From 368eacd6aeb9b724a7361289892efe34491b54b5 Mon Sep 17 00:00:00 2001 From: Ingo Schwarze Date: Tue, 16 Aug 2022 17:45:55 +0000 Subject: When starting a new input line, even when continuing the same output line, use the current output position as the reference position for tabs on that input line. This brings mandoc in line with the behaviour of GNU, Heirloom, and Plan 9 roff. --- man_term.c | 5 ++++- mdoc_term.c | 7 +++++-- term.c | 43 +++++++++++++++++++++++++++++++++++++------ term.h | 3 ++- term_ascii.c | 2 +- 5 files changed, 49 insertions(+), 11 deletions(-) diff --git a/man_term.c b/man_term.c index b5e0da70..33431749 100644 --- a/man_term.c +++ b/man_term.c @@ -910,8 +910,11 @@ print_man_node(DECL_ARGS) (p->flags & TERMP_NONEWLINE) == 0) term_newln(p); p->flags |= TERMP_BRNEVER; - } else + } else { + if (n->flags & NODE_LINE) + term_tab_ref(p); p->flags &= ~TERMP_BRNEVER; + } if (n->flags & NODE_ID) term_tag_write(n, p->line); diff --git a/mdoc_term.c b/mdoc_term.c index d221b61d..a9d67b0c 100644 --- a/mdoc_term.c +++ b/mdoc_term.c @@ -1,6 +1,6 @@ /* $Id$ */ /* - * Copyright (c) 2010, 2012-2020 Ingo Schwarze + * Copyright (c) 2010, 2012-2020, 2022 Ingo Schwarze * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2013 Franco Fichtner * @@ -320,8 +320,11 @@ print_mdoc_node(DECL_ARGS) (p->flags & TERMP_NONEWLINE) == 0) term_newln(p); p->flags |= TERMP_BRNEVER; - } else + } else { + if (n->flags & NODE_LINE) + term_tab_ref(p); p->flags &= ~TERMP_BRNEVER; + } if (n->type == ROFFT_COMMENT || n->flags & NODE_NOPRT) return; diff --git a/term.c b/term.c index 9823bb92..74d9db67 100644 --- a/term.c +++ b/term.c @@ -157,7 +157,11 @@ term_flushln(struct termp *p) /* Finally, print the field content. */ term_field(p, vbl, nbr); - p->tcol->taboff += vbr + (*p->width)(p, ' '); + if (vbr < vtarget) + p->tcol->taboff += vbr; + else + p->tcol->taboff += vtarget; + p->tcol->taboff += (*p->width)(p, ' '); /* * If there is no text left in the field, exit the loop. @@ -177,7 +181,9 @@ term_flushln(struct termp *p) vbr += (*p->width)(p, ' '); continue; case '\n': + case ASCII_NBRZW: case ASCII_BREAK: + case ASCII_TABREF: continue; default: break; @@ -258,9 +264,11 @@ term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget) size_t vn; /* Visual position of the next character. */ int breakline; /* Break at the end of this word. */ int graph; /* Last character was non-blank. */ + int taboff; /* Temporary offset for literal tabs. */ *nbr = *vbr = vis = 0; breakline = graph = 0; + taboff = p->tcol->taboff; for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) { switch (p->tcol->buf[ic]) { case '\b': /* Escape \o (overstrike) or backspace markup. */ @@ -306,12 +314,19 @@ term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget) *vbr = vis; continue; + case ASCII_TABREF: + taboff = -vis - (*p->width)(p, ' '); + continue; + default: switch (p->tcol->buf[ic]) { case '\t': - vis += p->tcol->taboff; + if (taboff < 0 && (size_t)-taboff > vis) + vis = 0; + else + vis += taboff; vis = term_tab_next(vis); - vis -= p->tcol->taboff; + vis -= taboff; break; case ASCII_NBRZW: /* Non-breakable zero-width. */ break; @@ -354,8 +369,10 @@ term_field(struct termp *p, size_t vbl, size_t nbr) size_t vis; /* Visual position of the current character. */ size_t vt; /* Visual position including tab offset. */ size_t dv; /* Visual width of the current character. */ + int taboff; /* Temporary offset for literal tabs. */ vis = 0; + taboff = p->tcol->taboff; for (ic = p->tcol->col; ic < nbr; ic++) { /* @@ -368,11 +385,17 @@ term_field(struct termp *p, size_t vbl, size_t nbr) case ASCII_BREAK: case ASCII_NBRZW: continue; + case ASCII_TABREF: + taboff = -vis - (*p->width)(p, ' '); + continue; case '\t': case ' ': case ASCII_NBRSP: if (p->tcol->buf[ic] == '\t') { - vt = p->tcol->taboff + vis; + if (taboff < 0 && (size_t)-taboff > vis) + vt = 0; + else + vt = vis + taboff; dv = term_tab_next(vt) - vt; } else dv = (*p->width)(p, ' '); @@ -437,10 +460,10 @@ endline(struct termp *p) void term_newln(struct termp *p) { - p->tcol->taboff = 0; p->flags |= TERMP_NOSPACE; if (p->tcol->lastcol || p->viscol) term_flushln(p); + p->tcol->taboff = 0; } /* @@ -801,6 +824,14 @@ bufferc(struct termp *p, char c) p->tcol->lastcol = p->col; } +void +term_tab_ref(struct termp *p) +{ + if (p->tcol->lastcol && p->tcol->lastcol <= p->col && + (p->flags & TERMP_NOBUF) == 0) + bufferc(p, ASCII_TABREF); +} + /* * See encode(). * Do this for a single (probably unicode) value. @@ -946,7 +977,7 @@ term_strlen(const struct termp *p, const char *cp) const char *seq, *rhs; enum mandoc_esc esc; static const char rej[] = { '\\', ASCII_NBRSP, ASCII_NBRZW, - ASCII_BREAK, ASCII_HYPH, '\0' }; + ASCII_BREAK, ASCII_HYPH, ASCII_TABREF, '\0' }; /* * Account for escaped sequences within string length diff --git a/term.h b/term.h index 721f0ccc..1a2580b1 100644 --- a/term.h +++ b/term.h @@ -1,7 +1,7 @@ /* $Id$ */ /* + * Copyright (c) 2011-2015,2017,2019,2022 Ingo Schwarze * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons - * Copyright (c) 2011-2015, 2017, 2019 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above @@ -150,6 +150,7 @@ size_t term_len(const struct termp *, size_t); void term_tab_set(const struct termp *, const char *); void term_tab_iset(size_t); +void term_tab_ref(struct termp *); size_t term_tab_next(size_t); void term_tab_free(void); diff --git a/term_ascii.c b/term_ascii.c index efab1fa6..5c105f20 100644 --- a/term_ascii.c +++ b/term_ascii.c @@ -196,7 +196,7 @@ terminal_sepline(void *arg) static size_t ascii_width(const struct termp *p, int c) { - return c != ASCII_BREAK && c != ASCII_NBRZW; + return c != ASCII_BREAK && c != ASCII_NBRZW && c != ASCII_TABREF; } void -- cgit