and Tuesday (and Sunday, depending on what day you think a week starts on) can plausibly claim to be "early". On that basis, I'm gonna say FC1 was at least a day late from the schedule in place a week before it came out, and on that basis...every release from FC1 to F27 was at least a day late. And F28 is the first one that's ever been on time.
:P
On Thu, Apr 26, 2018 at 4:03 PM Adam Williamson adamwill@fedoraproject.org wrote:
On Thu, 2018-04-26 at 20:41 +0200, Jan Kurik wrote:
The Fedora_28_RC_1.1 compose [1] is considered as GOLD and it is going to be shipped on 2018-May-01 as Fedora 28 Final release.
For more information please check the meeting minutes [2] from the Go/No-Go meeting.
For the record, I have determined to at least my own satisfaction that this is the first *ever* on-time Fedora release. Go team!
For releases from 11 onwards it's easy to demonstrate that they slipped: the original dates were kept in their wiki schedule pages with a strike-through each time they slipped, so you just go to each release's page and verify it has some strikethroughs for the 'Final release" date.
For releases from 7 to 10 this wasn't done - the 'official' schedule page was just silently edited when the schedule slipped, and as the wiki at that point in history was MoinMoin not Mediawiki, we don't have the edit histories any more. However, I've found references to earlier schedules around the place (meeting logs, mailing list archives, forum posts, sometimes John Poelstra's blog) that sufficiently indicate there *was* an 'official' schedule with an earlier release date than the actual one in each case. If anyone's as sadly nerdy as me, I can provide specific references for each of these releases.
For releases from FC2 to FC6 you can find the schedules in the Wayback Machine archives for http://fedora.redhat.com/participate/schedule/ :
https://web.archive.org/web/20030701000000*/http://fedora.redhat.com/partici...
For these releases, the schedule was never claimed to be 'official', it was always referred to as a 'draft'. But I came up with a pretty conservative definition of 'delayed': I looked at the page approx. 3 weeks before the *actual* release date for each of these releases. In each case, the Final release date that was scheduled 3 weeks before the *actual* release date didn't match, it was earlier. I think it's reasonable to consider this as a 'slip' in each case - if we didn't even meet the schedule we had planned less than a month before release, it's pretty hard to argue that's not a 'slip'.
FC1 is the trickiest. I don't think any FC1 development schedule was ever really made public. So for that one I got creative. There's an article on LWN - written by Joe Brockmeier no less! - around the time of the release:
It was written on Wednesday 2018-10-29, and states in part:
"With the first stable release of the Fedora Core scheduled for early next week..."
Now, the release actually happened on 2018-11-05. Which *is* 'next week' from 2018-10-29, but it's also Wednesday of the next week. I am going to hold that no-one can reasonably claim Wednesday is "early" in a given week. Surely only Monday and Tuesday (and Sunday, depending on what day you think a week starts on) can plausibly claim to be "early". On that basis, I'm gonna say FC1 was at least a day late from the schedule in place a week before it came out, and on that basis...every release from FC1 to F27 was at least a day late. And F28 is the first one that's ever been on time.
:P
Welp....
I don't know what to say except... We're going to get snark for this anyway, I bet!
On Thu, 2018-04-26 at 13:02 -0700, Adam Williamson wrote:
It was written on Wednesday 2018-10-29, and states in part:
"With the first stable release of the Fedora Core scheduled for early next week..."
Now, the release actually happened on 2018-11-05. Which *is* 'next week' from 2018-10-29
er, of course, change those 2018s to 2003s.
On 2018-04-26 16:02, Adam Williamson wrote:
FC1 is the trickiest. I don't think any FC1 development schedule was ever really made public. So for that one I got creative. There's an article on LWN - written by Joe Brockmeier no less! - around the time of the release:
https://lwn.net/Articles/56036/
It was written on Wednesday 2018-10-29, and states in part:
"With the first stable release of the Fedora Core scheduled for early next week..."
Now, the release actually happened on 2018-11-05. Which *is* 'next
s/2018/2003/ from what I can see. Fingers on auto-fill?
On to, 26 huhti 2018, Adam Williamson wrote:
On Thu, 2018-04-26 at 20:41 +0200, Jan Kurik wrote:
The Fedora_28_RC_1.1 compose [1] is considered as GOLD and it is going to be shipped on 2018-May-01 as Fedora 28 Final release.
For more information please check the meeting minutes [2] from the Go/No-Go meeting.
For the record, I have determined to at least my own satisfaction that this is the first *ever* on-time Fedora release. Go team!
Great!
For releases from 11 onwards it's easy to demonstrate that they slipped: the original dates were kept in their wiki schedule pages with a strike-through each time they slipped, so you just go to each release's page and verify it has some strikethroughs for the 'Final release" date.
For releases from 7 to 10 this wasn't done - the 'official' schedule page was just silently edited when the schedule slipped, and as the wiki at that point in history was MoinMoin not Mediawiki, we don't have the edit histories any more. However, I've found references to earlier schedules around the place (meeting logs, mailing list archives, forum posts, sometimes John Poelstra's blog) that sufficiently indicate there *was* an 'official' schedule with an earlier release date than the actual one in each case. If anyone's as sadly nerdy as me, I can provide specific references for each of these releases.
For releases from FC2 to FC6 you can find the schedules in the Wayback Machine archives for http://fedora.redhat.com/participate/schedule/ :
https://web.archive.org/web/20030701000000*/http://fedora.redhat.com/partici...
For these releases, the schedule was never claimed to be 'official', it was always referred to as a 'draft'. But I came up with a pretty conservative definition of 'delayed': I looked at the page approx. 3 weeks before the *actual* release date for each of these releases. In each case, the Final release date that was scheduled 3 weeks before the *actual* release date didn't match, it was earlier. I think it's reasonable to consider this as a 'slip' in each case - if we didn't even meet the schedule we had planned less than a month before release, it's pretty hard to argue that's not a 'slip'.
FC1 is the trickiest. I don't think any FC1 development schedule was ever really made public. So for that one I got creative. There's an article on LWN - written by Joe Brockmeier no less! - around the time of the release:
https://lwn.net/Articles/56036/
It was written on Wednesday 2018-10-29, and states in part:
"With the first stable release of the Fedora Core scheduled for early next week..."
Now, the release actually happened on 2018-11-05. Which *is* 'next week' from 2018-10-29, but it's also Wednesday of the next week. I am
I think it is a bit of a stretch for a release archeology to place events into the future, isn't it? :)_
going to hold that no-one can reasonably claim Wednesday is "early" in a given week. Surely only Monday and Tuesday (and Sunday, depending on what day you think a week starts on) can plausibly claim to be "early". On that basis, I'm gonna say FC1 was at least a day late from the schedule in place a week before it came out, and on that basis...every release from FC1 to F27 was at least a day late. And F28 is the first one that's ever been on time.
:P
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-leave@lists.fedoraproject.org
On Thu, Apr 26, 2018 at 01:02:32PM -0700, Adam Williamson wrote:
*was* an 'official' schedule with an earlier release date than the actual one in each case. If anyone's as sadly nerdy as me, I can provide specific references for each of these releases.
Um, yes please.
On Thu, 2018-04-26 at 16:54 -0400, Matthew Miller wrote:
On Thu, Apr 26, 2018 at 01:02:32PM -0700, Adam Williamson wrote:
*was* an 'official' schedule with an earlier release date than the actual one in each case. If anyone's as sadly nerdy as me, I can provide specific references for each of these releases.
Um, yes please.
FINE.
AdamW Industries hereby presents The Utterly Comprehensive And Definitive History Of Fedora Release Delays.
All actual release dates mentioned here are taken from https://fedoraproject.org/wiki/Releases/HistoricalSchedules , cross checked with https://distrowatch.com/table.php?distribution=fedora .
Fedora Core 1: https://www.redhat.com/archives/fedora-devel-list/2003-October/msg01178.html "We had to respin FC1 today for a non-technical issue...so we have to slip until after we hear back from them". Thanks to Thomas Moschny for spotting/remembering that one!
Fedora Core 2: https://web.archive.org/web/20040406202016/http://fedora.redhat.com/particip... That's the schedule on 2004-04-06 per Wayback Machine, the last snapshot taken before the actual release. It shows "Release open" on 2004-05-10. Actual release was 2004-05-18.
Fedora Core 3: https://web.archive.org/web/20041014022956/http://fedora.redhat.com/particip... That's the schedule on 2004-10-14 per Wayback Machine. It shows "Release open" on 2004-11-01. By 2004-10-22 that date had been changed to 2004-11-08, which was the actual final release date. So there was a one-week slip less than three weeks before release. (This wasn't the only slip; back in August, for instance, the date was shown as 2004-10-25: https://web.archive.org/web/20040810000233/http://fedora.redhat.com:80/parti... )
Fedora Core 4: https://web.archive.org/web/20050212015551/http://fedora.redhat.com/particip... On 2005-02-12, the "Release open" date was 2005-05-16. https://web.archive.org/web/20050517233132/http://fedora.redhat.com/particip... by 2005-05-17, it had changed to 2005-06-06 (which was still the target at least as late as 2005-05-29). Actual release was 2005-06-13.
Fedora Core 5: https://web.archive.org/web/20051211051141/http://fedora.redhat.com:80/About... On 2005-12-11, the "Release open" date was 2006-02-27. https://web.archive.org/web/20060217043822/http://fedora.redhat.com:80/About... By 2006-02-17, it had changed to 2006-03-15. Actual release date was 2006-03-20.
Fedora Core 6: https://web.archive.org/web/20060601062428/http://fedoraproject.org:80/wiki/... On 2006-06-01, GA date (this appears to be when we started calling it 'GA', fact fans!) was 2006-09-27. https://web.archive.org/web/20061010085518/http://fedoraproject.org:80/wiki/... by 2006-10-10, it had changed to 2006-10-17. We were still wrong just a week out, though: it actually was released 2006-10-24.
Fedora 7: I found various references for this, but carrying on using the Wayback Machine on the wiki seems like a good a plan as any. So: https://web.archive.org/web/20070319211234/http://fedoraproject.org/wiki/Rel... On 2007-03-19, GA date was 2007-05-24. It stayed that way till 2007-05-08: https://web.archive.org/web/20070508223935/http://fedoraproject.org/wiki/Rel... Then around 2007-05-10 the date was changed to 2007-05-?? (yes, really) with a note "Final freeze and GA have slipped by at least a week from their original dates of May 10 and May 24, respectively. When more firm dates are determined, this will be updated again.": https://web.archive.org/web/20070510074633/http://fedoraproject.org/wiki/Rel... indeed, the release finally happened on 2007-05-31. Note that in the initial tentative schedule discussed by the Board: https://lwn.net/Articles/213942/ , release was on 2007-04-24.
Fedora 8: https://web.archive.org/web/20170318225502/https://fedoraproject.org/wiki/Re... On 2007-06-02, GA was scheduled for 2007-10-31 (LWN posted a story about this on 2007-05-31: https://lwn.net/Articles/236468/), though note to be fair, the schedule was explicitly marked as a "draft" at this point. By 2007-06-30, the "draft" note was removed, and the GA date was set for 2007-11-07: https://web.archive.org/web/20070630004529/https://fedoraproject.org/wiki/Re... This is by far the closest we ever came to an 'on time' release before, because actual release happened just a day later, on 2007-11-08. So if we discount the 'draft' schedule, F8 was only delayed by one day.
Fedora 9: https://forums.fedoraforum.org/showthread.php?186443-Fedora-9-Release-Date-S... The forums actually have a thread recording that a two week slip was officially announced by FESCo around 2008-04-17. This is captured in the Wayback Machine: https://web.archive.org/web/20080409211329/https://fedoraproject.org/wiki/Re... on 2008-04-09, "final release" was listed as 2008-04-29, but by 2008-04-20, that had changed to 2008-05-13: https://web.archive.org/web/20080420060131/https://fedoraproject.org/wiki/Re... which was indeed the final release date.
Fedora 10: John Poelstra announced a three-week slip to the schedule due to infrastructure outages: http://johnpoelstra.com/fedora-10-schedule-update/ That post is unusefully not dated, but it was at least before 2008-09-09. It cites the new GA date as 2008-11-18, but the actual release was 2008-11-25, so there was an additional one week slip. The earliest 'official' schedule (without notes that it had not yet been ratified by FESCo) I can find is: https://web.archive.org/web/20080717094755/https://fedoraproject.org/wiki/Re... which lists GA as 2008-10-28.
From Fedora 11 onwards, the official schedule page in the wiki records all official schedule slips via strikethroughs, so I'll just link to each page and specify the earliest official GA date and the actual. For the record, no, I don't know why the hell whoever makes this pages couldn't make up their damn mind about what the URL ought to be for so long. :P
Fedora 11: https://fedoraproject.org/wiki/Releases/11/Schedule Earliest scheduled: 2009-05-26 Actual: 2009-06-09
Fedora 12: https://fedoraproject.org/wiki/Releases/12 Earliest scheduled: 2009-11-03 Actual: 2009-11-17
Fedora 13: https://fedoraproject.org/wiki/Releases/13/Schedule Earliest scheduled: 2010-05-11 Actual: 2010-05-25
Fedora 14: https://fedoraproject.org/wiki/Releases/14/Schedule Earliest scheduled: 2010-10-26 Actual: 2010-11-02
Fedora 15: https://fedoraproject.org/wiki/Releases/15 Earliest scheduled: 2011-05-10 Actual: 2011-05-24 (This was actually a minor miracle, as this was the release we introduced *both* systemd *and* GNOME 3...)
Fedora 16: https://fedoraproject.org/wiki/Releases/16/Schedule Earliest scheduled: 2011-10-25 Actual: 2011-11-08
Fedora 17: https://fedoraproject.org/wiki/Releases/17/Schedule Earliest scheduled: 2012-05-08 Actual: 2012-05-29
Fedora 18: https://fedoraproject.org/wiki/Releases/18/Schedule Earliest scheduled: 2012-11-06 Actual: 2013-01-15 (If you're wondering why this one was so rough: it was the release where we introduced the new anaconda UI.)
Fedora 19: https://fedoraproject.org/wiki/Releases/19/Schedule Earliest scheduled: 2013-06-25 Actual: 2013-07-02
Fedora 20: https://fedoraproject.org/wiki/Releases/20/Schedule Earliest scheduled: 2013-11-26 Actual: 2013-12-17
Fedora 21: https://fedoraproject.org/wiki/Releases/21/Schedule Earliest scheduled: 2014-11-18 Actual: 2014-12-09
Fedora 22: https://fedoraproject.org/wiki/Releases/22/Schedule Earliest scheduled: 2015-05-19 Actual: 2015-05-26
Fedora 23: https://fedoraproject.org/wiki/Releases/23/Schedule Earliest scheduled: 2015-10-27 Actual: 2015-11-03
Fedora 24: https://fedoraproject.org/wiki/Releases/24/Schedule Earliest scheduled: 2016-05-17 Actual: 2016-06-21 (I don't recall why this one was so rough...glibc and gcc bumps?)
Fedora 25: https://fedoraproject.org/wiki/Releases/25/Schedule Earliest scheduled: 2016-11-08 Actual: 2016-11-22
Fedora 26: https://fedoraproject.org/wiki/Releases/26/Schedule Earliest scheduled: 2017-06-06 Actual: 2017-07-11 (Again I forget what the deal was here, but it had a GCC bump, cflags updates, OpenSSL 1.1.0 and DNF 2.0, which could be it.)
Fedora 27: https://fedoraproject.org/wiki/Releases/27/Schedule Earliest scheduled: 2017-10-24 Actual: 2017-11-14
So there we go! Fedora 8 is clearly the closest we ever came before to shipping 'on time', with a slip of only one day from the first non- draft public schedule. There were a few other releases with slips of only a week. But F28 is truly the first to be right on time, assuming we don't run into any issues over the weekend. No pressure, releng/websites folks :P
Once upon a time, Adam Williamson adamwill@fedoraproject.org said:
For the record, I have determined to at least my own satisfaction that this is the first *ever* on-time Fedora release. Go team!
Ehh, release schedules are more like guidelines anyway... :)
Chris Adams linux@cmadams.net wrote:
Once upon a time, Adam Williamson adamwill@fedoraproject.org said:
For the record, I have determined to at least my own satisfaction that this is the first *ever* on-time Fedora release. Go team!
Ehh, release schedules are more like guidelines anyway... :)
In Fedora "guidelines" usually means "strict regulations". :-Þ
Björn Persson
Wow, that is some serious digging, and good to know that we are on time :)
Hope we continue it for future releases.
On Thu, Apr 26, 2018 at 4:03 PM Adam Williamson adamwill@fedoraproject.org wrote:
On Thu, 2018-04-26 at 20:41 +0200, Jan Kurik wrote:
The Fedora_28_RC_1.1 compose [1] is considered as GOLD and it is going to be shipped on 2018-May-01 as Fedora 28 Final release.
For more information please check the meeting minutes [2] from the Go/No-Go meeting.
For the record, I have determined to at least my own satisfaction that this is the first *ever* on-time Fedora release. Go team!
For releases from 11 onwards it's easy to demonstrate that they slipped: the original dates were kept in their wiki schedule pages with a strike-through each time they slipped, so you just go to each release's page and verify it has some strikethroughs for the 'Final release" date.
For releases from 7 to 10 this wasn't done - the 'official' schedule page was just silently edited when the schedule slipped, and as the wiki at that point in history was MoinMoin not Mediawiki, we don't have the edit histories any more. However, I've found references to earlier schedules around the place (meeting logs, mailing list archives, forum posts, sometimes John Poelstra's blog) that sufficiently indicate there *was* an 'official' schedule with an earlier release date than the actual one in each case. If anyone's as sadly nerdy as me, I can provide specific references for each of these releases.
For releases from FC2 to FC6 you can find the schedules in the Wayback Machine archives for http://fedora.redhat.com/participate/schedule/ :
https://web.archive.org/web/20030701000000*/http://fedora.redhat.com/partici...
For these releases, the schedule was never claimed to be 'official', it was always referred to as a 'draft'. But I came up with a pretty conservative definition of 'delayed': I looked at the page approx. 3 weeks before the *actual* release date for each of these releases. In each case, the Final release date that was scheduled 3 weeks before the *actual* release date didn't match, it was earlier. I think it's reasonable to consider this as a 'slip' in each case - if we didn't even meet the schedule we had planned less than a month before release, it's pretty hard to argue that's not a 'slip'.
FC1 is the trickiest. I don't think any FC1 development schedule was ever really made public. So for that one I got creative. There's an article on LWN - written by Joe Brockmeier no less! - around the time of the release:
https://lwn.net/Articles/56036/
It was written on Wednesday 2018-10-29, and states in part:
"With the first stable release of the Fedora Core scheduled for early next week..."
Now, the release actually happened on 2018-11-05. Which *is* 'next week' from 2018-10-29, but it's also Wednesday of the next week. I am going to hold that no-one can reasonably claim Wednesday is "early" in a given week. Surely only Monday and Tuesday (and Sunday, depending on what day you think a week starts on) can plausibly claim to be "early". On that basis, I'm gonna say FC1 was at least a day late from the schedule in place a week before it came out, and on that basis...every release from FC1 to F27 was at least a day late. And F28 is the first one that's ever been on time.
:P
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-leave@lists.fedoraproject.org
2018-04-26 22:02 GMT+02:00 Adam Williamson adamwill@fedoraproject.org:
On that basis, I'm gonna say FC1 was at least a day late from the schedule in place a week before it came out,
It even slipped 'officially': https://www.redhat.com/archives/fedora-devel-list/2003-October/msg01178.html :)
- Thomas
On Fri, 2018-04-27 at 18:29 +0200, Thomas Moschny wrote:
2018-04-26 22:02 GMT+02:00 Adam Williamson adamwill@fedoraproject.org:
On that basis, I'm gonna say FC1 was at least a day late from the schedule in place a week before it came out,
It even slipped 'officially': https://www.redhat.com/archives/fedora-devel-list/2003-October/msg01178.html :)
Wow, nice find, thank you Thomas! :)