Segmentation fault when using both webvfx and audiowaveform about mlt

[ddennedy]

This works for me using melt or qmelt from Shotcut 16.04. You must make sure that both the MLT qt module and WebVfx are built against the same version of Qt. Where did you get the WebVfx source? From https://github.com/mltframework/webvfx (not rectalogic)? You should try also with qmelt.

What is "DISPLAY=:1.0"? Are you using a headless system?

[geoffroymontel]

Thanks for your quick reply !

Yes, I'm using a headless system, compiling MLT on a bare Ubuntu 14.04 TLS server.

I'm building MLT and WebVfx with melt build-melt.sh script.
I just set ENABLE_WEBVFX=1 in build-melt.conf

I'm not yet familiar with qmelt, I will try it out, and will try to find out whether qt and webvfx are built against the same version of Qt.

Wil keep you posted.

[ddennedy]

What is DISPLAY pointing to exactly?

On a headless system, when using Qt, you typically use xvfb-run -a melt ... (or qmelt). qmelt is simply a melt wrapper provided by webvfx that addresses the warning you are getting "QApplication was not created in the main() thread." It creates a QApplication in the main thread.

I think whatever you are trying to do with DISPLAY is not working and causing your error "GLib-CRITICAL **: g_main_context_push_thread_default: assertion 'acquired_context' failed"

If you plan to use WebGL in WebVfx, there is more work to be done to ensure you have mesa with llvmpipe and add -s '-screen 0 640x480x24' to the xvfb-run command.

[geoffroymontel]

Dan, you were right once again ! It was a problem of compatibility between QT4 and QT5.
I reinstalled my dev box with QT5 development libraries only, recompiled melt and webvfx and now it's running smoothly.
I haven't tried yet using xvfb-run nor qmelt but I will investigate those too.

Thanks a lot !