From mboxrd@z Thu Jan 1 00:00:00 1970 MIME-Version: 1.0 From: hiro <23hiro@gmail.com> Date: Tue, 23 Jul 2019 13:38:06 +0200 Message-ID: To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net> Content-Type: text/plain; charset="UTF-8" Subject: [9fans] draw rpcs: 'o' and 'b' rectangles, on-screen, relative coordinates for screenimage Topicbox-Message-UUID: 01fbe38c-eada-11e9-9d60-3106f5b1d025 I've been working on some abstraction for attach/detach support in draw(3). To summarize the implementation: this entails mostly syncing draw(3) state from two different instances, one of which being the one from vncs, the other any terminal or drawterm (instead of running rio you run some script that attaches to the devdraw replicator in vncs). As all coordinates are by default absolute, right now i only see a way to have my program be the only client to draw(3), once i were to run it inside a rio window (like a subrio), the rio window would need to be in 0,0 coordinates. I have tried to put offsets into the rect in the 'o' rpc on all allocated images that have pointers to screenimage, because what i would really need is relative coordinates. i also tried to just give different rect or cliprects to an allocateimage 'b' rpc. But this results in weird image corruption: after some graphics operations, random parts of the images are either covered by black or white areas, or even some weird other stuff looking like some kind of random memory regions. Is it supported to set offsets onto the rect in b and o rpc, specifically targeting screenimage? or are these features that have never been finished? It's kinda hard to debug, so I would be very happy to get some higher level guidence.