Cybernethics / Cybernéthique

The Linux framebuffer console for the Jornada820 and other machines does not heed all the cursor controls that it has on the x86 console. Therefore, many programs that try to reprogram the cursor, including the login program and emacs, end up making the cursor invisible. A correct escape sequence to make the cursor appear, as per the Framebuffer-HOWTO, is ESC [ ? 2 c (thank you once more, pineal gland). Now, my problem, is to convince Emacs to output this control sequence instead of the one it uses to program the cursor, or sometimes after it makes the cursor invisible, so as to ensure the cursor is visible. Can any (X)Emacs guru out there help me find out how?

Yes, of course, an alternative take would be to edit the framebuffer driver and fix it so it works as advertised, or at least doesn't make the cursor invisible unless explicitly requested. However, I'd also like to convince Emacs to output other control sequences when it initializes its output: for instance, I sometimes found the console in the wrong output encoding mode, I'd like to ensure latin1 output encoding with the ESC ( ? B control sequence. The scary thing is I remembered enough of this sequence to retrieve it by testing things on the console, since I had spent quite some time fixing the same issues on a PC. (that is, I remembered ESC something, ( something and an uppercase letter.) My experimentation included the writing of a shell script to display those characters in the ASCII table that are not control codes; the script takes several seconds to run, spawning three subprocesses at every character, and using eval. Ain't Unix a wonderfully designed operating system? I don't think so. Just because I may prefer Linux to its current competition doesn't mean I'm not more and more of a Unix-hater.

( Read more... ) in the big mansion of the uncle who hosts me, with the WiFi router in a central location, I could walk around the whole house with my other uncle's iPAQ 5555, and stay connected to the Internet the whole time through ( Read more... )

In a c.l.l discussion about TAOUP (see also this LWN review), Dan Barlow of cirCLe fame (and who also has a nice diary) neatly summarizes the difference between the Unix way of programming and the Lisp way: Re: LISP & "The Art of Unix Programming". (As for Dan's final remark, how do you signal an exception over a pipe? You can't. And here's yet another place where the Unix way breaks apart.)