From: Bruno H. <br...@cl...> - 2009-08-19 08:01:40
|
Sam Steingold asked: > why wasn't this the original treatment of -I? > I mean, why does -I merely disable TAB, not the whole readline? The first use-case for -I is when running in a terminal window, but copy&paste of text containing tabs is desired. The second use-case for -I came later and is as documented: for better interoperation with ilisp and SLIME. > How about this patch? It would kill the first use-case. But you can save that by introducing another option. So that you have one option for the first use-case and another option for the second use-case. But I find two things surprising: 1) Why was same_tty apparently true in this sagemath environment? How are the I/O channels (pipes, ttys) set up? 2) The output '(* 4 5)\r\x1b[C\x1b[C\x1b[C\x1b[C\x1b[C\x1b[C\x1b[C\n20' (7 chacters, then CR, then 7 times move left, then newline) is way overkill for a normal terminal. It indicates that the value of TERM was not set to a meaningful value. So before changing the clisp options, I would ask - to test whether setting TERM to a different value (try 'dumb', 'vt100', 'xterm') would produce less control characters, - to test whether using a pipe instead of a tty would disable readline entirely (since that is desired)? Bruno |