Date: Fri, 6 Dec 1996 13:52:34 +0200 Message-Id: <199612061152.NAA01559@sc-uni.ktu.lt> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: "Salvador Eduardo Tropea (SET)" From: Algirdas DOT Kunigelis AT rf DOT ktu DOT lt (Algirdas Kunigelis) Subject: RHIDE/Linux Cc: djgpp AT delorie DOT com >>The >>overhead of using read()/write() is _minimal_ on 'device' files, but if >>you're still not satisfied, you can mmap() the file descriptor into your >>address space, and voila! >Then cames the virtual consoles problem ... > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ No way! There are different /dev/vcsa# files for different virtual consoles, namely the # in the filename above means the virtual console number. So if you open /dev/vcsa1 and mmap() it's file descriptor to a chunk of memory, you can access that memory as if it was a video buffer, and Linux will still handle virtual console switching, i.e. you can write stuff to the mmap()'ed buffer while a different console is on, and it will be there when you switch back. Quit looking for problems where they don't exist! Linux isn't Microsoft designed DPMI! And what is all this about anyway? Let's get this straight: THERE IS A WORKING TURBO VISION FOR LINUX (well, at version 0.2 currently) AND IT WORKS REALLY GOOD. It supports mouse and other stuff as well. What many people need and beg you (Robert? SET? others? me??) is to port the RHIDE to it (TV/Linux I mean). Is there really so much difference between standard TV and Robert's port? Another issue might be this 'I'd code for Linux if more people used it' attitude. Hey, man, CODE FOR LINUX AND MORE PEOPLE WILL USE IT! Huh. enoough of this capital letter madness. Don't get me wrong, I love you all! Martynas