- User Since
- Apr 15 2015, 5:07 PM (161 w, 6 d)
The purpose of this patch was to simply hide the standalone launcher functionality a little
Widget-based launchers do have jump lists, fwiw
And personally I consider pinning to be activity dependant way more complicated than having both options to pin and have a launcher
That's disappointing to hear and a bit frustrating considering that it never came up in the Bugzilla ticket I filed months 3 ago
I veto this patch, sorry. These are different features and they're both intended to be there. Task Manager pinning is not a replacement for permanent launchers.
Fri, May 18
Thu, May 17
I've cleaned this up/out now. About half or so got reassigned to either systemsettings:kcm_formats or frameworks-sonnet. Some got closed as FIXED and WORKSFORME. There's a bunch left. 2-3 have a point (e.g. the the language names), the rest is weird stuff.
Wed, May 16
I sometimes try IRC but that does not work very well, because a developer needs to be online at the same time you are online
Sorry, accident! The code in the Keyboard KCM is ... interesting and literally breaks my system trying to build it (it needs a two-digit amount of GB of memory to build thanks to a boost-based keyboard layout preview generator it even changes the LLVM template recursion depth for), so I usually disable it so I don't have to hard reset my PC and lose data.
Tue, May 15
It's worth noting we have this in a non-formalized way. The Getting Involved pages have always listed mentors you can contact for 1on1 questions (I've responsed to mails due to this for many years).
Mon, May 14
Move removal into the C++ model code to avoid another model reset
and speed up the UX.
Add back Move up/down actions.
Sun, May 13
Don't mutate source model while building list of langs to add to selection.
Remove width from delegate in sheet again.
This looks good, but I think there might be additional copies of these values in (a) the color KCM code and (b) the kdeglobals code in kdelibs4support, did you have a chance to look into those?
No worries, also my fault for not spotting it in review. :)
Sat, May 12
Fri, May 11
I currently don't have the time to make the showCount stuff per-browser to handle multi-browser usage. I think it might also not be worth it. People who use multiple browsers are probably power user enough to hear about and install the extension themselves.
Unload also on early abort.
- The SNI doesn't disappear when the browser is closed.
Always open the extension URL in the browser that was actually launched,
instead of relying on it being the same as the default browser.
Thu, May 10
@mart Do you think I should push this before we work on the DND stuff?
Wed, May 9
There's some conceptual problems with this remaining:
Clean up auto usage
Fix logic error.
The code simplification in the above changes also gets rid of the possible C++20 usage.
Add .desktop file names for Firefox Nightly and Google Chrome Beta.
- Revise naming and make consistent
- Fix copyrights
- Don't use other people's branding
- Simplify (BrowserInfo is gone)
At the same time, I'm not sure using the browser icon straight in an SNI is appropriate either. Upstreams might not appreciate their branding being slapped onto something they didn't originate. I think we should use internet-web-browser here (the same icon that's used for the default browser KCM in System Settings).
I'm not a fan of the combined icon idea. On face of it it's a neat "go the extra mile" kind of thing, on the other hand it creates follow-up which I'm certain will not be done in a timely manner, or even forgotten about: Updating those assets whenever the upstream branding is revised. It means initially the visual is slightly nicer, but in the future it will look outdated and crap. It's not a good idea to bring this about for such a minor thing, I think it's better to keep it simple and agree with Vlad here.
Tue, May 8
Set explicit width for Material style
- Right-align 'Add languages ...' button
- Clean up cruft from a pre-ListView revision, thanks Marco
I'm left to interpret what David meant with "I don't like the module names".
The code looks fine to me. I was questioning if config.writeEntry("shownCount", 100); should be MAX_SHOW_COUNT + 1 instead for a moment, but this is safer in case it gets bumped later.
I'll commandeer, Kai asked me to take this due to being really sick.
Right-aligned button: Can do, sure.
Looks good to me, thanks for working on this.
Mon, May 7
- Fix settings saving
- String tweak
Sat, May 5
Fri, Apr 27
Thu, Apr 26
I also just realized we seem to have zero keyboard nav support ...
- Port to ScrollVieWKCM
- Add list label at the top and show conditionally
- Actions are now 'Make default' and 'Remove' with appropriate tooltips
- Topmost language has a "(Default)" behind it
- Code cleanups
I can't quite follow the discussion so far (also because my Konsole already seems to paste URLs on drop). Are you saying the current default is to show a Dolphin/FV-like drop menu with actions? I think that's nice for consistency, no? Then again apps like Konversation don't and Konsole is often used for chats, too.
Same as before, we're just refactoring to clean some code :)
Wed, Apr 25
Model looks good! :)
In general: This code would be cleaner if it wasn't using QStandardItemModel but just a QAbstractListModel subclass. Then stuff like the "selected theme index" could use either a role or a QItemSelectionModel, and the role enum wouldn't need to live outside of the model. :)
Expand documentation as requested by Marco.
It looks good otherwise and I really want to accept this so I can progress on the Language KCM, but this magic value thing gives me a headache. It's going to break with other QStyles with a different value. We need some sort of fix for that ...
Fix the animating prop
Unlikely, Gnome's shell runs in-process with the compositor, they don't need any IPC for the same use case.
Does the magic pixel size number for the padding get scaled correctly on hidpi?
Tue, Apr 24
Does it still have the issues mentioned above? Because then not really ...
Mon, Apr 23
Apr 22 2018
Apr 21 2018
In principle I'm inclined to side with Martin here. At this point, I no longer want the extra burden/distraction of working on newly-introduced X11-related bugs. It's frustrating to spend manhours on code you know will be obsolete, and that delays getting work done to make it obsolete. I still have many scaling problems on Wayland that I think should be higher priority.
Adding 'Default' is pretty easy, no problem.