On Tue, Oct 23, 2007 at 10:32:28PM +0200, Bernd Zeimetz wrote: > after reading through the dvi specs and several pages of information > about the dvi special commands I came to the conclusion that 'sh 1' and > 'ip' must be the result of groff not converting .SH and .IP correctly. > Reassigning the package therefore.
You are mistaken; from the file attached to the bug report, 'sh 1' and 'ip' appear in the rendering of a pic table. They have nothing to do with the .SH and .IP man macros. The pieces of code in question in groff are these, from src/devices/grodvi/dvi.cc: void draw_dvi_printer::fill_next(const environment *env) { unsigned int g; if (env->fill->is_default()) g = 0; else { // currently, only BW support env->fill->get_gray(&g); } char buf[256]; sprintf(buf, "sh %.3g", 1 - double(g)/color::MAX_COLOR_VAL); do_special(buf); } ... and: case 'p': { if (np & 1) { error("even number of arguments required for polygon"); break; } if (np == 0) { error("no arguments for polygon"); break; } moveto(env->hpos, env->vpos); if (fill_flag) fill_next(env); else set_line_thickness(env); do_special("pa 0 0"); int h = 0, v = 0; for (int i = 0; i < np; i += 2) { h += p[i]; v += p[i+1]; sprintf(buf, "pa %d %d", milliinches(h), milliinches(v)); do_special(buf); } do_special("pa 0 0"); do_special(fill_flag ? "ip" : "fp"); break; } Do you still think that this is a groff bug? If so, bearing in mind that I don't know a whole lot about DVI, I'd like to know more about why. -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]