[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#3102: top has problems when not on console



Michael Alan Dorman writes ("Bug#3102: top has problems when not on console"):
> In message <[🔎] 199605231554.IAA08325@bb29c.mdd.comm.mot.com>, Bill Mitchell writes:
> >`infocmp linux` says that el=\EK, vs. el=\E[K$<3> for 'infocmp vt100.`
> >I don't have a vt100 manual handy to check what it says that the
> >el capability for TERM=vt100 should be set to.  If it's not
> >"^[[K$<3>", then this is a ncurses-base bug.  if it is, then
> >it looks like this is not a bug.
>
> If someone with access to a real vt100 and/or vt100 manual can confirm
> whether this string is wrong, do please reassign this bug to ncurses, so
> I'll take care of it.
>
> I would stress, though, that this should only be done if the person so
> doing has come up with an _authoritative_ answer---and that doesn't just
> mean saying it like you mean it. :-) Otherwise I end up bugging Eric
> Raymond with that turns out to be a non-bug---and Eric has been very
> helpful when it comes to Debian + ncurses, so I certainly don't want to
> send him on unnecessary wild-goose-chases.

terminfo(5) says:
       A delay in milliseconds may appear anywhere  in  a  string
       capability,  enclosed in $<..> brackets, as in el=\EK$<5>,
       and padding characters are supplied by  tputs  to  provide
       this  delay.   [ much deleted ]

So, this $<3> isn't part of the string to be output to the terminal -
it's meant to be used to generate padding.

So, this may be a bug in ncurses or procps, depending how procps is
using the capability - in particular, whether it really is calling
whatever ncurses routine is supposed to translate the string.

Ian.


Reply to: