We could activate -Wunreachable-code so the compiler tells you where break isn't needed.
- Queries
- All Stories
- Search
- Advanced Search
Advanced Search
Sep 21 2018
Sep 20 2018
A small additional change to the plugin json, and a request for a bit of help.
C is not a subset of C++ (let me point to Mr. Stroustrup who should know it :)
... shared-mime-info has this questionable inheritage for C++ and Objective-C from C.
Updated as requested, waiting for final a-ok.
Sep 19 2018
Possibly because Qt on macOS uses the macOS mimettype database
Not really: https://repology.org/metapackage/astyle/versions
Normally I'd agree, but here we're dealing with really a tiny amount of additional code and a project which isn't exactly being kept up-to-date by distributions either (on Linux I'd get 2.04 from the system, on Mac 2.05 from MacPorts). Bundling the few files involved means that at least everyone gets the astyle plugin and with the same version.
Code looks fine to me, modulo all the comments made, and works where I tested it. Please also update the summary text and title of the patch.
BTW, this change is still marked as approved, is it?
The libastyle upgrade was split off as D15605, as requested.
(Fortunately the 2 changesets were orthogonal.)
It's just that I use the 5.3 branch.
Yeah, I was expecting this, though I hoped it could just be separate commit, not a separate review.
I added ObjC previews as requested, and doing so I noticed I had to make a few more changes, justifying a ticket name change. I did mention the change was going to be a bit more elaborate than one might have thought, didn't I? :)
Sep 18 2018
seems to still lead us here into a situation where only "smallItems" exists?
I would suspect some rounding errors.
No, I haven't pushed a fix, but I have one which seems to work just fine. Feel free to clean it up as you see fit (I see an accidental indentation change) and push it yourself.
On it, but the changes go a bit further than I (and you?) might have guessed ([private] method renaming e.g.). I hope to have something by the end of day, but if you think it'd end up having to be split into a separate commit I could just as well hear that now ;)
Sep 15 2018
Thanks. Learned about ObjC++ this way... the world is full of strange things :)
astye change split off as D15532
Sep 5 2018
it should be connected destroyed to containerEmpty too.
I say that because in one of my konsole branches I removed that and it indeed worked.
I think these do different things; containerEmpty is called when a container (Konsole window, presumably) closed its last tab but doesn't unregister the container. I'd guess that containerEmpty will be called first, and then containerDestroyed when the container is being disposed off (because it's empty).
There may not be an explicit current need to handle these separately, but I don't see why one should NOT do that.
No need to specify Qt::DirectConnection explicitly.
Aug 22 2018
Food for thought: what ‘s the KF5 frameworks version in that CentOS?
Aug 21 2018
Why does that build not use the current hunspell? The old api has been obsolete for quite a while IIRC and hunspell is not a core/base system library
Aug 14 2018
Downloading already didn't work for me since quite a while (the operation never completed) but what does this mean for highlighting files I write myself (or someone's mom does)?
Aug 13 2018
Use a current instead of an obsolete platform token.
I'll test the patch ASAP, but I think there's a simple guideline to observe here (apologies if it's as much an open door as I think it is):
Aug 6 2018
OK, if no one else cares :)
I'd still like that message not to appear if I just dismissed DrKonqi myself.
I don't see why. We try to start drkonqi even the very first time.
The crash recursion counter is increased when a SEGV happens inside the crash handler itself.
Aug 5 2018
I'm abandoning this for now because my assumption was wrong even though my patch had the effect I intended. I'll reopen if/when I have a real fix, because there are still 2 issues here:
Actually, it's even worse than that: I didn't double-check my assumptions about the use of DrKonqi's return/exit code. Looking at the code again there is actually no way that information is even obtained. The startProcess function simply starts DrKonqi and then waits for it to exit if so instructed; the only information obtained about the crashreporter is its PID.
> if DrKonqi gives an exit code that means it was started It could have crashed though...
Aug 4 2018
But there's a catch, on my system I get some plasma popup on which I have to click "Report Bug" in order to get the usual drkonqi dialog. Do you get that too?
When I click "Close" in drkonqi (there's no Cancel button), crashtest resumes and exits, without any error message from KCrash.
When I click "Close" in drkonqi (there's no Cancel button), crashtest resumes and exits, without any error message from KCrash.
Aug 1 2018
That was already mentioned :)
Jul 31 2018
I really don't see what the big fuss is all about.
I have no other objections than the ones above, as long as there's a fallback to NOT using libcanberra which can be enforced without having to patch the code (= via CMAKE_DISABLE_FIND_PACKAGE_FOO or a dedicated option).
Jul 30 2018
I don't really care about your artificially created problems. You *have* libcanberra but want to use it for one project and not the other. Sorry, but no.
(Oh btw it's not like I didn't try, I originally ported all of this to QtMultimedia but `QAudioEffect` which is for low-latency sound effects only supports WAV sounds and `QMediaPlayer` had significant overhead and initialization times as well which is why I eventually ended up using canberra)
Jul 29 2018
Jul 28 2018
[I didn't get to send this yesterday]
Jul 27 2018
That's reinventing the wheel to a point.
Jul 26 2018
However I rather like the idea of improving QTMultiMedia's sound handling features rather than working around the lack of them by using a different library.
Anyway, if using QtMultiMedia cannot be considered please make it a build option to use just QApplication::beep() for sound notifications so that packagers can decide whether or not they want to introduce dependencies on libcanberra and everything it needs. On Mac at least QApplication::beep() will trigger the user-selected system notification sound (which can but doesn't have to be a horrible beep).
That's a nicety, not necessarily a requirement, at least not for playing notification sounds. Just how often do you get sound alerts at such high rates that caching becomes unavoidable, before you get mad and resolve the issue or yank the plug? I think that proper cross-platform support without requiring multiple 3rd-party libraries is more important than being able to leverage PA caching directly.
According to its website "It comes with several backends (ALSA, PulseAudio, OSS, GStreamer, null)"
- gstreamer is technically x-platform, whether one actually wants to attempt building and using it on these platforms is another question I suppose
We probably lose the ability to play sound on Windows and Mac?
Jul 25 2018
This is an even lower-level fix which applies to all cases where the temporary directory is obtained by reading TMPDIR. That variable contains the path containing the /var symlink; the patch basically sets the variable to the canonical representation of itself, or else sets it to the fallback temp. dir.
This should be a real fix for the issue I've been seeing.
What is your own answer to this? ;)
Jul 24 2018
If you mean that the commit message should be something like "require KDecoration > 5.12.90" then yes, that's what I'd write.
FWIW, without this check (and an older KDecoration2), clang first complained about the override tag on non-virtual methods, and then failed to link the decoration autotest when I removed the overrides because supposedly the two request* methods weren't found. I understand the former error, the latter not so much (why wouldn't requestShowTooltip() be treated as an extension?)
Jul 23 2018
Patch for the 5.2 branch (fully tested):
Refactored for the current master branch
Jul 22 2018
I'm just waiting to know what branch to push to. I'm using it in 5.2.x so technically it could go there.
Jul 16 2018
Using QApplication::style()
This is what approach 1 looks like with (clockwise) Breeze Dark, Oxygen and Wonton Soup:
Jul 11 2018
build on Mac too
Ping?
Ping?