<<< Date Index >>>     <<< Thread Index >>>

Re: S-Lang -> ncurses: Lost bold attribute



On Sun, Oct 12, 2003 at 11:13:23AM -0400, Thomas Dickey wrote:
> On Sun, 12 Oct 2003, Axel Thimm wrote:
> 
> > On Sun, Oct 12, 2003 at 06:34:06AM -0400, Thomas Dickey wrote:
> > > On Sun, 12 Oct 2003, Axel Thimm wrote:
> > >
> > > > On Sat, Oct 11, 2003 at 12:35:24PM -0400, Thomas Dickey wrote:
> > > > So is this expected behaviour, e.g. mutt&ncurses have no bright/bold
> > > > attribute? I couldn't find any reference to this problem elsewhere.
> > >
> > > it really depends on how the terminal description is treated by the
> > > ncurses/slang library.  slang has some hardcoded behavior to treat
> > > the blink attribute as "bright".  you may be seeing that.  otherwise,
> > > "bright" may refer to colors 8-15, which (given a correct terminfo),
> > > is treated the same in ncurses and slang.
> >
> > I did replace all brightsomething entries with colorN entries without
> > effect. E.g. color14 and color6 are the same non-bold cyan under
> > ncurses, while they are bold and non-bold under slang.
> >
> > I am using xterm-16color under xterm if that matters. Maybe ncurses
> > and slang have a different interpretation of the terminal
> > capabilities?
> 
> It does - you can test this easily:
> 
> > $ infocmp xterm xterm-16color
> > comparing xterm to xterm-16color.
> >     comparing booleans.
> >     comparing numbers.
> >         colors: 8, 16.
> >         ncv: NULL, 32.
> 
> ncv#32 means that the bold attribute doesn't work properly in combination
> with colors.  It's partly dependent on the resource settings - when I
> implemented that one I was probably considering only the aspect of mapping
> colors 8-15 as bold, but there's also the chance that one would be using
> a bold font, etc.  If you tweak the terminfo to cancel that, you should
> see ncurses combining bold+colors (unless it's not done someplace in mutt,
> of course).  slang ignores ncv, ncurses looks at it and suppresses video
> attributes which are indicated to conflict with colors.

Thanks Thomas, I can confirm this to be working.

For the people reading this in the archives, here is what to do: I
used

        infocmp xterm-16color > xterm-16color
        cp xterm-16color xterm-16color2

to get a source representation of xterm-16color, edited
the ncv attribute away, compiled it with tic to
xterm-16color2.

        tic xterm-16color2

TERM=xterm-16color2 gives me my colors back in mutt.

--- xterm-16color       2003-10-16 18:19:47.000000000 +0200
+++ xterm-16color2      2003-10-16 18:20:15.000000000 +0200
@@ -1,7 +1,7 @@
-#      Reconstructed via infocmp from file: /usr/share/terminfo/x/xterm-16color
-xterm-16color|xterm with 16 colors, 
+#      Reconstructed via infocmp from file: /usr/share/terminfo/x/xterm-16color
+xterm-16color2|xterm with 16 colors, 
        am, bce, km, mc5i, mir, msgr, xenl, 
-       colors#16, cols#80, it#8, lines#24, ncv#32, pairs#256, 
+       colors#16, cols#80, it#8, lines#24, pairs#256, 
        acsc=``aaffggiijjkkllmmnnooppqqrrssttuuvvwwxxyyzz{{||}}~~, 
        bel=^G, blink=\E[5m, bold=\E[1m, cbt=\E[Z, civis=\E[?25l, 
        clear=\E[H\E[2J, cnorm=\E[?25h, cr=^M, 
-- 
Axel.Thimm@xxxxxxxxxxxxxxxxxxx

Attachment: pgpTyjHRMPQxr.pgp
Description: PGP signature