On Fri, Mar 30, 2012 at 04:43:21PM +0200, Vincent Lefevre wrote:
> On 2012-03-27 17:22:01 -0400, Thomas Dickey wrote:
> > As I recall, the main reason why the change for #515609 was accepted was
> > considering the non-ncurses uses (such as termcap applications and slang).
> 
> Not sure what you mean by "non-ncurses uses", but scripts that
> use the tput command were affected by bug #515609, and tput is
> provided by ncurses-bin.

I distinguish things by what features they use:
        ncurses - fullscreen applications (things that rely on libncurses in the
                current packaging scheme).
        terminfo - things that use the database but do not use ncurses screen
                optimization (this includes all of the command-line tools)
        termcap - interface provided for compatibility (applications such as
                vim and w3m).

terminfo/termcap applications have their own logic.

Elimar's probably can probably be solved by setting TERM to linux2.2

comparing linux2.2 to linux3.0.
    comparing booleans.
    comparing numbers.
    comparing strings.
        rmacs: '\E[10m', '^O'.
        sgr: 
'\E[0;10%?%p1%t;7%;%?%p2%t;4%;%?%p3%t;7%;%?%p4%t;5%;%?%p5%t;2%;%?%p6%t;1%;%?%p9%t;11%;m',
 
'\E[0;10%?%p1%t;7%;%?%p2%t;4%;%?%p3%t;7%;%?%p4%t;5%;%?%p5%t;2%;%?%p6%t;1%;m%?%p9%t\016%e\017%;'.
        sgr0: '\E[0;10m', '\E[m\017'.
        smacs: '\E[11m', '^N'.

Sure, \e[11m is "wrong".  But a lot of Linux console is wrong, and
has been for more than 15 years.  SI/SO was deliberately broken in
the late 1990s by the people who added UTF-8 support to Linux console,
and it remained broken for about 10 years.

However, I don't have enough information to see whether changing the
default behavior was a bad idea or a good one.  That comes down to
seeing which group is largest:

        a) ncurses users are largely unaffected (except for those doing
           remote logins from one Linux console to another Linux system
           with different kernel versions hence different default TERM
           settings).

           I did spend some time attempting to identify the first kernel
           version in which the SI/SO change _actually_ appeared.  Let's
           not go there - it's enough to point out that by the time it
           was actually delivered, it was at least 5 years overdue ;-)

        b) users with workarounds for Linux console quirks (as Elimar
           appears to be) are affected.

        c) some unspecified number of users would like to use tput for
           drawing lines (they're affected).  I'd expect only a small
           fraction of those to be slang users (simply because they
           generally would use slang's quirks rather than script with
           tput).

-- 
Thomas E. Dickey <dic...@invisible-island.net>
http://invisible-island.net
ftp://invisible-island.net

Attachment: signature.asc
Description: Digital signature

Reply via email to