[spam comment removed by sysadmin]
- Queries
- All Stories
- Search
- Advanced Search
Advanced Search
Aug 29 2019
Aug 19 2019
Hey I noticed the torrent didn't generate anymore:
Aug 18 2019
Aug 14 2019
+1 for the placeholder version; that looks nicer IMO.
The UI for the latest implementation is very good. I have only one suggestion. When no valid USB disks are mounted, instead of displaying a blank combobox with no entries, instead hide the combobox and show a centered label that says, "Please plug in a USB disk"
In T11388#195423, @kives wrote:Great work. Really impressed by the turnaround time on this and the fixes to KTorrent, those are awesome.
I noticed when I click on the .torrent file from files.kde.org sometimes it redirects to a 404, I assume some mirrors haven't updated yet?
Great work. Really impressed by the turnaround time on this and the fixes to KTorrent, those are awesome.
Aug 13 2019
torrents are now generated when isos are built and then get their mirror lists updated every 3 hours. thanks again for the generator code @kives!
Aug 12 2019
Wow really great work I just got a chance to test the KTorrent changes they work awesome!
This is in review in debian
I'm looking at the KTorrent bug and trying to fix it. I've got a working setup right now to build and debug libKF5Torrent.
In T11388#195179, @kives wrote:in light of ktorrent being bust, should we maybe run a seed? it would kinda defeat the purpose of load balancing through mirrors though
In my opinion it moves the logic to the clients to do the load balancing. It's unlikely that a client will actually download from all of the mirrors at once but rather will connect to them and use the ones that are the fastest. From the perspective of most torrent clients, the mirrors are just another peer, and the normal bandwidth optimization strategies are used. It should still distribute the load to many mirrors.
what a time to be alive
updated https://github.com/pangea-project/pangea-tooling/commit/c4abd6389c009f296f8a57137b0548769e8600e6 will be on next images
in light of ktorrent being bust, should we maybe run a seed? it would kinda defeat the purpose of load balancing through mirrors though
@jriddell thoughts?
Aug 11 2019
Aug 10 2019
+1, I think this would be nice as well. Needs VDG input and work, probably.
Aug 9 2019
Aug 5 2019
thanks much!
In T11318#194216, @jriddell wrote:What does this error mean @kossebau ?
'-- No such target KF5PulseAudioQt_QCH when calling ecm_install_qch_export().
- No target exported for KF5PulseAudioQt_QCH. '
We are getting this on pulseaudio-qt itself.
What does this error mean @kossebau ?
'-- No such target KF5PulseAudioQt_QCH when calling ecm_install_qch_export().
- No target exported for KF5PulseAudioQt_QCH.
'
Aug 2 2019
Putting in normal -dev files is also an option, sure.
How the packages are split is generally done by the Debian packagers and we just copy them.
Small feedback/info from my discussions with opensuse packagers (who sadly still need to finish execution/implementation), which might help on designing solutions:
[11:29] <sitter> I do rather think that this also needs a fix in ECM, no? [11:29] <JonathanRiddell-> sitter: ecm does complain about it, what more should it do? [11:30] <sitter> the complaint should be a feature_info I would think [11:30] <sitter> 02:22:21 -- The following features have been enabled: [11:30] <sitter> 02:22:21 [11:30] <sitter> 02:22:21 * QCH, API documentation in QCH format (for e.g. Qt Assistant, Qt Creator & KDevelop) [11:30] <sitter> ther should also be a feature QCH-Qt5 linkage or somesuch [11:30] <JonathanRiddell-> mm right [11:31] <DavidRedondo-M> frinring [11:31] <DavidRedondo-M> one could also change the code not to skip silently, but make noise or even fail [11:31] <DavidRedondo-M> at the time when I wokred on that I did not want to be too aggressive here, to not have people run into lots of issues before they could see something... [11:31] <DavidRedondo-M> yesterday [11:31] <DavidRedondo-M> I would agree that a warning would be good [11:33] --> dharman (~dharman@93-42-171-52.ip87.fastwebnet.it) has joined this channel. [11:34] <sitter> JonathanRiddell-: in fact it seems our tooling nowadays ignores Warnings altogether, supposedly because the cmake files are often in a horrible state and raise warnings when they should raise author_warning [11:34] <sitter> that is something we could revisit [11:34] <sitter> obviously if we do we also need to deal with that very problem that often times people use WARNING when they should have used AUTHOR_WARNING ^^
Aug 1 2019
Jul 31 2019
Prototype in action now. Needs some additional QA tech to finds possible overlaps and warn so we can actually blacklist them.
Should be fixed with https://invent.kde.org/kde/ktrip/commit/8a81ecf99a1391663da72ff97af6541c2460e44d
package looks lovely. thanks.
Jul 29 2019
Jul 26 2019
Another debian-pm package that could be useful for Neon: https://gitlab.com/debian-pm/libs/qml-box2d.
(Probably also another candidate for upstreaming to debian)
Jul 24 2019
Jul 22 2019
My comment is that the final page once it has succeeded a write to USB should have buttons "Write Another USB" or "Close". Current I expect there to be buttons there and the green tick box looks like a button.
The UI for the latest implementation is very good. I have only one suggestion. When no valid USB disks are mounted, instead of displaying a blank combobox with no entries, instead hide the combobox and show a centered label that says, "Please plug in a USB disk"
Jul 21 2019
You can use https://gitlab.com/debian-pm/apps/ktrip
Jul 17 2019
we have the data now but I'm unsure how to proceed
[18:22] <sitter> JonathanRiddell-: I'm not quite sure what to do with the html we could pop it into the metadata domain and iframe it on the website, but then I am also not sure where to put it on the website https://build.neon.kde.org/view/mgmt/job/mgmt_version_list_bionic_user/2/artifact/versions.html [18:22] <sitter> also the squishing problem I mentioned earlier screws this plan, so that needs fixing IMO
it didn't like comments in cmake-ignore
yeah seems sorted
Apps / Frameworks / Plasma seems the main use case, although individual extragear apps would also be interesting
I do wonder if we even should care about extragear. That may just complicate things?
This is already solved.
watcher.rb checks which projects.kde.org product a build belongs to and if it is one of the bundles it sends a warning mail and refuses to auto-bump.
kde4 translations are dead from a kde-i18n team POV if I followed some phab tasks correctly. phonon is now using kf5 translations, and I would presume kdeedu too?
Jul 16 2019
kicon cache is a huge drag (10MiB per kde snap). in point of fact I suspect all of .cache is being fairly awkward but I also do not think this is all too solvable. there is no way to ensure that the cache format of snap X is the same as Y, so they couldn't share the cache.
no reply, so inactive, we won't package it
Closing this task for now. We've missed the mark by a bit, and from the neon side things should be more or less in place right now. So, this is mostly blocked on actually having the plasma wayland UX ready for primetime. Or perhaps more importantly, should the requirements change we need to reshuffle the stack anyway, so having a perpetually stuck task serves no purpose.
Jul 15 2019
Jul 10 2019
Yup.
Jul 9 2019
Do existing installs get migrated, or is this only for new installs?
All ISOs respun and openqa updated as well. Transition complete.
Jul 8 2019
Hi Sam!
Reassigning to Neon.
Jul 1 2019
Jun 28 2019
I just find the phab board a bit easier to keep track of but I'll keep the two in sync if ever I set up a phab task for a bug report
Works great for me! Maybe the CMS hadn't updated yet.