Re: lynx-dev Relatively mild dev13 build problems

[Klaus Weide]

On Sat, 23 Oct 1999, Michael Warner wrote:

> Building on "OSF1 V4.0 878 alpha" with SLang...
>[...]
 
> Next, I was using the configure option  "--enable-internal-links"
> and got the error:
> 
> gcc -DNO_CHARSET_utf_8=0 -DNO_CHARSET_us_ascii=0
> -DNO_CHARSET_iso_8859_1=0 -DHAVE_CONFIG_H -DLOCALEDIR=\
> "/users/wsunix/warner/share/locale\" -I. -I.. -Ichrtrans
> -I./chrtrans -I.. -I../src -I../WWW/Library/Imple mentation
> -g -O2 -DDSYS5 -W -Wall -Wbad-function-cast -Wcast-align
> -Wcast-qual -DXTSTRINGDEFINES -Winlin e -Wmissing-declarations
> -Wmissing-prototypes -Wnested-externs -Wpointer-arith -Wshadow
> -Wstrict-prototype s -c ./HTML.c
> ./HTML.c: In function `strtolower':
> ./HTML.c:160: warning: subscript has type `char'

For this, and lots of similar warnings below:
There's a bunch of isdigit() atc. and tolower() that should have
their argument cast to (unsigned char)... (or use one of the lynx
macros instead, like TOLOWER).

(This is different from what I complained about recently, i.e.
unnecessary casts - in theses cases, casts *should* be there afaik.)

> ./HTML.c: In function `LYHandleFIG':
> ./HTML.c:831: `intern_flag' undeclared (first use in this function)
> ./HTML.c:831: (Each undeclared identifier is reported only once
> ./HTML.c:831: for each function it appears in.)

That should be fixed now with the patches in
<http://enteract.com/~kweide/lynx/>.

> ./HTML.c: In function `HTML_start_element':
> ./HTML.c:2161: warning: subscript has type `char'
> ./HTML.c:2259: warning: subscript has type `char'
> ./HTML.c:2275: warning: subscript has type `char'
> ./HTML.c:2816: warning: subscript has type `char'
> ./HTML.c:2818: warning: subscript has type `char'
> make[1]: *** [HTML.o] Error 1
> make[1]: Leaving directory `/tmp/warner/lynx2-8-3/src'
> make: *** [all] Error 2
> 
> Configuring without that option let the compile finish.
> 
> Thirdly, the -nounderline command line argument doesn't seem
> to do anything for me.  I only tried it on one page, though --
> <http://www.drudgereport.com/matt.htm>, I think it was.  Pretty
> much the whole page is in <B></B> tags.  Does -nounderline depend
> on SLang vs ncurses, or terminfo settings, or something?

Those no* flags seem to act at a later level - not before the
{bold,reverse,underline} -> COLOR:n translation that happens in
color mode, but after (if at all).

I see the same as you (if I understand right) with -nounderline,
with ncurses:  on that page, all the main text is still shown as
COLOR:4.

But when I force *mono* mode, i.e. combine -nounderline with -nocolor
(or -term=vt100 for ncurses), the -nounderline has the expected
effect.

The only way I know to get what you want, removal of the bold-bit
prior to COLOR-translation, is to redefine the COLORs -
make 4,5,6,7 the same as 0,1,2,3.  Actually that's more flexible
since you don't have to do it for all four of them (but less convenient
since it's not a command line option).


   Klaus