For Gtk+ main.cpp of the TestNetscapePlugin plugin was forked (to not interfere with mac), merge the changes from main.cpp back to the Gtk+ version.
Created attachment 30329 [details] Properly unregister the plugin instance Copied from a recent Qt bugfix.
Created attachment 30330 [details] Merge r39877 Merge r39877 and log through the JavaScript Console.
Comment on attachment 30330 [details] Merge r39877 Wrong patch. Invoke testNPRuntime and remove the test from the skipped list.
Created attachment 30331 [details] Log through JavaScript Console (r41819) Merge r41819 and log through the javascript console.
Comment on attachment 30329 [details] Properly unregister the plugin instance Clearing review, landed in r43697.
Comment on attachment 30331 [details] Log through JavaScript Console (r41819) Please give a brief summary of why we're not using printf in this case prior to landing this - thanks. r=me.
Comment on attachment 30330 [details] Merge r39877 r=me.
(In reply to comment #6) > (From update of attachment 30331 [details] [review]) > Please give a brief summary of why we're not using printf in this case prior to > landing this - thanks. The only reference in r41819 is a radar bug. The reason for us to follow is two fold. First of all the -expected result has a "CONSOLE MESSAGE: line 0: PLUGIN:" the second is we might run the plugin in a different process and then the stdout's might not be the same, in this case we still want the output to come from the DRT process.
Landed in r43800 and r43801.