On Thursday March 29 2018 12:00:00 [email protected] wrote:
>My problem with that explanation is that I see no reason at all to >think Chrome ever connects to that socket. Why would it? =20 Again, it doesn't have to itself. It may trigger something that ultimately leads to a connection being made on the socket. Let me return your question: why would Chrome do anything that will make XQuartz think it needs to come to the front? It has even less reason to do so (Chrome exists for X11 environments, so it *might* contain code somewhere that reads the DISPLAY variable if it's set and could even try to do something with the returned value; Chrome is almost an OS in itself inside.) And as Wayne pointed out: the actual access and connection attempt might come from code loaded from a remote site (Javascript, Java, whatever). >> I don't know of any feature to make XQuarz bring itself to the front >> in reaction to certain events > >Every windowing API has a function to move the application to the >front. Apple's seems to be "makeKeyAndOrderFront"=20 I meant a feature exposed to the user, evidently. >seems likely to me that XQuartz contains calls to such functions in >response to particular events. Normally applications do not have to do this (and it's even frowned upon I think), but XQuartz might have specific needs. If I were to guess I'd say it might be related to the rootless/rooted modes. >interacting with it (or launchd) in any way. Also we know other >browers provoke the same behavior. =20 Whatever the explanation is, it must be something that doesn't apply to my set-up. I'm not ever seeing XQuartz come to the front - not with Chrome, not with any of the other browsers I use. > Hmm, that's not the only reason. Another is that the behavior could > occur without any assistance from launchd. Excuse me? If unloading the plist and disabling the auto-activation via launchd resolved the issue, this would have been the most parsimonious explanation (however launchd were triggered). Unloading the plist doesn't help, so yes, there must be another explanation, one that doesn't include the launchd (at least, not the feature where it starts an application if a connection is made to one of its sockets). LaunchServices can still be involved as I explained earlier, if somehow XQuartz ends up as the application registered to open certain kinds of douments, and if somehow it's OK for web browsers to have such documents opened without user intervention. Or do all ye affected users have their browsers configured to open "safe" documents automatically? >occur without any assistance from launchd. All that is needed is for >Chrome (et alia) to send a message to "all windows". It appears in You can't do that on OS X, you can only treat with other applications. (WIthout "voodoo" in any case.) And why would a web browser send a message to all other applications (or windows)? Why would it not do that on my system, or why would XQuartz not react to it on my system (and those of the majority, I assume, since the bug is still there)? Why would XQuartz be the only affected application that reacts this way? Has anyone verified if this issue also occurs when using another window manager (I think the legacy twm is installed by default, if not MacPorts has a selection of X11 window managers that are for the most part guaranteed not to contain any code that calls Mac SDKs directly. R. _______________________________________________ Do not post admin requests to the list. They will be ignored. X11-users mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/x11-users/archive%40mail-archive.com This email sent to [email protected]
