On Tue, Oct 11, 2005 at 03:34:49PM +0000, Bryan Henderson wrote: > >But it's missing a piece needed to make curses (and ncurses) decide that > >it's really color-capable. See > > > > http://invisible-island.net/ncurses/ncurses.faq.html#no_color > > > >Specifically, I notice that it's missing one of > > orig_pair (op) > > orig_colors (oc) > > The referenced description says these aren't necessary for Ncurses 5, > which is what I have.
hmm - so it does (I'd forgotten my amendment to that ;-). > Indeed, tput and "worm" (demonstration program that comes with Ncurses) > work fine. Even Lynx is able to display two colors with this eterm > definition that it doesn't with the older no-color one: yellow > foreground and blue background. That reminds me to ask: Is lynx built with ncurses or slang? (While I would recall offhand most of what I wrote on that page 5-6 years ago, I don't recall what slang needs from op/oc). > Is there a way to trace what characters or ncurses capability strings > are being sent to the screen? I suspect Lynx is getting the colors > it's asking for, but just asking for the wrong ones. Lynx has some relevant stuff in it's Lynx.trace file (lynx -trace), but not really enough detail to see more than that it's parsing its config file. Generally what I do in that situation is run the application linked to a debug version of ncurses, and turn on ncurses' trace (which shows lots of detail). For starters, look in lynx's trace for "color", since that doesn't require any recompiling. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net _______________________________________________ Lynx-dev mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/lynx-dev
