<@salimma:fedora.im>
17:00:41
!startmeeting FESCO (2025-04-22)
<@meetbot:fedora.im>
17:00:43
Meeting started at 2025-04-22 17:00:41 UTC
<@meetbot:fedora.im>
17:00:43
The Meeting name is 'FESCO (2025-04-22)'
<@fale:fale.io>
17:00:49
!hi
<@decathorpe:fedora.im>
17:00:53
!hi
<@zodbot:fedora.im>
17:00:54
Fabio Alessandro Locati (fale) - he / him / his
<@zodbot:fedora.im>
17:00:54
Fabio Valentini (decathorpe) - he / him / his
<@salimma:fedora.im>
17:00:59
!meetingname fesco
<@meetbot:fedora.im>
17:01:01
The Meeting Name is now fesco
<@salimma:fedora.im>
17:01:02
!group members fesco
<@zodbot:fedora.im>
17:01:05
Members of fesco: David Cantrell, Fabio Valentini, Fabio Alessandro Locati, Tomáš Hrčka, Kevin Fenzi, Matthew Miller, ngompa (@conan_kudo:matrix.org, @ngompa:fedora.im, @pharaoh_atem:opensuse.org, @ngompa:kde.org, @ngompa:almalinux.im), salimma (@michel-slm:matrix.org, @salimma:fedora.im, @michel:one.ems.host), Stephen Gallagher, Zbigniew Jędrzejewski-Szmek
<@dcantrell:fedora.im>
17:01:06
!hi
<@zodbot:fedora.im>
17:01:07
David Cantrell (dcantrell) - he / him / his
<@salimma:fedora.im>
17:01:08
!topic Init Process
<@salimma:fedora.im>
17:01:13
!hi
<@zodbot:fedora.im>
17:01:14
Michel Lind (salimma) - he / him / his
<@salimma:fedora.im>
17:01:27
hello everyone! long(er) time no see
<@jvanek:matrix.org>
17:02:26
hello!
<@nirik:matrix.scrye.com>
17:02:28
morning
<@salimma:fedora.im>
17:02:28
let's wait a couple of minutes - oh hi jvanek
<@humaton:fedora.im>
17:03:34
hello
<@zbyszek:fedora.im>
17:03:45
!hi
<@zodbot:fedora.im>
17:03:47
Zbigniew Jędrzejewski-Szmek (zbyszek)
<@salimma:fedora.im>
17:04:11
we have 7 members already, if I don't miscount
<@brakoose:matrix.org>
17:04:26
!hi
<@salimma:fedora.im>
17:04:29
Conan Kudo: Stephen Gallagher joining?
<@zodbot:fedora.im>
17:04:29
Godfred Addai Amoako (brakoose) - he / him / his
<@conan_kudo:matrix.org>
17:04:41
coming
<@conan_kudo:matrix.org>
17:04:44
!hi
<@zodbot:fedora.im>
17:04:48
Neal Gompa (ngompa) - he / him / his
<@salimma:fedora.im>
17:05:27
it's 5 past, let's start
<@salimma:fedora.im>
17:05:55
we can start with the JDK issue first if nobody objects, so Jiri doesn't have to sit through one hour of meeting (I doubt it will be that long anyway)
<@salimma:fedora.im>
17:06:18
!topic #3385 Change: Java25 And No More System JDK
<@salimma:fedora.im>
17:06:29
!fesco 3385
<@zodbot:fedora.im>
17:06:30
● **Assignee:** jvanek
<@zodbot:fedora.im>
17:06:30
● **Last Updated:** 54 minutes ago
<@zodbot:fedora.im>
17:06:30
● **Opened:** 3 weeks ago by amoloney
<@zodbot:fedora.im>
17:06:30
<@zodbot:fedora.im>
17:06:30
**fesco #3385** (https://pagure.io/fesco/issue/3385):**Change: Java25 And No More System JDK**
<@sgallagh:fedora.im>
17:06:39
!hi
<@zodbot:fedora.im>
17:06:40
Stephen Gallagher (sgallagh) - he / him / his
<@sgallagh:fedora.im>
17:07:04
Sorry, it's been a day...
<@salimma:fedora.im>
17:07:05
we have all 9! perfect timing
<@salimma:fedora.im>
17:07:25
Fabio Valentini: you have some reservations about the process, IIRC
<@decathorpe:fedora.im>
17:07:54
I'm mostly confused about some of the choices made for the CP since they don't make sense to me
<@decathorpe:fedora.im>
17:09:49
or rather, confused in general might describe my reaction better
<@conan_kudo:matrix.org>
17:10:20
same
<@conan_kudo:matrix.org>
17:10:26
I don't really understand the Change
<@nirik:matrix.scrye.com>
17:10:56
it's basically moving from a model where there is a system jdk, to where there's versioned ones and a particular one is the default right?
<@decathorpe:fedora.im>
17:11:05
- and v21 will go away at a later point in time
<@decathorpe:fedora.im>
17:11:05
- there will be no system JDK
<@decathorpe:fedora.im>
17:11:05
- but effectively v25 will be *"the one"* because it will be the one installed by default, and every Java package will be patched to build with it?
<@decathorpe:fedora.im>
17:11:05
- so later, there will be no "system JDK" but there will only be *one*?
<@salimma:fedora.im>
17:11:22
right... I think I also got a bit confused. I *think* the idea is to manage the versions similar to how Python is managed, but I'm not sure
<@zbyszek:fedora.im>
17:11:22
??
<@zbyszek:fedora.im>
17:11:22
> JDK21 will remain in Fedoras, until it is newest JDK providing java in any live Fedora (f44)
<@zbyszek:fedora.im>
17:11:22
<@decathorpe:fedora.im>
17:11:26
I don't particularly like that this implies patching *all* Java packages every time
<@jvanek:matrix.org>
17:11:55
I think you misunderstood Fabio
<@conan_kudo:matrix.org>
17:12:00
also note that another aspect that was mentioned is maintaining this -latest openjdk package, rather than switching that to versioned too
<@salimma:fedora.im>
17:12:12
right. when I tried bootstrapping some Java packages in EPEL10 I noticed some packages are still broken on anything newer than Java 11, which is worrying
<@jvanek:matrix.org>
17:12:14
The -latest will never provide java
<@conan_kudo:matrix.org>
17:12:15
I'd rather the unversioned jdk package go away completely
<@jvanek:matrix.org>
17:12:19
as it is STS only
<@conan_kudo:matrix.org>
17:12:35
yes, but I don't see a reason to not just have versioned packages for both "STS" and "LTS" ones
<@salimma:fedora.im>
17:12:45
so I think there's terminology confusion for anyone not familiar with the ecosystem
<@conan_kudo:matrix.org>
17:12:49
we already have policy in place to make it easy to introduce them
<@salimma:fedora.im>
17:12:54
STS is not defined in the proposal :)
<@jvanek:matrix.org>
17:13:03
Te rolling latest is slightly out of the cope oft his meeting
<@salimma:fedora.im>
17:13:03
short term support?
<@jvanek:matrix.org>
17:13:11
yy,
<@brakoose:matrix.org>
17:13:13
Little do I know..but replacing java with kotlin will be great. Since they are both based on jdk.
<@jvanek:matrix.org>
17:13:21
:D
<@fale:fale.io>
17:13:26
JDK 21 is supported until September 2031, so it is ok for it to be the default for a while, imho
<@jvanek:matrix.org>
17:13:38
they both still needs jvm and java runtime
<@jvanek:matrix.org>
17:13:56
Supportede by who and where?
<@jvanek:matrix.org>
17:14:10
do not confuse Oralce paid support x live of OpenJDK
<@salimma:fedora.im>
17:14:16
yeah. also - not relevant here, but iirc you need Gradle to build Kotlin and that's a nightmare of a dependency cycle
<@fale:fale.io>
17:14:39
uh right! Oracle support Oracle JDK untile 2031, RH supports OpenJDK (for its own products) until 2029
<@jvanek:matrix.org>
17:15:05
Anybody of them can start to eork behind the firewall any moment
<@jvanek:matrix.org>
17:15:23
Public OpenJDK is hard to declare
<@salimma:fedora.im>
17:15:52
to make things less confusing - can we do it like Python. the source package is always versioned? or is that already planned
<@fale:fale.io>
17:15:52
so who dis the F44 decision was made?
<@fale:fale.io>
17:16:06
s/who/how/
<@salimma:fedora.im>
17:16:06
whether the binary package should be versioned or not or we can have '-latest' is then another question
<@jvanek:matrix.org>
17:16:35
To much entwined threads
<@jvanek:matrix.org>
17:16:43
What f44 decision?
<@jvanek:matrix.org>
17:16:59
And what is issue with two jdks providne "java" - which I belive is what confuses Fabio
<@jvanek:matrix.org>
17:17:16
And what is issue with two jdks providing "java" - which I belive is what confuses Fabio
<@salimma:fedora.im>
17:17:31
on the subject of Fabios there are also two of them :)
<@conan_kudo:matrix.org>
17:17:37
multiple jdks providing java isn't a big deal
<@conan_kudo:matrix.org>
17:17:44
since there's alternatives in place to make it work
<@jvanek:matrix.org>
17:17:49
that was the last showstopper IIUC
<@decathorpe:fedora.im>
17:18:06
this is not confusing me
<@decathorpe:fedora.im>
17:19:11
from the change proposal (just above [this anchor](https://fedoraproject.org/wiki/Changes/Java25AndNoMoreSystemJdk#Schedule):
<@decathorpe:fedora.im>
17:19:11
<@decathorpe:fedora.im>
17:19:11
> Any packager can decide to keep any (available) JDK version, as long as it is in system, so if there is major breakage for them, they have more than a year to fix it. In scope of https://fedoraproject.org/wiki/Changes/ThirdPartyLegacyJdks, the JDK will be removed from Fedora and once it lost status of "system jdk" in any live Fedora (so JDK21 should no longer be available in f45, and will be deprecated in live f43+f44 in them moment of removal). For those who wish to continue using OpenJDK 21 after that time, the recommendation is to use the third party Temurin JDK 21"
<@decathorpe:fedora.im>
17:19:20
I do not understand this ***at all***
<@jvanek:matrix.org>
17:19:50
Canyou place more concrete question please?
<@nirik:matrix.scrye.com>
17:19:59
Thats what java is used for BuildRequires no?
<@zbyszek:fedora.im>
17:20:02
I think the meta-problem with this CP is that it is very hard to understand.
<@zbyszek:fedora.im>
17:20:16
Already at the level of the language.
<@jvanek:matrix.org>
17:20:18
BuildRequires - yes
<@jvanek:matrix.org>
17:20:22
and javac
<@sgallagh:fedora.im>
17:20:24
I think there's also a language barrier happening here.
<@decathorpe:fedora.im>
17:20:41
> Fedora will stay on top with fresh technologies by having newest JDK available immediately and having new system JDK as soon as possible.
<@decathorpe:fedora.im>
17:20:41
so there *will be* a system JDK? I thought the CP was about getting rid of that?
<@decathorpe:fedora.im>
17:20:41
further down there is
<@decathorpe:fedora.im>
17:20:41
<@decathorpe:fedora.im>
17:20:41
<@zbyszek:fedora.im>
17:20:55
And if _we_ have a problem with understanding the text, the average user of Fedora will do so too.
<@sgallagh:fedora.im>
17:21:07
jvanek: I think it might be easier if you described the exact lifecycle plan of JDK 21 and 25?
<@decathorpe:fedora.im>
17:21:17
it is really hard for me to follow what is happening (not even mentioning *why*)
<@fale:fale.io>
17:21:28
should this CP be splitted in multiple CPs?
<@jvanek:matrix.org>
17:21:42
hm. Two people ahd read it and send notes. Both of them are english natives, nut unluckily both are deep in java world
<@sgallagh:fedora.im>
17:21:55
Like, in what capacity they will each exist in F43, F44, F45 and F46
<@conan_kudo:matrix.org>
17:21:57
yeah I'm honestly confused too
<@jvanek:matrix.org>
17:22:12
Wre are adding jdk25, that si celar I think
<@jvanek:matrix.org>
17:22:29
And we are aligning the java/ajvac requires with rest of the world
<@sgallagh:fedora.im>
17:22:33
jvanek: Not sufficient.
<@jvanek:matrix.org>
17:22:54
where the huga benefit is, that there will be year to adjust packages
<@jvanek:matrix.org>
17:23:02
not jsut the moment of mass rebuild
<@sgallagh:fedora.im>
17:23:10
jvanek: What will each package be called, where will it live on the filesystem, will it own /usr/lib/java, will it own /usr/lib/java[version]?
<@sgallagh:fedora.im>
17:23:29
For each Fedora release I listed above
<@jvanek:matrix.org>
17:24:06
usr/lib/java[version] as always. usr/lib/java is symlink
<@sgallagh:fedora.im>
17:24:11
Because there's a huge knowledge gap somewhere. Your change proposal makes a number of leaps that are probably obvious to someone who has been working on it closely, but are alien to us
<@jvanek:matrix.org>
17:24:16
always and everywhere
<@decathorpe:fedora.im>
17:24:41
I think I am getting the picture now. You will be making changes to allow building Java packages with *either* openjdk 21 *or* openjdk 25, making it possible to transition over the course over a longer period of time?
<@jvanek:matrix.org>
17:24:52
in f41+f42 jdk25 will be there, but not providing java. jdk21 will remain here and will be providing java. f43+f44 both jdk25 will be there and both will provide java. oin f45 only jdk25 will remain and will of course provide java
<@decathorpe:fedora.im>
17:24:56
vs. instead of switching the default system-wide at a single point in time
<@salimma:fedora.im>
17:25:05
are there anyone else in the Java team that could help explain the proposal? Maybe make them co-owner of the change?
<@nirik:matrix.scrye.com>
17:25:09
FWIW, I think I get it... and was +1 in the discussion in the end. But if lots of others don't, more clarity is good for sure.
<@jvanek:matrix.org>
17:25:15
Stephen Gallagher: rigth, thats why I'm here today
<@jvanek:matrix.org>
17:25:31
Fabio Valentini: right
<@sgallagh:fedora.im>
17:25:39
On F43 and F44 will they `Conflicts:`, use `alternatives`, or something else?
<@conan_kudo:matrix.org>
17:25:55
that makes a ton of sense
<@conan_kudo:matrix.org>
17:26:01
it's essentially moving to the Python model
<@nirik:matrix.scrye.com>
17:26:07
yes
<@jvanek:matrix.org>
17:26:10
no conflict in alternatives
<@salimma:fedora.im>
17:26:21
that's my guess but I must admit it's not clear from the initial reading
<@sgallagh:fedora.im>
17:26:24
jvanek: Thank you for doing so. I apologize if I come across as patronizing, but we're obviously missing details, so I'm trying to be as pedantic as possible so we will understand.
<@decathorpe:fedora.im>
17:27:02
I feel like the Change proposal is *much too complicated* compared to the simple change that it's going to make 😅
<@nirik:matrix.scrye.com>
17:27:06
alternatives... java has used it forever.
<@jvanek:matrix.org>
17:27:26
they will coexsits jsut find
<@jvanek:matrix.org>
17:27:33
the provides here,is reallythe virtual provides in specfile. nothing more
<@jvanek:matrix.org>
17:27:36
they will coexsits jsut fine
<@jvanek:matrix.org>
17:27:38
..nothing more.. but important:)
<@sgallagh:fedora.im>
17:28:13
jvanek: Will Java packages be expected to use `/usr/bin/java` or `/usr/bin/java[version]` in Fedora?
<@sgallagh:fedora.im>
17:28:29
Let me rephrase that, actually.
<@jvanek:matrix.org>
17:28:31
that is up to them
<@jvanek:matrix.org>
17:28:45
If they use java-packages-tools proeprly, it will be handled for them
<@sgallagh:fedora.im>
17:28:55
jvanek: Will we expect them to use `/usr/bin/java` unless they are unable to run on the latest available?
<@jvanek:matrix.org>
17:29:01
as java-packages-tools are bale to determine build jdk, and set up runtime jdk acordingly
<@decathorpe:fedora.im>
17:29:21
so the generated wrapper script will use the correct one?
<@jvanek:matrix.org>
17:29:22
one fix on latest *LTS*
<@sgallagh:fedora.im>
17:29:25
Ah, now that is critical information I was missing
<@jvanek:matrix.org>
17:29:41
yes
<@jvanek:matrix.org>
17:29:43
it will
<@sgallagh:fedora.im>
17:29:46
jvanek: Sorry, I can't parse that
<@jvanek:matrix.org>
17:29:59
/usr/bin/java is only for user space
<@jvanek:matrix.org>
17:30:05
as it can be affected by alternatives
<@nirik:matrix.scrye.com>
17:30:10
so as a java packager you can just 'go with the flow' and keep on 21 for now, until it switches to 25, if it breaks too badly, you can manually opt to go back to 21
<@nirik:matrix.scrye.com>
17:30:16
(if I understand correctly)
<@jvanek:matrix.org>
17:30:37
launchers should go dierectly to /usr/lib/jvm/java[version] and that is what generated launcher is doing
<@jvanek:matrix.org>
17:30:46
nirik: correct
<@decathorpe:fedora.im>
17:30:51
Stephen Gallagher: for context, Java *"programs"* are actually generated shell wrapper scripts that call some JRE with a JAR file
<@jvanek:matrix.org>
17:30:52
only oposite
<@jvanek:matrix.org>
17:31:02
you will stay ion jdk21, unless you willingly move to 25
<@jvanek:matrix.org>
17:31:11
Which I'm going to try during mass ebuild
<@sgallagh:fedora.im>
17:31:14
jvanek: Do packagers need to manually move their packages forward to Java 25 or is that automatic and they'd need to manually pin to the older version?
<@nirik:matrix.scrye.com>
17:31:21
for now, but at some point 21 goes away and 25 becomes the default right?
<@jvanek:matrix.org>
17:31:22
waht will be possible to move to jdk25 wihtout issues, I will move to it
<@jvanek:matrix.org>
17:31:43
Stephen Gallagher: you mentioned latest. Taht sugests java-latest-opendjk, and that is out of scope
<@decathorpe:fedora.im>
17:31:55
Yeah, this is one thing that I don't understand. Why make the move to openjdk 25 opt-in instead of opt-out?
<@sgallagh:fedora.im>
17:32:34
Yeah, I'm with Fabio Valentini here. That's the opposite of how we do it for every other language ecosystem in Fedora
<@sgallagh:fedora.im>
17:32:43
And kind of disagrees with one of our Four Foundations.
<@salimma:fedora.im>
17:32:52
if this is just a new package and is opt-in, what does it mean to be 'preferred'
<@jvanek:matrix.org>
17:33:06
only latest LTS is in scope
<@decathorpe:fedora.im>
17:33:13
maybe higher priority in alternatives?
<@jvanek:matrix.org>
17:33:20
once java-latest-openjdk willmvoe to jdk26, it will nto affect anything
<@sgallagh:fedora.im>
17:33:39
jvanek: When I said "latest", I meant "latest packaged"
<@sgallagh:fedora.im>
17:33:50
But if that's also packaged... where does that fit in?
<@nirik:matrix.scrye.com>
17:34:25
it's the java-devel provides right? ie, packages that buildrequire that now get 21.
<@sgallagh:fedora.im>
17:34:34
(and why would 26 not be packaged the same way as the other Java releases?)
<@jvanek:matrix.org>
17:34:48
java-latest-openjdk donot provide java, nor java-devel and is rolling up, and is harbouring STS. is important for any future development
<@jvanek:matrix.org>
17:35:01
java-xyz-openjd harbours LTSs
<@decathorpe:fedora.im>
17:35:09
as I understand it, unversioned java-devel would go away, and packages would be changed to explicitly require 21 or 25?
<@jvanek:matrix.org>
17:35:17
jdk26 will be STS and will dvell jsut in java-latest-openjdk
<@conan_kudo:matrix.org>
17:35:22
what's the reason for not just doing STSes the same way as LTSes and retiring them when they EOL
<@nirik:matrix.scrye.com>
17:35:22
no, the provide just moves I thought?
<@sgallagh:fedora.im>
17:35:28
Fabio Valentini: That's not what was just described, no
<@jvanek:matrix.org>
17:35:29
next LTS will be jdk29
<@decathorpe:fedora.im>
17:35:45
oh well, then I am confused :)
<@decathorpe:fedora.im>
17:36:22
(I gather only LTS versions can be used to build packages, but non-LTS versions can be used for "user space"?)
<@sgallagh:fedora.im>
17:36:39
Yeah, that's what it sounds like they want to achieve.
<@nirik:matrix.scrye.com>
17:36:40
STS is... shorter than a fedora cycle I think?
<@jvanek:matrix.org>
17:36:41
Fabio Valentini: that is correct way to do it. However the versionless provides will stay there. And may be used, only you will then get whatever JDK dnf willc ome with
<@jvanek:matrix.org>
17:37:00
STS is short term support
<@decathorpe:fedora.im>
17:37:00
what does that mean?
<@sgallagh:fedora.im>
17:37:02
I think that splitting the packaging paradigm is a painfully overengineered way to do it, though
<@jvanek:matrix.org>
17:37:13
jdk 8,11,17,21,25 and 29 are LTS
<@decathorpe:fedora.im>
17:37:14
"whatever dnf gives you" is not well defined :D
<@salimma:fedora.im>
17:37:17
- when will such defaults change
<@salimma:fedora.im>
17:37:17
- will any of either source or binary packages provide a default (e.g. '-latest')?
<@salimma:fedora.im>
17:37:17
- what are source packages called, will they be versioned?
<@salimma:fedora.im>
17:37:17
- what is their lifecycle / when will they go EOL
<@salimma:fedora.im>
17:37:17
- when will new Java releases be introduced
<@salimma:fedora.im>
17:37:17
<@salimma:fedora.im>
17:37:17
so I feel like the proposal should clarify a few things
<@jvanek:matrix.org>
17:37:42
Stephen Gallagher: Splitting the packaging paradigma? Now I can not parse, sorry
<@salimma:fedora.im>
17:37:45
explicit is better than implicit, since I doubt any of us would remember this in a few days
<@zbyszek:fedora.im>
17:37:53
Hmm, I don't think it makes sense to try to "decipher" the proposal during a meeting. The proposal **must** be written in a way where the basic idea can be understood by a casual Fedora user.
<@decathorpe:fedora.im>
17:38:06
jvanek: I think Stephen meant that LTS and non-LTS versions are packaged differently
<@sgallagh:fedora.im>
17:38:15
jvanek: You have one style of packaging to be used for LTS releases and an entirely different one (with different rules) for the STS releases.
<@zbyszek:fedora.im>
17:38:22
Even the title doesn't make sense gramatically.
<@salimma:fedora.im>
17:38:24
so... we probably should timebox this. maybe provide recommendations of what we want to see clarified, and let's say move on to the next topic at 45 past?
<@jvanek:matrix.org>
17:38:40
I think more confusion would casue to include new JDK every 6months, and then dropping it again
<@sgallagh:fedora.im>
17:38:46
I have a proposal
<@jvanek:matrix.org>
17:38:52
in additionlifecycle of STS is shorter then single fedora life span
<@jvanek:matrix.org>
17:39:14
zbyszek: There is quite a wrong vocabulary on our sides
<@salimma:fedora.im>
17:39:15
I'm ok with just packaging LTSes then. but it probably should be stated
<@jvanek:matrix.org>
17:39:28
I'm tro much deep in java, and tyu are to much deep in fedora:(
<@salimma:fedora.im>
17:39:36
maybe provide STSes in COPR for people who want to test
<@pboy:fedora.im>
17:39:38
Folks, from a java deloper pov you are discussing on the wrong foot. The problem is not latest or the name, the issue is we will have too few Java versions. E.G. wildfly is release for a specific LTS version, with Fedora will not provide after a while.
<@jvanek:matrix.org>
17:39:53
I would regret to miss the STS from fedora
<@jvanek:matrix.org>
17:39:56
its usefull apckage
<@sgallagh:fedora.im>
17:40:01
Proposal: We request that the Java SIG writes up a new version of the Java Packaging Guidelines for Fedora that can be understood by other Java package maintainers. This must be provided alongside the Change Proposal to be considered for acceptance.
<@jvanek:matrix.org>
17:40:06
and esential for testing next LTS
<@nirik:matrix.scrye.com>
17:40:33
are the current guidelines not clear enough? I haven't looked at them...
<@jvanek:matrix.org>
17:40:39
but it is offtopic jhrere a bi
<@sgallagh:fedora.im>
17:40:41
We want to make sure the Change can be understood by the people who need to understand it, so asking them to write the Packaging Guidelines seems like an appropriate method.
<@decathorpe:fedora.im>
17:41:00
Stephen Gallagher: the Java SIG has been defunct for a decade, I don't think that makes sense
<@sgallagh:fedora.im>
17:41:08
...
<@zbyszek:fedora.im>
17:41:10
Stephen Gallagher: sorry, but seems completely beside the point.
<@zbyszek:fedora.im>
17:41:37
Sorry, let me rephrase that.
<@jvanek:matrix.org>
17:41:54
Stephen Gallagher: The java apckaging guidelines have to be adjsute
<@jvanek:matrix.org>
17:42:07
it is even written in the proposal, with suggested steps
<@zbyszek:fedora.im>
17:42:11
It'd be possibly good to update the Java Packaging Guidelines, but we cannot require that here. It's not directly related to the CP.
<@jvanek:matrix.org>
17:42:32
They hae to be updated
<@decathorpe:fedora.im>
17:42:45
Anyway, I think it is clear that the proposal as-is is too confusing, and several things need to be clarified. I propose that we go back to discussion.fp.o with specific suggestions, and once the wiki page is updated, we can discuss it in meeting again?
<@salimma:fedora.im>
17:42:53
jvanek: I'm confused. you yourself said it will be confusing to include a new JDK then drop it again, but you said you want to package STSes
<@zbyszek:fedora.im>
17:42:56
"adjusted" != "new version of"
<@jvanek:matrix.org>
17:43:01
And will be updated (if we agree on the several "java" provides
<@sgallagh:fedora.im>
17:43:14
zbyszek: I think that might just be a poor choice of words on my part.
<@nirik:matrix.scrye.com>
17:43:16
the STSes are always in the 'latest' package
<@jvanek:matrix.org>
17:43:32
Michel Lind UTC-6: I pack them in tha rolling java-latest-opendjk
<@sgallagh:fedora.im>
17:43:35
"new version of" shouldn't be read to imply a green-field rewrite.
<@jvanek:matrix.org>
17:43:49
it is jsut adjsuting. The change is smaller then it seems
<@sgallagh:fedora.im>
17:43:56
But whatever modifications are needed to adapt to this significant change in packaging
<@nirik:matrix.scrye.com>
17:43:58
what LTS version does that require? out of curiosity?
<@sgallagh:fedora.im>
17:44:24
jvanek: I think that seeing the delta of the packaging guidelines would go a LONG way to making this more understandable.
<@pboy:fedora.im>
17:44:42
nirik: was release fo 17, now certified for 21 imho.
<@salimma:fedora.im>
17:44:49
latest always providing the current STS, and if necessary getting a waiver for incompatible updates, and each LTS being packaged with the version explicitly stated, seems reasonable
<@salimma:fedora.im>
17:45:05
though I'd question what happens when latest == LTS every four releases
<@jvanek:matrix.org>
17:45:06
But wildfly is no longer packed, right? It was impossible to mantain/ And if you run 3rd party jar/war, you may happily install 3rd party jdk
<@salimma:fedora.im>
17:45:38
so yeah if the packaging guidelines is currently out of date, perhaps updating them as part of this CP *is* in scope - or should be done anyway
<@nirik:matrix.scrye.com>
17:45:39
Michel Lind UTC-6: I think the idea is to test it some in fedora, then create a new LTS named package and the STS one moves on
<@jvanek:matrix.org>
17:46:03
right.
<@nirik:matrix.scrye.com>
17:46:05
The change includes that the guidelines need updating. ;)
<@salimma:fedora.im>
17:46:09
let's start wrapping up this topic though... any explicit recommendations we can make for next steps?
<@conan_kudo:matrix.org>
17:46:13
the thing is -latest packages are not allowed in fedora :)
<@conan_kudo:matrix.org>
17:46:20
because of the mess that was docker-latest
<@nirik:matrix.scrye.com>
17:46:26
"the https://docs.fedoraproject.org/en-US/packaging-guidelines/Java/ will need adjusting. "
<@jvanek:matrix.org>
17:46:30
Only this time I have to include java-25-opendjk very early, because the timing of jdk25 release and f43 forking is unhappy\
<@decathorpe:fedora.im>
17:46:50
I suggested something here
<@jvanek:matrix.org>
17:47:03
and package review would take away to long time
<@pboy:fedora.im>
17:47:54
I would be willing to help to rewrite/revise both and bring my docs experience in
<@salimma:fedora.im>
17:48:17
yeah, let's vote on Fabio's suggestion and take this to Discourse
<@nirik:matrix.scrye.com>
17:48:18
If it's a forward compat package it could use the exception to review...
<@salimma:fedora.im>
17:48:21
I'm +1
<@conan_kudo:matrix.org>
17:48:33
you can ping me or any of the rest of us if you need package reviews to go quickly
<@conan_kudo:matrix.org>
17:48:35
we can do the thing
<@conan_kudo:matrix.org>
17:48:45
but reviews are not required for versioned forks of packages
<@conan_kudo:matrix.org>
17:49:09
unless they are substantially different, or you may want to do it anyway to have someone check over it
<@jvanek:matrix.org>
17:49:11
hmhmh. I consider java-xyz-as fresh pacage
<@jvanek:matrix.org>
17:49:21
and during the review always some rotten things are found
<@conan_kudo:matrix.org>
17:49:26
that's fine, you can always ask one of us to do a package review
<@jvanek:matrix.org>
17:49:31
hmhmh. I consider java-xyz-openjdk as fresh pacage
<@emma:rory.gay>
17:49:32
how does this affect applications like minecraft?
<@conan_kudo:matrix.org>
17:49:35
the dotnet people do the same thing and I always help them out on this
<@emma:rory.gay>
17:50:09
im guessing one would need to build JRE8 from source?
<@decathorpe:fedora.im>
17:50:09
Emma [it/its]: minecraft uses /usr/bin/java AFAIK so it will get the one that's set up by the alternatives system
<@nirik:matrix.scrye.com>
17:50:15
it should keep working, if it doesn't it can optionally build against the old one...
<@jvanek:matrix.org>
17:50:18
Emma [it/its]: it should not affect it.
<@jvanek:matrix.org>
17:50:30
especially if it is on java on path
<@emma:rory.gay>
17:50:37
that hasnt been the case since 1.6 (2013?)
<@nirik:matrix.scrye.com>
17:50:46
anyhow, thanks Peter Boy for offering to work on docs/helping clarify change. ;)
<@emma:rory.gay>
17:51:10
i was more asking in the sense of older versions of the game depending on ie. java 8 (java 17, 21 for newer versions)
<@salimma:fedora.im>
17:51:11
ok, people finish what you're typing then we'll move on
<@jvanek:matrix.org>
17:51:20
Peter Boy: can yuo pelase reach me on jvanek@redhat.com ? And we can walk ti paragraph by paragraph
<@zodbot:fedora.im>
17:51:47
kevin gave a cookie to pboy. They now have 34 cookies, 1 of which were obtained in the Fedora 42 release cycle
<@pboy:fedora.im>
17:51:49
jvanek: I'll do
<@jvanek:matrix.org>
17:51:50
Emma [it/its]: older jdks are provided by temurin project. and temurin repos are part of fedora.
<@emma:rory.gay>
17:52:13
ah i see, then i suppose my concern is invalid
<@jvanek:matrix.org>
17:52:14
Michel Lind UTC-6: for OC.. I'm not sure what the outcome here is for me
<@jvanek:matrix.org>
17:52:20
I hope I clarified all I could
<@jvanek:matrix.org>
17:52:30
but obviously some demands are here
<@jvanek:matrix.org>
17:52:39
So those will be summed up on @devel list?
<@salimma:fedora.im>
17:52:46
jvanek: I think we'll ask clarifying questions in Discourse, then revisit this once the proposal is updated with them
<@salimma:fedora.im>
17:53:03
and with all the questions asked/answered here
<@salimma:fedora.im>
17:53:24
I'll try to summarize as best as I can after this meeting
<@jvanek:matrix.org>
17:54:01
ok. So right now ballis in your hands, and some time later, I will recieve the lsit of requiremnts
<@jvanek:matrix.org>
17:54:03
oook?
<@jvanek:matrix.org>
17:54:23
ok. So right now ballis in your hands, and some time later, I will receive the list of requirements
<@decathorpe:fedora.im>
17:54:32
👍️
<@zbyszek:fedora.im>
17:55:11
jvanek: thank you for being here and patiently answering
<@nirik:matrix.scrye.com>
17:55:24
yes, thanks jvanek!
<@jvanek:matrix.org>
17:55:24
Unable to answer all :((
<@salimma:fedora.im>
17:55:28
I'll let nirik, pboy and Stephen finish then let's move on
<@jvanek:matrix.org>
17:55:28
thanx for invitation!
<@jvanek:matrix.org>
17:55:33
Thanx for fedora
<@pboy:fedora.im>
17:55:34
jvanek: We should skip throught the zocbozprotocol and work on each question
<@jvanek:matrix.org>
17:55:36
and good speed!
<@jvanek:matrix.org>
17:55:41
//me off and go
<@salimma:fedora.im>
17:55:42
and yes, thanks jvanek for your time and helping answer these questions
<@zodbot:fedora.im>
17:55:56
No Fedora Accounts users have the @jvanek:matrix.org Matrix Account defined
<@zodbot:fedora.im>
17:56:01
No Fedora Accounts users have the @jvanek:matrix.org Matrix Account defined
<@jvanek:matrix.org>
17:56:06
they were unsufficient :(
<@jvanek:matrix.org>
17:56:34
zocbozprotocol? I will learn somethign it seems :)
<@jvanek:matrix.org>
17:56:36
bye!
<@nirik:matrix.scrye.com>
17:56:37
sidenote: for cookies, add you matrix id to your fedora account system account.
<@salimma:fedora.im>
17:57:06
!topic #3391 Exemption for deprecated Aspell dependency
<@salimma:fedora.im>
17:57:14
!fesco 3391
<@zodbot:fedora.im>
17:57:15
● **Opened:** a week ago by fed500
<@zodbot:fedora.im>
17:57:15
● **Last Updated:** 15 hours ago
<@zodbot:fedora.im>
17:57:15
● **Assignee:** Not Assigned
<@zodbot:fedora.im>
17:57:15
<@zodbot:fedora.im>
17:57:15
**fesco #3391** (https://pagure.io/fesco/issue/3391):**Exemption for depcrecated Aspell dependency**
<@zbyszek:fedora.im>
17:57:32
… so your matrix id is in the fedora account system account matrix.
<@zbyszek:fedora.im>
17:57:39
(for symmetry)
<@salimma:fedora.im>
17:57:41
(sorry, Stephen Gallagher but idk if you're really typing in this chat or it's somewhere else, since Matrix is weird)
<@sgallagh:fedora.im>
17:58:01
No, I wasn't typing. Probably a network glitch
<@salimma:fedora.im>
17:58:06
so for this topic - we agreed the exemption for this particular package, but agree this should just be documented as a general case
<@nirik:matrix.scrye.com>
17:58:07
I don't think you can depend on typing notifications too heavily. ;)
<@salimma:fedora.im>
17:58:49
so - take this back to FPC, I guess? maybe we should just discuss which documents need to be updated now, while we're all here
<@decathorpe:fedora.im>
17:59:00
yeah, I basically volunteered to make a PR add language that says what I proposed to the Package Review Guidelines
<@salimma:fedora.im>
17:59:05
so far I see review guidelines, and there is the deprecation guideline as well
<@emma:rory.gay>
17:59:13
i dont even send them ;)
<@decathorpe:fedora.im>
17:59:35
the exceptions to the package review process are documented on the Package Review Guidelines page, so I think it makes sense to add them there
<@salimma:fedora.im>
17:59:40
yeah, I mean for people who do send them I do try to wait a bit, but it has to be a bounded wait
<@salimma:fedora.im>
17:59:59
ah, so just there and not on the deprecation page? that's fine, one less place to check
<@emma:rory.gay>
18:00:02
they expire for a reason :) (or at least, are supposed to, if not, your client does it wrong)
<@nirik:matrix.scrye.com>
18:00:04
right. it's a per server option, many don't send it... and clients seem to sometimes think you are typing into your matrix client when you are doing something else, or typing into some other channel, etc.
<@zbyszek:fedora.im>
18:00:12
Should we just bounce this to the FPC?
<@nirik:matrix.scrye.com>
18:00:22
(but lets take that discussion elsewhere)
<@salimma:fedora.im>
18:00:41
yeah, my plan is to bounce, just do a last discussion here in case we need to clarify anything. the issue will remain open until it's documented I guess
<@nirik:matrix.scrye.com>
18:00:58
This page? https://docs.fedoraproject.org/en-US/fesco/Package_review_policy/
<@salimma:fedora.im>
18:01:04
last time FPC just bounced this to us to check if we think it's a good idea, but FPC seems better positioned to review PRs to guidelines
<@decathorpe:fedora.im>
18:01:04
!action decathorpe to file a PR with FPC
<@salimma:fedora.im>
18:01:12
ok, we can move on then
<@zodbot:fedora.im>
18:01:18
salimma gave a cookie to decathorpe. They now have 123 cookies, 2 of which were obtained in the Fedora 42 release cycle
<@nirik:matrix.scrye.com>
18:01:19
note that this page is in the fesco space. ;)
<@salimma:fedora.im>
18:01:22
thanks Fabio
<@salimma:fedora.im>
18:01:34
last scheduled item
<@salimma:fedora.im>
18:01:39
!topic #3388 DOCUMENTATION_URL is 404
<@salimma:fedora.im>
18:01:42
!fesco 3388
<@zodbot:fedora.im>
18:01:43
● **Assignee:** Not Assigned
<@zodbot:fedora.im>
18:01:43
**fesco #3388** (https://pagure.io/fesco/issue/3388):**DOCUMENTATION_URL is 404**
<@zodbot:fedora.im>
18:01:43
<@zodbot:fedora.im>
18:01:43
● **Opened:** 2 weeks ago by sgallagh
<@zodbot:fedora.im>
18:01:43
● **Last Updated:** 15 hours ago
<@nirik:matrix.scrye.com>
18:01:49
before we get into the weeds discussing this....
<@salimma:fedora.im>
18:02:04
people voted on the suggested idea, before Stephen Gallagher pointed out that the URL proposed is not actually live at beta time
<@nirik:matrix.scrye.com>
18:02:16
does this need to be a fesco ticket? isn't this just a bug on fedora-release for interested parties to discuss and fix there?
<@salimma:fedora.im>
18:02:21
and Conan Kudo suggested this should be added to the SOP for launching a beta
<@salimma:fedora.im>
18:02:42
agreed on tabling Matrix grouching to another time and place. Or I'll never stop :)
<@fale:fale.io>
18:02:53
I think Conan Kudo's approach would be the best one imho :)
<@decathorpe:fedora.im>
18:03:12
why are the package review policies in both fesco and fpc space? ...
<@salimma:fedora.im>
18:03:50
the original issue, probably not a fesco ticket, but if we want to provide them a URL to use and that URL is not live during the beta... who owns the SOP for what to do for the beta? us, or release engineering
<@nirik:matrix.scrye.com>
18:04:01
the freedesktop thing says: "DOCUMENTATION_URL= should refer to the main documentation page for this operating system"
<@nirik:matrix.scrye.com>
18:04:24
so, why not direct them just to the top level?
<@salimma:fedora.im>
18:04:24
I think this one? https://docs.fedoraproject.org/en-US/packaging-guidelines/ReviewGuidelines/
<@salimma:fedora.im>
18:04:45
and apologies, if we still need to discuss the previous topic I can change the topic back
<@nirik:matrix.scrye.com>
18:04:46
(and ... how many people ever look at or follow this link? )
<@fale:fale.io>
18:05:07
nirik: I'm worried that some users might be confused on their version and therefore read the "wrong" one (or give up)
<@sgallagh:fedora.im>
18:05:18
nirik: Traditionally, FESCo is the set of "interested parties" for `fedora-release`
<@sgallagh:fedora.im>
18:05:44
nirik: That link is populated into numerous other apps
<@conan_kudo:matrix.org>
18:05:46
it's used in desktops to create a working link for the OS
<@nirik:matrix.scrye.com>
18:05:54
seems pretty micromanagery to me, but if you all want to... sure.
<@sgallagh:fedora.im>
18:05:57
It's meant to be the canonical location for other tools to use
<@decathorpe:fedora.im>
18:06:31
Well ... what would it take to make "this is a working and up-to-date link" a beta criterion instead?
<@decathorpe:fedora.im>
18:06:49
that would avoid micro-managery things :)
<@sgallagh:fedora.im>
18:06:59
Fabio Valentini: We (FESCo) can declare it thus, and QA will block on it.
<@salimma:fedora.im>
18:07:06
should not be too hard right? that page can start with a disclaimer that it's not finalized yet
<@sgallagh:fedora.im>
18:07:33
But I'm also fine with nirik 's suggestion of "just use the top-level link"
<@nirik:matrix.scrye.com>
18:07:35
why don't we point it to https://docs.fedoraproject.org/en-US/fedora/latest/ ?
<@sgallagh:fedora.im>
18:07:55
nirik: Because even on release day, that was still pointing to the previous release
<@decathorpe:fedora.im>
18:08:00
well ... that will be wrong?
<@nirik:matrix.scrye.com>
18:08:08
it's not wrong?
<@decathorpe:fedora.im>
18:08:08
there isn't even an f42 version of those docs yet
<@sgallagh:fedora.im>
18:08:14
And it might be wrong if you're running an older release.
<@zbyszek:fedora.im>
18:08:14
I think we want it to point the the versioned docs, if at all possible.
<@salimma:fedora.im>
18:08:25
yeah, and also if you install when f41 is the latest, then latest points to f42, you'll be reading the wrong document
<@nirik:matrix.scrye.com>
18:08:26
what does a f42 version look like?
<@sgallagh:fedora.im>
18:08:28
Fabio Valentini: F42's version is available now
<@nirik:matrix.scrye.com>
18:08:29
I am confused
<@decathorpe:fedora.im>
18:08:35
<del>there isn't even an f42 version of those docs yet</del> oh wait, there is, the UI is just very confusing
<@nirik:matrix.scrye.com>
18:08:38
it's pointing to the landing page
<@sgallagh:fedora.im>
18:08:40
But it wasn't on release dat
<@salimma:fedora.im>
18:08:45
at some point the links will need to be pinned to point to specific releases, it's just a question of when. only having to do it once seems useful
<@sgallagh:fedora.im>
18:08:45
But it wasn't on release day
<@nirik:matrix.scrye.com>
18:08:59
but we shouldn't point to release notes because those are not... for this?
<@salimma:fedora.im>
18:09:04
right now f42 redirects to latest, but at least f42 points to something that's not 404
<@nirik:matrix.scrye.com>
18:09:25
The only release specific things I see are release notes
<@nirik:matrix.scrye.com>
18:09:27
this is not that
<@sgallagh:fedora.im>
18:09:29
Proposal: Just set it to https://docs.fedoraproject.org/
<@nirik:matrix.scrye.com>
18:09:39
"should refer to the main documentation page for this operating system"
<@nirik:matrix.scrye.com>
18:09:43
how is the latest page not that?
<@nirik:matrix.scrye.com>
18:09:56
the top page is not the operating system...
<@nirik:matrix.scrye.com>
18:10:01
it's all the fedora project stuff
<@decathorpe:fedora.im>
18:10:02
"latest" means f42 on a f40 system too, which would be ... weird
<@salimma:fedora.im>
18:10:18
exactly
<@nirik:matrix.scrye.com>
18:10:24
it means "the landing page" you can choose the f41 release notes... they are right there...
<@zbyszek:fedora.im>
18:10:38
Stephen Gallagher: -1
<@salimma:fedora.im>
18:10:57
we can decide to either always point to an explicit per-release page, but then it can't be latest, and we need to make that part of the beta SOP
<@nirik:matrix.scrye.com>
18:10:58
it does mean at beta you won't see a 'f43' one there... but it will have at least upgrade docs and other things.
<@zbyszek:fedora.im>
18:11:02
I think Neal's original proposal is good.
<@salimma:fedora.im>
18:11:03
or we can decide to point to just the docs landing page
<@sgallagh:fedora.im>
18:11:03
nirik: You didn't link the landing page, you linked https://docs.fedoraproject.org/en-US/fedora/latest/
<@sgallagh:fedora.im>
18:11:11
Which is currently the "f42" page
<@nirik:matrix.scrye.com>
18:11:36
thats the latest page for the fedora linux user docs
<@nirik:matrix.scrye.com>
18:11:43
how is it a f42 page?
<@zbyszek:fedora.im>
18:11:51
Also, we the redirect from `/f42` to `/latest` needs to be dropped.
<@nirik:matrix.scrye.com>
18:11:53
oh, I see, there's a pull down
<@sgallagh:fedora.im>
18:11:54
My first choice would be for https://docs.fedoraproject.org/en-US/fedora/fNN to always be correct by Beta and to use that. My second choice is to just use https://docs.fedoraproject.org/ like I proposed above.
<@zbyszek:fedora.im>
18:12:00
It makes no sense.
<@nirik:matrix.scrye.com>
18:12:08
but I don't think that's actually used/matters/this landing page changes too much
<@pboy:fedora.im>
18:12:19
Folks, our "docs team" is basically me and Petr. And we are overbooked and coulnd kdeep pace.
<@sgallagh:fedora.im>
18:12:51
Peter Boy: And we are deeply grateful to you for what you do!
<@nirik:matrix.scrye.com>
18:12:55
yeah. So, IMHO this should just become a unversioned page...
<@sgallagh:fedora.im>
18:13:09
nirik: Hence my second-choice proposal above.
<@nirik:matrix.scrye.com>
18:13:20
The only thing that changes is the release notes I think.
<@salimma:fedora.im>
18:13:25
if we don't want to overload the docs people, we can just make 'fXX' point to latest first at beta time. just make sure it's not 404
<@pboy:fedora.im>
18:13:27
nirik: agreed! But it's a löot of work.
<@salimma:fedora.im>
18:13:36
if we don't want to require that then yes, suggest people use the landing page instead
<@nirik:matrix.scrye.com>
18:13:39
Stephen Gallagher: but thats not the page for the "operating system"
<@sgallagh:fedora.im>
18:13:57
nirik: True enough.
<@nirik:matrix.scrye.com>
18:14:56
proposal: I'll talk to Petr and see if we can make this page unversioned/latest always, and add a historical release notes section if people want
<@sgallagh:fedora.im>
18:14:59
Proposal: Set DOCUMENTATION_URL to https://docs.fedoraproject.org/en-US/fedora/fNN/ (and if it doesn't exist prior to GA, that's okay)
<@nirik:matrix.scrye.com>
18:15:18
so at branching?
<@sgallagh:fedora.im>
18:15:20
(Is there a way to have it autodetect the language?)
<@pboy:fedora.im>
18:15:25
Stephen Gallagher: the page for the docs about operating system Fedora is docs.fedoraperoject.org, imho
<@nirik:matrix.scrye.com>
18:15:25
yet another thing to change. :(
<@sgallagh:fedora.im>
18:15:30
nirik: Yes, at branching
<@sgallagh:fedora.im>
18:15:47
nirik: No, I'll write a patch for the specfile. It will be automatic.
<@pboy:fedora.im>
18:16:00
The language is autodetected (or should be, if it works)
<@nirik:matrix.scrye.com>
18:16:21
sure, I think we could do better, but I won't stand in the way.
<@sgallagh:fedora.im>
18:16:25
Peter Boy: Even if the link says "en-us"?
<@salimma:fedora.im>
18:16:28
I'm ok with Stephen's proposal
<@salimma:fedora.im>
18:16:59
so it's OK to have a 404 - wait, even at GA? or are we saying it has to exist at GA, but prior to that it's fine
<@pboy:fedora.im>
18:17:10
Hm, in that case not, i think. links should not contain language items
<@sgallagh:fedora.im>
18:17:23
Michel Lind UTC-6: It should exist at GA, but I'm not proposing we BLOCK for that.
<@salimma:fedora.im>
18:17:34
got it. so a SHOULD not a MUST
<@sgallagh:fedora.im>
18:17:44
Peter Boy: I don't know of a link that will lead to that location with the language auto-detected. If you do, please suggest it.
<@salimma:fedora.im>
18:17:57
should we vote? and if it is approved Stephen has an action item to do it since he volunteered
<@nirik:matrix.scrye.com>
18:18:11
yeah, using en-US there is not great.
<@sgallagh:fedora.im>
18:18:18
Yes, I'll make the appropriate change, once agreed.
<@nirik:matrix.scrye.com>
18:18:29
but I suppose it can be changed after.
<@pboy:fedora.im>
18:18:38
Stephen Gallagher: docs.fedoraproject.org is meant to autodetect. I didn't test for a long thime thouhg
<@nirik:matrix.scrye.com>
18:18:39
and I am sceptical that very many apps use this. ;)
<@decathorpe:fedora.im>
18:19:02
Michel Lind UTC-6: this proposal? I'm +1
<@sgallagh:fedora.im>
18:19:08
Peter Boy: The root docs.fp.o works fine. I was wondering if we have another similar redirect for the lower level.
<@sgallagh:fedora.im>
18:19:13
But for now, let's not worry about it.
<@sgallagh:fedora.im>
18:19:19
There's a language select at the top of the page
<@pboy:fedora.im>
18:19:47
Fabio Valentini: your proposal is dangerous!
<@decathorpe:fedora.im>
18:20:03
hey, it's not mine 😆
<@salimma:fedora.im>
18:20:05
Peter Boy: what proposal? we're voting on Stephen's
<@salimma:fedora.im>
18:20:14
what's dangerous about it?
<@pboy:fedora.im>
18:20:22
Any change on the docs pages should be evaluated first!
<@decathorpe:fedora.im>
18:20:30
isn't this just codifying the status quo?
<@pboy:fedora.im>
18:20:39
And probably you wait for my Flock talks :-)
<@salimma:fedora.im>
18:21:04
if Peter Boy has a better idea I'm sure we're all ears
<@conan_kudo:matrix.org>
18:21:08
... wat?
<@nirik:matrix.scrye.com>
18:21:19
it's changing it... the current link is broken completely/doesn't exist
<@salimma:fedora.im>
18:21:20
otherwise we probably should not waste too much time on this
<@nirik:matrix.scrye.com>
18:21:40
we have already been at it for over 20min. ;)
<@sgallagh:fedora.im>
18:21:45
Fabio Valentini: It's slightly different. It's the general Fedora guide, not the Sysadmin guide.
<@decathorpe:fedora.im>
18:21:48
I meant, DOCUMENTATION_URL being potentially HTTP 404 wouldn't change
<@salimma:fedora.im>
18:21:50
right, the current link is completely broken, so pointing to something that works is better. let's not bikeshed it right now
<@pboy:fedora.im>
18:21:50
I didnÄ't follox this completely. I can make a proposal / comment on this, but need some toime (days, not years)
<@conan_kudo:matrix.org>
18:22:10
we're just setting it to `https://docs.fedoraproject.org/en-US/fedora/f<releasever>/`
<@conan_kudo:matrix.org>
18:22:18
which is the fedora linux documentation place
<@pboy:fedora.im>
18:22:24
The docs system is a bit fragile
<@salimma:fedora.im>
18:22:28
yeah. let's vote, and if it does not pass, then let's discuss alterantives
<@salimma:fedora.im>
18:22:44
(let's discuss alternatives in the ticket, I mean, since we only have one solid proposal now)
<@sgallagh:fedora.im>
18:22:48
I'll restate the proposal for clarity, if desired.
<@salimma:fedora.im>
18:22:55
I see a Fabio plus 1. I'm +1 too
<@nirik:matrix.scrye.com>
18:22:57
FWIW, the rawhide one works
<@salimma:fedora.im>
18:23:19
yes, please, plus the proposal is up there in the scrollback anyway
<@pboy:fedora.im>
18:23:31
The proposed link makes any translations useless. And the work of our lacalization SIG
<@decathorpe:fedora.im>
18:23:47
Peter Boy: is it worse than what's currently there?
<@salimma:fedora.im>
18:24:09
seems like the right thing to do is to make that link auto-pick the right language then?
<@sgallagh:fedora.im>
18:24:10
Proposal: Set DOCUMENTATION_URL to https://docs.fedoraproject.org/en-US/fedora/fNN/ This link SHOULD be available at the launch of GA, but if it's delayed, it does not block release.
<@salimma:fedora.im>
18:24:18
oh, the en-US part
<@zbyszek:fedora.im>
18:24:26
Peter Boy: if you can come up with a better URL, we can delay the vote until next week certainly.
<@pboy:fedora.im>
18:24:28
Fabio Valentini: it makes the mistake permanent. That's the problem.
<@dcantrell:fedora.im>
18:24:28
+1 to the proposal
<@sgallagh:fedora.im>
18:24:36
Peter Boy: I'd like to fix that, and we can work on that in parallel. But this is no worse than the current state.
<@sgallagh:fedora.im>
18:24:50
Which is also pointing at an `en-US` link.
<@decathorpe:fedora.im>
18:24:53
it's not like we can't change this *again* once there's a better solution ...
<@pboy:fedora.im>
18:25:04
I can come up, but I have th check it
<@zbyszek:fedora.im>
18:25:24
OK, so +1 to Stephen Gallagher
<@sgallagh:fedora.im>
18:25:27
One more revised proposal
<@salimma:fedora.im>
18:25:30
we can make it explicit that "the hardcoded language path will be updated once a better link is available"
<@salimma:fedora.im>
18:25:47
to clarify that we're also not too happy with hardcoding it but there are no alternatives now
<@sgallagh:fedora.im>
18:25:50
Proposal: Set DOCUMENTATION_URL to https://docs.fedoraproject.org/en-US/fedora/fNN/ This link SHOULD be available at the launch of GA, but if it's delayed, it does not block release. This link will be updated to the same comment, but using a language-aware link once that becomes available.
<@salimma:fedora.im>
18:25:57
yes
<@fale:fale.io>
18:26:05
+1 on this last version
<@salimma:fedora.im>
18:26:08
ok - let's timebox it so we have up to half an hour for open floor
<@zbyszek:fedora.im>
18:26:09
"comment"?
<@sgallagh:fedora.im>
18:26:12
err
<@decathorpe:fedora.im>
18:26:14
s/somment/content?
<@salimma:fedora.im>
18:26:19
fesco members, vote please
<@sgallagh:fedora.im>
18:26:19
Proposal: Set DOCUMENTATION\_URL to https://docs.fedoraproject.org/en-US/fedora/fNN/ This link SHOULD be available at the launch of GA, but if it's delayed, it does not block release. This link will be updated to the same document, but using a language-aware link once that becomes available.
<@nirik:matrix.scrye.com>
18:26:26
weak +1 I guess...
<@salimma:fedora.im>
18:26:28
document - that looks fine
<@salimma:fedora.im>
18:26:29
+1
<@zbyszek:fedora.im>
18:26:29
And it's missing a dot after the URL.
<@zbyszek:fedora.im>
18:26:32
But +1
<@decathorpe:fedora.im>
18:26:35
+1
<@sgallagh:fedora.im>
18:26:35
It was supposed to be "content", but I went to "document" instead
<@dcantrell:fedora.im>
18:26:36
+1
<@humaton:fedora.im>
18:26:39
+1
<@sgallagh:fedora.im>
18:27:09
zbyszek: I left that out intentionally so it didn't get treated as part of the link by clients
<@zbyszek:fedora.im>
18:27:29
Smart clients will omit the dot.
<@zbyszek:fedora.im>
18:27:41
https://fedoraproject.org/.
<@salimma:fedora.im>
18:28:28
I'm assuming Stephen Gallagher is a +1 on his own proposal
<@salimma:fedora.im>
18:28:41
Conan Kudo: ? and I'm missing someone
<@zbyszek:fedora.im>
18:28:43
Michel Lind UTC-6: that's implicit in the voting rules.
<@sgallagh:fedora.im>
18:28:47
Michel Lind UTC-6: That's always implicit, yes
<@salimma:fedora.im>
18:29:03
oh zbyszek. so yes, we're +9 by my count
<@salimma:fedora.im>
18:29:24
!agreed APPROVED (+9, 0, -0) Set DOCUMENTATION_URL to https://docs.fedoraproject.org/en-US/fedora/fNN/ This link SHOULD be available at the launch of GA, but if it's delayed, it does not block release. This link will be updated to the same document, but using a language-aware link once that becomes available
<@nirik:matrix.scrye.com>
18:29:29
~30min per bug. We are on 🔥!
<@zbyszek:fedora.im>
18:29:42
Actually, thinking about this more, "that's implied by the voting rules" is a better way to phrase this.
<@salimma:fedora.im>
18:29:43
!action sgallagh will update the DOCUMENTATION_URL in os-release
<@conan_kudo:matrix.org>
18:29:47
+1
<@salimma:fedora.im>
18:29:49
!topic Open Floor
<@salimma:fedora.im>
18:30:01
sorry this is dragging on folks
<@zbyszek:fedora.im>
18:30:13
https://pagure.io/fesco/fesco-docs/pull-request/100, it had +4
<@zodbot:fedora.im>
18:30:14
decathorpe gave a cookie to salimma. They now have 59 cookies, 1 of which were obtained in the Fedora 42 release cycle
<@zbyszek:fedora.im>
18:30:18
Can I merge it?
<@zodbot:fedora.im>
18:30:23
sgallagh gave a cookie to salimma. They now have 60 cookies, 2 of which were obtained in the Fedora 42 release cycle
<@salimma:fedora.im>
18:30:44
lgtm
<@salimma:fedora.im>
18:31:19
-7 is the dual of +7 so if +7 and no downvote means go for it, -7 and no +1 should also mean can it
<@salimma:fedora.im>
18:31:36
oh there is a typo in the first section
<@salimma:fedora.im>
18:31:47
"Otherwise, if there is at least one '-1' vote - that should be 'at least one +1 vote'
<@salimma:fedora.im>
18:32:17
wait, never mind, I read that wrong
<@zbyszek:fedora.im>
18:33:38
OK, I'll merge it soon.
<@salimma:fedora.im>
18:34:42
do we do action items for this too? we do I guess
<@salimma:fedora.im>
18:34:56
!action zbyszek will merge https://pagure.io/fesco/fesco-docs/pull-request/100
<@zbyszek:fedora.im>
18:34:58
We also have https://pagure.io/fesco/fesco-docs/pull-request/101.
<@zbyszek:fedora.im>
18:35:08
Stephen Gallagher: can we close it?
<@sgallagh:fedora.im>
18:35:42
zbyszek: I'm outvoted, so I suppose so?
<@salimma:fedora.im>
18:37:31
!action https://pagure.io/fesco/fesco-docs/pull-request/101 will be closed as it lacks support
<@salimma:fedora.im>
18:38:13
so related to docs review - we're waiting for a suggested wording to add it to the meeting process, IIRC
<@decathorpe:fedora.im>
18:38:25
that's my TODO item yeah
<@salimma:fedora.im>
18:38:40
which is why it's not in the agenda for today yet. apologies. we do have one PR open https://pagure.io/fesco/fesco-docs/pull-request/94
<@decathorpe:fedora.im>
18:38:45
it's on *"The Pile" ™️*
<@salimma:fedora.im>
18:38:50
seems like nobody cares to weigh in on devel and on Discourse?
<@decathorpe:fedora.im>
18:39:20
I can open a discussion post if that is preferred?
<@decathorpe:fedora.im>
18:39:32
though I would think this an uncontroversial (and net-positive) change
<@salimma:fedora.im>
18:39:37
Fabio Valentini: fun fact: piles means something else https://www.medicalnewstoday.com/articles/239454
<@salimma:fedora.im>
18:39:59
I don't think that's necessary - we are the people voting on that anyway
<@nirik:matrix.scrye.com>
18:40:14
There was some discussion on list I thought?
<@salimma:fedora.im>
18:40:15
oh, for the lightweight stalled request? it's already on discourse
<@salimma:fedora.im>
18:40:31
!link https://discussion.fedoraproject.org/t/rfc-lightweight-stalled-request-process/148733
<@salimma:fedora.im>
18:40:35
let me check the list again
<@decathorpe:fedora.im>
18:40:54
oh, sorry, I missed that link in the scrollback, I thought it was for fesco meeting process change.
<@salimma:fedora.im>
18:40:57
ah right, the problem with posting this on tax day
<@decathorpe:fedora.im>
18:41:08
my trusty dictionary tells me I'm correct too: https://dict.leo.org/german-english/pile
<@salimma:fedora.im>
18:41:18
yeah for the fesco meeting process change I think keeping it in the ticket is fine, since we're the ones voting
<@salimma:fedora.im>
18:41:33
though I don't think anyone will stop you filing a discussion item :)
<@decathorpe:fedora.im>
18:42:47
IMO the discussion on the devel list was mostly useless ...
<@salimma:fedora.im>
18:42:54
ok, caught up on the questions. for Simon's suggestion - the problem with lack of comaintainership is that, of course, unless you're a provenpackager you can't build it
<@salimma:fedora.im>
18:43:45
yeah, the discussion is not helping that much, I can respond to them to close the loop though. do we want to wait on that another week, or should we just say, eh, the issues don't seem consequential?
<@nirik:matrix.scrye.com>
18:43:52
well, you can via a pr no?
<@salimma:fedora.im>
18:44:17
landing the PR does not trigger a build, right? Simon was talking about the case where FESCo intervened to approve the PR
<@decathorpe:fedora.im>
18:44:22
for rawhide, that would work, for stable releases, no
<@salimma:fedora.im>
18:44:28
surely we don't want to be on the hook to land and build it
<@nirik:matrix.scrye.com>
18:44:45
ah right, because you can't make the update
<@salimma:fedora.im>
18:44:46
rawhide does not auto build either right
<@salimma:fedora.im>
18:44:53
only if you have packit set up to build it?
<@decathorpe:fedora.im>
18:45:06
in general I also don't like situations where the person who makes a package update won't be notified for new bugs on that package
<@nirik:matrix.scrye.com>
18:45:15
it doesn't, but anyone can build a thing from scm...
<@nirik:matrix.scrye.com>
18:45:21
(as long as it's not been built before)
<@decathorpe:fedora.im>
18:45:27
Michel Lind UTC-6: the issue is not koji, but bodhi
<@salimma:fedora.im>
18:45:42
I can reply to devel pointing out that python SIG for example require people to commit to comaintaining if they want to build a change (instead of just submitting a PR and waiting for the maintainer to do something)
<@salimma:fedora.im>
18:46:14
ah.. right. so anyone can build for any branch, but you can't put up an update
<@salimma:fedora.im>
18:46:22
so it's not a problem for rawhide
<@nirik:matrix.scrye.com>
18:46:23
yep
<@salimma:fedora.im>
18:46:51
I wish there's a way so you have to explicitly opt in to use provenpackager access, because I'm a bit rusty about how it works with a normal account :P
<@decathorpe:fedora.im>
18:47:27
I try to remind myself every time I use them. just to make sure I don't do anything that oversteps
<@salimma:fedora.im>
18:47:27
ok... if there is no other suggested change from the discussion, I'll rebase the PR and wait a few days for comments then maybe land it ... say this Friday
<@salimma:fedora.im>
18:47:49
and I'll follow up with Michael and Simon
<@decathorpe:fedora.im>
18:47:50
assuming we vote in favor, no? :)
<@salimma:fedora.im>
18:48:10
right. of course
<@nirik:matrix.scrye.com>
18:48:35
I have more comments on the list to do, so I would prefer waiting
<@nirik:matrix.scrye.com>
18:48:41
but if everyone else wants to go, ok.
<@salimma:fedora.im>
18:49:05
do please send your comments - I'll hold off on rebasing until then so we just need to do one pass
<@salimma:fedora.im>
18:49:19
we've waited long enough, what's another week
<@salimma:fedora.im>
18:49:30
anyone else has a topic? we have 10 mins left
<@salimma:fedora.im>
18:50:33
sorry :(
<@salimma:fedora.im>
18:50:37
haven't had lunch either myself
<@salimma:fedora.im>
18:50:47
ok -- if nobody else has anything, let's close this then
<@salimma:fedora.im>
18:51:11
!action salimma will follow up on the mailing list questions and then rebase the PR and address any additional comments later this week
<@nirik:matrix.scrye.com>
18:51:15
oh, who is next chair?
<@salimma:fedora.im>
18:51:20
oh right
<@salimma:fedora.im>
18:51:24
!topic next chair
<@fale:fale.io>
18:52:11
if there is no volunteer, I can volunteer
<@salimma:fedora.im>
18:53:02
no other taker?
<@salimma:fedora.im>
18:53:30
!action fale will chair next meeting
<@salimma:fedora.im>
18:53:45
thanks all for coming and for all the discussions
<@salimma:fedora.im>
18:53:51
!endmeeting