On Mon, Mar 25, 2019 at 7:50 PM Thomas Stüfe <[email protected]> wrote: > > > > On Mon 25. Mar 2019 at 19:42, Volker Simonis <[email protected]> wrote: >> >> Hi Thomas, Christoph, >> >> thanks for looking into my fix. I hope I've addressed all your >> concerns with my new webrev: >> >> http://cr.openjdk.java.net/~simonis/webrevs/2019/8220528.v4 >> >> By the way, I've tested the new implementation by simulating a >> multi-screen environment on my Linux box with Xephyr: >> >> $ Xephyr +xinerama -screen 1024x768 -screen 1024x768+1024+0 -ac -listen tcp >> :1 >> >> And started a Swing application from AIX where I've redirected the >> DISPLAY to myhost:1: >> >> $ DISPLAY=myhost:1 ./images/jdk/bin/java -showversion >> -Dsun.awt.nativedebug=true -Dawtdebug.trace=true -Dawtdebug.on=true >> -Dawtdebug.ctrace=true -cp ~/Java HelloSwing >> >> The various debug properties will lead to the following output: >> >> openjdk version "13-internal" 2019-09-17 >> OpenJDK Runtime Environment (slowdebug build 13-internal+0-adhoc.xxx.jdk-jdk) >> OpenJDK 64-Bit Server VM (slowdebug build >> 13-internal+0-adhoc.xxx.jdk-jdk, mixed mode) >> Xinerama extension is available >> calling XineramaQueryScreens func >> Enabling Xinerama support >> num screens = 2 >> allocating 2 screens >> >> The Swing frame will appear on one of the Xinerama screens and can be >> move with the mouse between the two available screens. >> >> If I hear no further objections, I plan to push this by the end of the week. > > > All good now. Thank you. >
Thanks. > Strange that this error was not detected before, what with the masses of > people running swing apps on AIX. > It was not really an error. Basic Swing/AWT worked (we actually passed the TCK). What did not work was some advanced transparency/translucency support (due to failing Xrender) and multiscreen support (due to failing Xinerama). > Cheers, thomas > > >> >> Thank you and best regards, >> Volker >> >> >> >> On Mon, Mar 18, 2019 at 8:59 PM Thomas Stüfe <[email protected]> wrote: >> > >> > Sorry, not awt/2d group, but to me it looks fine too. Only nit: the error >> > printout at 495 I would put right after the associated dlsym at 456. You >> > want to make sure dlerror() is not stale at the point where you call it, >> > since it would return NULL. >> > >> > Cheers, Thomas >> > >> > >> > >> > On Mon, Mar 18, 2019 at 3:17 PM Volker Simonis <[email protected]> >> > wrote: >> >> >> >> Ping... >> >> >> >> Can I please also get a review from the awt/2d group? >> >> >> >> Thanks, >> >> Volker >> >> >> >> On Wed, Mar 13, 2019 at 11:09 AM Volker Simonis >> >> <[email protected]> wrote: >> >> > >> >> > Hi, >> >> > >> >> > can I please have a review for the following change which fixes >> >> > Xinerama and Xrender support on AIX: >> >> > >> >> > http://cr.openjdk.java.net/~simonis/webrevs/2019/8220528/ >> >> > https://bugs.openjdk.java.net/browse/JDK-8220528 >> >> > >> >> > The change basically fixes the way how the corresponding libraries >> >> > (libXext and libXrender) get loaded on AIX and does some cleanups with >> >> > regards to the various platform macros. I've also added some debug >> >> > tracing to the Xrender initialization code similar to the Xinerama >> >> > initialization. >> >> > >> >> > Thank you and best regards, >> >> > Volker
