Allowing the leak tracer to report/log leaks. From https://bugs.webkit.org/show_bug.cgi?id=217655#c42
Created attachment 452976 [details] Patch
Comment on attachment 452976 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=452976&action=review > Source/WebKit/WebProcess/glib/WebProcessGLib.cpp:88 > +#if USE(GSTREAMER) > + gst_deinit(); > +#endif I think the right place for this is WebProcessMainGtk.cpp and WebProcessMainWPE.cpp. You need to override platformFinalize(). Closing the pages is done here because m_pageMap is private.
Created attachment 452977 [details] Patch
Comment on attachment 452977 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=452977&action=review > Source/WebKit/WebProcess/gtk/WebProcessMainGtk.cpp:75 > +#if USE(GSTREAMER) > + gst_deinit(); > +#endif Don't we need to include a gst header for this? > Source/WebKit/WebProcess/wpe/WebProcessMainWPE.cpp:107 > +#if USE(GSTREAMER) > + gst_deinit(); > +#endif Ditto.
Somehow they were pulled in, the magics of unified builds I suppose. Anyway yes, I'll add explicit includes, it won't hurt.
Created attachment 452979 [details] Patch
(In reply to Philippe Normand from comment #5) > Somehow they were pulled in, the magics of unified builds I suppose. Anyway > yes, I'll add explicit includes, it won't hurt. You'll save Adrian from fixing it after you ;)
Committed r290375 (247691@main): <https://commits.webkit.org/247691@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 452979 [details].
<rdar://problem/89357593>
(In reply to Michael Catanzaro from comment #7) > (In reply to Philippe Normand from comment #5) > > Somehow they were pulled in, the magics of unified builds I suppose. Anyway > > yes, I'll add explicit includes, it won't hurt. > > You'll save Adrian from fixing it after you ;) Indeed! Thanks for adding the explicit includes 😁️