Hi Mads, im a big fan of Kick Assembler. i believe this issue is fixed in trunk (upstream dep issue, actually), any chance you can verify?
macOS bindist no longer require platypus. And other bindist issues fixed.
I'm not sure i'm seeing the same problem in trunk Would you be able to try to repo it using the latest code from SVN, without hardcoding the path to the dylib? I'm wondering if the real problem is that a signed .app is trying to load the dylib from the system, rather than it being due to needing the full path.
GTK Crashes on Font Selection in Preferences.
I guess the main question is, if you install deps from homebrew and compile and install VICE trunk on your machine - does the problem still occur? macOS Build instructions are at doc/building/macOS-Howto.txt I just retested using deps I installed a few months ago, and again using updated deps, and no crash. So I think it was just a bug in upstream that has since been fixed. Feel free to reopen if you can repro in trunk
MAC Vice 3.8 / GTK / WiC64 crashes right after startup
Unless it's replicable in trunk, I think we have to assume this was inadvertently fixed already by another change. Feel free to reopen if you can find a way to replicate.
I can't replicate this anymore. Seems the -wic64colorizetrace option has been removed in the meantime also
I actually did check this when I saw the email. On my machine 3.9 crashes deep inside pango when doing this, but a build i have from WIP joystick work doesn't and shows a font selection dialog. So it might be that it's a problem fixed by a newer version of the dependencies.
Add link to GitHub actions run to all build failure notifications
Fixed macOS build (currently no host joystick support)
Will at least get it building again today.
Update macOS bindist to latest notarisation tool
Update macOS bindist for latest GTK3
lack of code-signing might be related too
ugh, this is the test that the homebrew guy forced me to add when I last wanted to fix the homebrew build formula. And now the test itself breaks the build and creating work, classic.
I just tried this morning and I found that x64sc won't even start on Windows for me. The other emulators do. So there is a bigger issue to investigate first
Forgot that replies to the mailing list emails are disabled. I'm probably the last person who worked on the mouse code. It would be good to get a description of how it behaves for you. Does it work if you don't make fast/big mouse movements?
(macOS) Register all the ttf fonts, not just C64_Pro_Mono-STYLE
Fixed unused variable in non-debug builds
(macOS) Moved system includes to end of generated CMake include list. Fixes hvsc #include "md5.h"
are you able to test that adapter on a linux or windows machine?
still here if something serious needs fixing on mac, sounds like this would be hard to replicate though.
MAC Vice 3.8 / GTK / WiC64 crashes right after startup
this one is tricky to find - it fails in userport_wic64_store_pbx() but only if optimisation is enabled.
MAC Vice 3.8 / GTK / WiC64 crashes right after startup
thanks for the report, ill look at it tomorrow
Fixed the ritual busting of vice.text iso-8859-x
Partial revert
Changed macOS monitor hotkey from reserved Cmd-h to Cmd-Escape. Thanks to Oskar Linde for the patch!
Change default hotkey for opening Monitor on Mac
Yeah, not sure why this was changed, Cmd-Escape is a good change, thanks.
Update macOS bindist build flags to closer match windows builds
Updated macOS bindist to require macOS 12. Let's see if this upsets people, building deps for old macOS support is a hassle.
Fixed github workflow not waiting for build_deb. It happens to work already because build_deb is faster than build.
Fixed release_id var in make release github workflow
configure fails on MacOS X
i couldnt replicate it and it sounds like its fixed :)
bindist requiring vice.pdf is fixed in r44747 @goran - looks like you are using Homebrew on Intel, is that correct? Do you have pkg-config installed?
Fixed macOS bindist when vice.pdf isn't built
im out of town for another week or so but i'll take a look when i get back. I only have PS and XB controllers to test with, though.
sure ill look. very occupied by new job but am still here :)
commit++
Fixed shutdown locking bug found by an assertion
what OS, CPU and GPU do you have? does changing the sound settings (also try disabling sound) change anything? SDL or GTK?
What speed are you running the emulator at? What is the refresh rate of your monitor?
Fixed a couple of bad reads
Fixed single threaded build
Fixed 64-bit printfy things
Added assert to catch unsafe mainlock_yield() on UI thread
fixed memory leak
Prevent sound flush on UI thread, which can happen in response to settings changing resources
Fixed not generating sources until after they're looked for
Fixed cmake_bootstrap.sh when EXTRA_DIST lists files that don't exist
Yeah, I can see a few things wrong with the SDL dual window stuff in my branch. I'll clean it up.
yikes. I just tested this on macOS and none of these issues exist. I think the fix is for users to upgrade to macOS
I can't remember if I did anything for this on trunk, but it shouldn't be an issue in the vice-4 branch. Would you be able to test both trunk and vice-4 to figure out the current status? My understanding is that i'd be able to merge the vice-4 stuff once we get 3.7.x point releases out of the way, are we still planning a point release?
Fix 'echo -n' not working inside configure script on macOS and probably other platforms
Do not build PDF docs unless --enable-pdf-docs (yay!)
A better fix for x64 autostart when setting VICIIborders via config/cli
Fixed x64 autostart when setting VICIIborders via config/cli
commit++
Fixed missing void in macOS code that somehow never triggered a warning before
Fixed macOS bindist. Probably doesn't work without code signing, however an attempt is now made to detect and use a signing identity.
Fixed annoying warning about unused mem_burst_read
Fixed dtv crash on shutdown. There was an assumption that 4 printer drivers were set up that had a form feed callback, in this case there were only 3
another unsigned -> signed warning
Fixed a bunch of intptr_t / void * comparison warnings by defining CBM_FILE as void *
Fixed unsigned -> signed warnings
Removed pointless assignment
I've added an attempt to make the bindist script compatible with newer platypus in r43597. Can you please test?
Attempt to make macOS bindist script work with multiple platypus versions
fixed GNU sed assumption
So, I'm not seeing this issue. I see that your version of platypus is newer than mine, I have version 5.3. How did you install a newer platypus?
Make binary fails with Add: "CFBundlePackageType" Entry Already Exists
It wasn't when I made the offical 3.7.1 binaries, r43150 is a seemingly random early december revision. @goran can you please let us know if this is still an issue with the latest trunk, or if you prefer, 3.7.1 ?
Fixed platpus output redirection, thanks Patrice Duroux for the patch!
Added a log interface locking to threaded builds
Fixed debug logging of pointer without cast to void*
Fixed debug logging of intptr_t without cast to void*
Fixed debug logging of uninitialised data
Gtk3 mouse emulation broken under Windows
This should be fixed as of r43149
Fixed trailing spaces
Fixed the windows mouse warp thing
We're now in a bit of a weird weird place where cmd-w closes the monitor window (which is right), but also toggles warp when used on the emu windows. Perhaps mac should use the same (general) shortcuts as win/linux - i.e. use alt instead of cmd.
We're now in a bit of a weird weird place where cmd-w closes the monitor window (which is right), but also toggles warp when used on the emu windows. Perhaps mac should use the same shortcuts as win/linux - i.e. use alt instead of cmd.
Commit 40432 causes timeouts using the GEOS fast loader
A big thank you from me for your work on this so far. Could you please test r42844 to see if it changes anything for you?
Experimental fix for bug https://sourceforge.net/p/vice-emu/bugs/1706/
Removed another undefined behaviour
Deliberate integer sign conversion made explicit
Deliberate integer overflow made explicit
Deliberate integer overflow made explicit
Deliberate integer overflow made explicit
Fixed a few undefined behaviours due to unexpected rvalue promotion to signed int, noticed by when running x128 with clang -fsanitize=undefined.
Updated website macOS binary links and *cough* added a link to the discord server
xvic ntsc weird display
xvic NTSC crashes when enabling interlace mode
Thanks for rechecking!