=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.45 retrieving revision 1.267 diff -u -p -r1.45 -r1.267 --- mandoc/term.c 2009/03/08 13:57:07 1.45 +++ mandoc/term.c 2017/06/07 20:01:19 1.267 @@ -1,1711 +1,938 @@ -/* $Id: term.c,v 1.45 2009/03/08 13:57:07 kristaps Exp $ */ +/* $Id: term.c,v 1.267 2017/06/07 20:01:19 schwarze Exp $ */ /* - * Copyright (c) 2009 Kristaps Dzonsons + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons + * Copyright (c) 2010-2017 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 copyright notice and this permission notice appear in all - * copies. + * purpose with or without fee is hereby granted, provided that the above + * copyright notice and this permission notice appear in all copies. * - * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL - * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED - * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE - * AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL - * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR - * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER - * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR - * PERFORMANCE OF THIS SOFTWARE. + * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES + * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR + * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN + * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF + * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ +#include "config.h" + +#include + #include #include -#include #include #include #include +#include "mandoc.h" +#include "mandoc_aux.h" +#include "out.h" #include "term.h" +#include "main.h" -/* - * Performs actions on nodes of the abstract syntax tree. Both pre- and - * post-fix operations are defined here. - */ +static size_t cond_width(const struct termp *, int, int *); +static void adjbuf(struct termp_col *, size_t); +static void bufferc(struct termp *, char); +static void encode(struct termp *, const char *, size_t); +static void encode1(struct termp *, int); +static void endline(struct termp *); -/* FIXME: macro arguments can be escaped. */ -#define TTYPE_PROG 0 -#define TTYPE_CMD_FLAG 1 -#define TTYPE_CMD_ARG 2 -#define TTYPE_SECTION 3 -#define TTYPE_FUNC_DECL 4 -#define TTYPE_VAR_DECL 5 -#define TTYPE_FUNC_TYPE 6 -#define TTYPE_FUNC_NAME 7 -#define TTYPE_FUNC_ARG 8 -#define TTYPE_LINK 9 -#define TTYPE_SSECTION 10 -#define TTYPE_FILE 11 -#define TTYPE_EMPH 12 -#define TTYPE_CONFIG 13 -#define TTYPE_CMD 14 -#define TTYPE_INCLUDE 15 -#define TTYPE_SYMB 16 -#define TTYPE_SYMBOL 17 -#define TTYPE_DIAG 18 -#define TTYPE_NMAX 19 - -/* - * These define "styles" for element types, like command arguments or - * executable names. This is useful when multiple macros must decorate - * the same thing (like .Ex -std cmd and .Nm cmd). - */ - -/* TODO: abstract this into mdocterm.c. */ - -const int ttypes[TTYPE_NMAX] = { - TERMP_BOLD, /* TTYPE_PROG */ - TERMP_BOLD, /* TTYPE_CMD_FLAG */ - TERMP_UNDERLINE, /* TTYPE_CMD_ARG */ - TERMP_BOLD, /* TTYPE_SECTION */ - TERMP_BOLD, /* TTYPE_FUNC_DECL */ - TERMP_UNDERLINE, /* TTYPE_VAR_DECL */ - TERMP_UNDERLINE, /* TTYPE_FUNC_TYPE */ - TERMP_BOLD, /* TTYPE_FUNC_NAME */ - TERMP_UNDERLINE, /* TTYPE_FUNC_ARG */ - TERMP_UNDERLINE, /* TTYPE_LINK */ - TERMP_BOLD, /* TTYPE_SSECTION */ - TERMP_UNDERLINE, /* TTYPE_FILE */ - TERMP_UNDERLINE, /* TTYPE_EMPH */ - TERMP_BOLD, /* TTYPE_CONFIG */ - TERMP_BOLD, /* TTYPE_CMD */ - TERMP_BOLD, /* TTYPE_INCLUDE */ - TERMP_BOLD, /* TTYPE_SYMB */ - TERMP_BOLD, /* TTYPE_SYMBOL */ - TERMP_BOLD /* TTYPE_DIAG */ -}; - -static int arg_hasattr(int, const struct mdoc_node *); -static int arg_getattr(int, const struct mdoc_node *); -static size_t arg_offset(const struct mdoc_argv *); -static size_t arg_width(const struct mdoc_argv *); -static int arg_listtype(const struct mdoc_node *); - -/* - * What follows describes prefix and postfix operations for the abstract - * syntax tree descent. - */ - -#define DECL_ARGS \ - struct termp *p, \ - struct termpair *pair, \ - const struct mdoc_meta *meta, \ - const struct mdoc_node *node - -#define DECL_PRE(name) \ -static int name##_pre(DECL_ARGS) -#define DECL_POST(name) \ -static void name##_post(DECL_ARGS) -#define DECL_PREPOST(name) \ -DECL_PRE(name); \ -DECL_POST(name); - -DECL_PREPOST(termp__t); -DECL_PREPOST(termp_aq); -DECL_PREPOST(termp_bd); -DECL_PREPOST(termp_bq); -DECL_PREPOST(termp_d1); -DECL_PREPOST(termp_dq); -DECL_PREPOST(termp_fd); -DECL_PREPOST(termp_fn); -DECL_PREPOST(termp_fo); -DECL_PREPOST(termp_ft); -DECL_PREPOST(termp_in); -DECL_PREPOST(termp_it); -DECL_PREPOST(termp_op); -DECL_PREPOST(termp_pf); -DECL_PREPOST(termp_pq); -DECL_PREPOST(termp_qq); -DECL_PREPOST(termp_sh); -DECL_PREPOST(termp_ss); -DECL_PREPOST(termp_sq); -DECL_PREPOST(termp_vt); - -DECL_PRE(termp_ar); -DECL_PRE(termp_at); -DECL_PRE(termp_bf); -DECL_PRE(termp_bsx); -DECL_PRE(termp_bt); -DECL_PRE(termp_cd); -DECL_PRE(termp_cm); -DECL_PRE(termp_em); -DECL_PRE(termp_ex); -DECL_PRE(termp_fa); -DECL_PRE(termp_fl); -DECL_PRE(termp_fx); -DECL_PRE(termp_ic); -DECL_PRE(termp_ms); -DECL_PRE(termp_nd); -DECL_PRE(termp_nm); -DECL_PRE(termp_ns); -DECL_PRE(termp_nx); -DECL_PRE(termp_ox); -DECL_PRE(termp_pa); -DECL_PRE(termp_pp); -DECL_PRE(termp_rs); -DECL_PRE(termp_rv); -DECL_PRE(termp_sm); -DECL_PRE(termp_st); -DECL_PRE(termp_sx); -DECL_PRE(termp_sy); -DECL_PRE(termp_ud); -DECL_PRE(termp_ux); -DECL_PRE(termp_va); -DECL_PRE(termp_xr); - -DECL_POST(termp___); -DECL_POST(termp_bl); -DECL_POST(termp_bx); -DECL_POST(termp_lb); - -const struct termact __termacts[MDOC_MAX] = { - { NULL, NULL }, /* \" */ - { NULL, NULL }, /* Dd */ - { NULL, NULL }, /* Dt */ - { NULL, NULL }, /* Os */ - { termp_sh_pre, termp_sh_post }, /* Sh */ - { termp_ss_pre, termp_ss_post }, /* Ss */ - { termp_pp_pre, NULL }, /* Pp */ - { termp_d1_pre, termp_d1_post }, /* D1 */ - { termp_d1_pre, termp_d1_post }, /* Dl */ - { termp_bd_pre, termp_bd_post }, /* Bd */ - { NULL, NULL }, /* Ed */ - { NULL, termp_bl_post }, /* Bl */ - { NULL, NULL }, /* El */ - { termp_it_pre, termp_it_post }, /* It */ - { NULL, NULL }, /* Ad */ - { NULL, NULL }, /* An */ - { termp_ar_pre, NULL }, /* Ar */ - { termp_cd_pre, NULL }, /* Cd */ - { termp_cm_pre, NULL }, /* Cm */ - { NULL, NULL }, /* Dv */ - { NULL, NULL }, /* Er */ - { NULL, NULL }, /* Ev */ - { termp_ex_pre, NULL }, /* Ex */ - { termp_fa_pre, NULL }, /* Fa */ - { termp_fd_pre, termp_fd_post }, /* Fd */ - { termp_fl_pre, NULL }, /* Fl */ - { termp_fn_pre, termp_fn_post }, /* Fn */ - { termp_ft_pre, termp_ft_post }, /* Ft */ - { termp_ic_pre, NULL }, /* Ic */ - { termp_in_pre, termp_in_post }, /* In */ - { NULL, NULL }, /* Li */ - { termp_nd_pre, NULL }, /* Nd */ - { termp_nm_pre, NULL }, /* Nm */ - { termp_op_pre, termp_op_post }, /* Op */ - { NULL, NULL }, /* Ot */ - { termp_pa_pre, NULL }, /* Pa */ - { termp_rv_pre, NULL }, /* Rv */ - { termp_st_pre, NULL }, /* St */ - { termp_va_pre, NULL }, /* Va */ - { termp_vt_pre, termp_vt_post }, /* Vt */ - { termp_xr_pre, NULL }, /* Xr */ - { NULL, termp____post }, /* %A */ - { NULL, termp____post }, /* %B */ - { NULL, termp____post }, /* %D */ - { NULL, termp____post }, /* %I */ - { NULL, termp____post }, /* %J */ - { NULL, termp____post }, /* %N */ - { NULL, termp____post }, /* %O */ - { NULL, termp____post }, /* %P */ - { NULL, termp____post }, /* %R */ - { termp__t_pre, termp__t_post }, /* %T */ - { NULL, termp____post }, /* %V */ - { NULL, NULL }, /* Ac */ - { termp_aq_pre, termp_aq_post }, /* Ao */ - { termp_aq_pre, termp_aq_post }, /* Aq */ - { termp_at_pre, NULL }, /* At */ - { NULL, NULL }, /* Bc */ - { termp_bf_pre, NULL }, /* Bf */ - { termp_bq_pre, termp_bq_post }, /* Bo */ - { termp_bq_pre, termp_bq_post }, /* Bq */ - { termp_bsx_pre, NULL }, /* Bsx */ - { NULL, termp_bx_post }, /* Bx */ - { NULL, NULL }, /* Db */ - { NULL, NULL }, /* Dc */ - { termp_dq_pre, termp_dq_post }, /* Do */ - { termp_dq_pre, termp_dq_post }, /* Dq */ - { NULL, NULL }, /* Ec */ - { NULL, NULL }, /* Ef */ - { termp_em_pre, NULL }, /* Em */ - { NULL, NULL }, /* Eo */ - { termp_fx_pre, NULL }, /* Fx */ - { termp_ms_pre, NULL }, /* Ms */ - { NULL, NULL }, /* No */ - { termp_ns_pre, NULL }, /* Ns */ - { termp_nx_pre, NULL }, /* Nx */ - { termp_ox_pre, NULL }, /* Ox */ - { NULL, NULL }, /* Pc */ - { termp_pf_pre, termp_pf_post }, /* Pf */ - { termp_pq_pre, termp_pq_post }, /* Po */ - { termp_pq_pre, termp_pq_post }, /* Pq */ - { NULL, NULL }, /* Qc */ - { termp_sq_pre, termp_sq_post }, /* Ql */ - { termp_qq_pre, termp_qq_post }, /* Qo */ - { termp_qq_pre, termp_qq_post }, /* Qq */ - { NULL, NULL }, /* Re */ - { termp_rs_pre, NULL }, /* Rs */ - { NULL, NULL }, /* Sc */ - { termp_sq_pre, termp_sq_post }, /* So */ - { termp_sq_pre, termp_sq_post }, /* Sq */ - { termp_sm_pre, NULL }, /* Sm */ - { termp_sx_pre, NULL }, /* Sx */ - { termp_sy_pre, NULL }, /* Sy */ - { NULL, NULL }, /* Tn */ - { termp_ux_pre, NULL }, /* Ux */ - { NULL, NULL }, /* Xc */ - { NULL, NULL }, /* Xo */ - { termp_fo_pre, termp_fo_post }, /* Fo */ - { NULL, NULL }, /* Fc */ - { termp_op_pre, termp_op_post }, /* Oo */ - { NULL, NULL }, /* Oc */ - { NULL, NULL }, /* Bk */ - { NULL, NULL }, /* Ek */ - { termp_bt_pre, NULL }, /* Bt */ - { NULL, NULL }, /* Hf */ - { NULL, NULL }, /* Fr */ - { termp_ud_pre, NULL }, /* Ud */ - { NULL, termp_lb_post }, /* lb */ -}; - -const struct termact *termacts = __termacts; - - -static size_t -arg_width(const struct mdoc_argv *arg) +void +term_free(struct termp *p) { - size_t v; - int i, len; - - assert(*arg->value); - if (0 == strcmp(*arg->value, "indent")) - return(INDENT); - if (0 == strcmp(*arg->value, "indent-two")) - return(INDENT * 2); - - len = (int)strlen(*arg->value); - assert(len > 0); - - for (i = 0; i < len - 1; i++) - if ( ! isdigit((u_char)(*arg->value)[i])) - break; - - if (i == len - 1) { - if ('n' == (*arg->value)[len - 1]) { - v = (size_t)atoi(*arg->value); - return(v); - } - - } - return(strlen(*arg->value) + 1); + for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++) + free(p->tcol->buf); + free(p->tcols); + free(p->fontq); + free(p); } - -static int -arg_listtype(const struct mdoc_node *n) +void +term_begin(struct termp *p, term_margin head, + term_margin foot, const struct roff_meta *arg) { - int i, len; - assert(MDOC_BLOCK == n->type); - - len = n->args ? n->args->argc : 0; - - for (i = 0; i < len; i++) - switch (n->args->argv[i].arg) { - case (MDOC_Bullet): - /* FALLTHROUGH */ - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Enum): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - /* FALLTHROUGH */ - case (MDOC_Tag): - /* FALLTHROUGH */ - case (MDOC_Inset): - /* FALLTHROUGH */ - case (MDOC_Diag): - /* FALLTHROUGH */ - case (MDOC_Item): - /* FALLTHROUGH */ - case (MDOC_Ohang): - return(n->args->argv[i].arg); - default: - break; - } - - errx(1, "list type not supported"); - /* NOTREACHED */ + p->headf = head; + p->footf = foot; + p->argf = arg; + (*p->begin)(p); } - -static size_t -arg_offset(const struct mdoc_argv *arg) +void +term_end(struct termp *p) { - /* TODO */ - assert(*arg->value); - if (0 == strcmp(*arg->value, "indent")) - return(INDENT); - if (0 == strcmp(*arg->value, "indent-two")) - return(INDENT * 2); - return(strlen(*arg->value)); + (*p->end)(p); } - -static int -arg_hasattr(int arg, const struct mdoc_node *n) +/* + * Flush a chunk of text. By default, break the output line each time + * the right margin is reached, and continue output on the next line + * at the same offset as the chunk itself. By default, also break the + * output line at the end of the chunk. + * The following flags may be specified: + * + * - TERMP_NOBREAK: Do not break the output line at the right margin, + * but only at the max right margin. Also, do not break the output + * line at the end of the chunk, such that the next call can pad to + * the next column. However, if less than p->trailspace blanks, + * which can be 0, 1, or 2, remain to the right margin, the line + * will be broken. + * - TERMP_BRTRSP: Consider trailing whitespace significant + * when deciding whether the chunk fits or not. + * - TERMP_BRIND: If the chunk does not fit and the output line has + * to be broken, start the next line at the right margin instead + * of at the offset. Used together with TERMP_NOBREAK for the tags + * in various kinds of tagged lists. + * - TERMP_HANG: Do not break the output line at the right margin, + * append the next chunk after it even if this one is too long. + * To be used together with TERMP_NOBREAK. + * - TERMP_NOPAD: Start writing at the current position, + * do not pad with blank characters up to the offset. + */ +void +term_flushln(struct termp *p) { + size_t vis; /* current visual position on output */ + size_t vbl; /* number of blanks to prepend to output */ + size_t vend; /* end of word visual position on output */ + size_t bp; /* visual right border position */ + size_t dv; /* temporary for visual pos calculations */ + size_t j; /* temporary loop index for p->tcol->buf */ + size_t jhy; /* last hyph before overflow w/r/t j */ + size_t maxvis; /* output position of visible boundary */ + int ntab; /* number of tabs to prepend */ - return(-1 != arg_getattr(arg, n)); -} + vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ? + 0 : p->tcol->offset - p->viscol; + if (p->minbl && vbl < p->minbl) + vbl = p->minbl; + maxvis = p->tcol->rmargin > p->viscol + vbl ? + p->tcol->rmargin - p->viscol - vbl : 0; + bp = !(p->flags & TERMP_NOBREAK) ? maxvis : + p->maxrmargin > p->viscol + vbl ? + p->maxrmargin - p->viscol - vbl : 0; + vis = vend = 0; + if (p->lasttcol == 0) + p->tcol->col = 0; + while (p->tcol->col < p->lastcol) { -static int -arg_getattr(int arg, const struct mdoc_node *n) -{ - int i; + /* + * Handle literal tab characters: collapse all + * subsequent tabs into a single huge set of spaces. + */ - if (NULL == n->args) - return(-1); - for (i = 0; i < (int)n->args->argc; i++) - if (n->args->argv[i].arg == arg) - return(i); - return(-1); -} + ntab = 0; + while (p->tcol->col < p->lastcol && + p->tcol->buf[p->tcol->col] == '\t') { + vend = term_tab_next(vis); + vbl += vend - vis; + vis = vend; + ntab++; + p->tcol->col++; + } + /* + * Count up visible word characters. Control sequences + * (starting with the CSI) aren't counted. A space + * generates a non-printing word, which is valid (the + * space is printed according to regular spacing rules). + */ -/* ARGSUSED */ -static int -termp_dq_pre(DECL_ARGS) -{ + jhy = 0; + for (j = p->tcol->col; j < p->lastcol; j++) { + if (p->tcol->buf[j] == ' ' || p->tcol->buf[j] == '\t') + break; - if (MDOC_BODY != node->type) - return(1); + /* Back over the last printed character. */ + if (p->tcol->buf[j] == '\b') { + assert(j); + vend -= (*p->width)(p, p->tcol->buf[j - 1]); + continue; + } - word(p, "``"); - p->flags |= TERMP_NOSPACE; - return(1); -} + /* Regular word. */ + /* Break at the hyphen point if we overrun. */ + if (vend > vis && vend < bp && + (p->tcol->buf[j] == ASCII_HYPH|| + p->tcol->buf[j] == ASCII_BREAK)) + jhy = j; + /* + * Hyphenation now decided, put back a real + * hyphen such that we get the correct width. + */ + if (p->tcol->buf[j] == ASCII_HYPH) + p->tcol->buf[j] = '-'; -/* ARGSUSED */ -static void -termp_dq_post(DECL_ARGS) -{ + vend += (*p->width)(p, p->tcol->buf[j]); + } - if (MDOC_BODY != node->type) - return; + /* + * Find out whether we would exceed the right margin. + * If so, break to the next line. + */ - p->flags |= TERMP_NOSPACE; - word(p, "''"); -} + if (vend > bp && jhy == 0 && vis > 0 && + (p->flags & TERMP_BRNEVER) == 0) { + if (p->lasttcol) + return; + endline(p); + vend -= vis; -/* ARGSUSED */ -static int -termp_it_pre_block(DECL_ARGS) -{ + /* Use pending tabs on the new line. */ - newln(p); - if ( ! arg_hasattr(MDOC_Compact, node->parent->parent)) - if (node->prev || node->parent->parent->prev) - vspace(p); + vbl = 0; + while (ntab--) + vbl = term_tab_next(vbl); - return(1); -} + /* Re-establish indentation. */ + if (p->flags & TERMP_BRIND) + vbl += p->tcol->rmargin; + else + vbl += p->tcol->offset; + maxvis = p->tcol->rmargin > vbl ? + p->tcol->rmargin - vbl : 0; + bp = !(p->flags & TERMP_NOBREAK) ? maxvis : + p->maxrmargin > vbl ? p->maxrmargin - vbl : 0; + } -/* ARGSUSED */ -static int -termp_it_pre(DECL_ARGS) -{ - const struct mdoc_node *bl; - char buf[7]; - int i, type; - size_t width, offset; + /* + * Write out the rest of the word. + */ - if (MDOC_BLOCK == node->type) - return(termp_it_pre_block(p, pair, meta, node)); + for ( ; p->tcol->col < p->lastcol; p->tcol->col++) { + if (vend > bp && jhy > 0 && p->tcol->col > jhy) + break; + if (p->tcol->buf[p->tcol->col] == '\t') + break; + if (p->tcol->buf[p->tcol->col] == ' ') { + j = p->tcol->col; + while (p->tcol->col < p->lastcol && + p->tcol->buf[p->tcol->col] == ' ') + p->tcol->col++; + dv = (p->tcol->col - j) * (*p->width)(p, ' '); + vbl += dv; + vend += dv; + break; + } + if (p->tcol->buf[p->tcol->col] == ASCII_NBRSP) { + vbl += (*p->width)(p, ' '); + continue; + } + if (p->tcol->buf[p->tcol->col] == ASCII_BREAK) + continue; - /* Get ptr to list block, type, etc. */ + /* + * Now we definitely know there will be + * printable characters to output, + * so write preceding white space now. + */ + if (vbl) { + (*p->advance)(p, vbl); + p->viscol += vbl; + vbl = 0; + } - bl = node->parent->parent->parent; - type = arg_listtype(bl); - - /* Save parent attributes. */ - - pair->offset = p->offset; - pair->rmargin = p->rmargin; - pair->flag = p->flags; - - /* Get list width and offset. */ - - i = arg_getattr(MDOC_Width, bl); - width = i >= 0 ? arg_width(&bl->args->argv[i]) : 0; - - i = arg_getattr(MDOC_Offset, bl); - offset = i >= 0 ? arg_offset(&bl->args->argv[i]) : 0; - - /* - * List-type can override the width in the case of fixed-head - * values (bullet, dash/hyphen, enum). Tags need a non-zero - * offset. - */ - - switch (type) { - case (MDOC_Bullet): - /* FALLTHROUGH */ - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Enum): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - width = width > 4 ? width : 4; - break; - case (MDOC_Tag): - if (width) - break; - errx(1, "need non-zero %s for list type", - mdoc_argnames[MDOC_Width]); - default: - break; + (*p->letter)(p, p->tcol->buf[p->tcol->col]); + if (p->tcol->buf[p->tcol->col] == '\b') + p->viscol -= (*p->width)(p, + p->tcol->buf[p->tcol->col - 1]); + else + p->viscol += (*p->width)(p, + p->tcol->buf[p->tcol->col]); + } + vis = vend; } - /* - * Whitespace control. Inset bodies need an initial space. - */ - - switch (type) { - case (MDOC_Diag): - /* FALLTHROUGH */ - case (MDOC_Inset): - if (MDOC_BODY == node->type) - p->flags &= ~TERMP_NOSPACE; - else - p->flags |= TERMP_NOSPACE; - break; - default: - p->flags |= TERMP_NOSPACE; - break; - } - /* - * Style flags. Diagnostic heads need TTYPE_DIAG. + * If there was trailing white space, it was not printed; + * so reset the cursor position accordingly. */ - switch (type) { - case (MDOC_Diag): - if (MDOC_HEAD == node->type) - p->flags |= ttypes[TTYPE_DIAG]; - break; - default: - break; - } + if (vis > vbl) + vis -= vbl; + else + vis = 0; - /* - * Pad and break control. This is the tricker part. Lists with - * set right-margins for the head get TERMP_NOBREAK because, if - * they overrun the margin, they wrap to the new margin. - * Correspondingly, the body for these types don't left-pad, as - * the head will pad out to to the right. - */ + p->col = p->lastcol = 0; + p->minbl = p->trailspace; + p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD); - switch (type) { - case (MDOC_Bullet): - /* FALLTHROUGH */ - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Enum): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - /* FALLTHROUGH */ - case (MDOC_Tag): - if (MDOC_HEAD == node->type) - p->flags |= TERMP_NOBREAK; - else - p->flags |= TERMP_NOLPAD; - if (MDOC_HEAD == node->type && MDOC_Tag == type) - if (NULL == node->next || - NULL == node->next->child) - p->flags |= TERMP_NONOBREAK; - break; - case (MDOC_Diag): - if (MDOC_HEAD == node->type) - p->flags |= TERMP_NOBREAK; - break; - default: - break; - } + /* Trailing whitespace is significant in some columns. */ - /* - * Margin control. Set-head-width lists have their right - * margins shortened. The body for these lists has the offset - * necessarily lengthened. Everybody gets the offset. - */ + if (vis && vbl && (TERMP_BRTRSP & p->flags)) + vis += vbl; - p->offset += offset; - - switch (type) { - case (MDOC_Bullet): - /* FALLTHROUGH */ - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Enum): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - /* FALLTHROUGH */ - case (MDOC_Tag): - if (MDOC_HEAD == node->type) - p->rmargin = p->offset + width; - else - p->offset += width; - /* FALLTHROUGH */ - default: - break; - } - - /* - * The dash, hyphen, bullet and enum lists all have a special - * HEAD character. Print it now. - */ - - if (MDOC_HEAD == node->type) - switch (type) { - case (MDOC_Bullet): - word(p, "\\[bu]"); - break; - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - word(p, "\\-"); - break; - case (MDOC_Enum): - /* TODO: have a wordfmt or something. */ - (pair->ppair->ppair->count)++; - (void)snprintf(buf, sizeof(buf), "%d.", - pair->ppair->ppair->count); - word(p, buf); - break; - default: - break; - } - - /* - * If we're not going to process our header children, indicate - * so here. - */ - - if (MDOC_HEAD == node->type) - switch (type) { - case (MDOC_Bullet): - /* FALLTHROUGH */ - case (MDOC_Item): - /* FALLTHROUGH */ - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - /* FALLTHROUGH */ - case (MDOC_Enum): - return(0); - default: - break; - } - - return(1); + /* If the column was overrun, break the line. */ + if ((p->flags & TERMP_NOBREAK) == 0 || + ((p->flags & TERMP_HANG) == 0 && + vis + p->trailspace * (*p->width)(p, ' ') > maxvis)) + endline(p); } - -/* ARGSUSED */ static void -termp_it_post(DECL_ARGS) +endline(struct termp *p) { - int type; - - if (MDOC_BODY != node->type && MDOC_HEAD != node->type) - return; - - type = arg_listtype(node->parent->parent->parent); - - switch (type) { - case (MDOC_Diag): - /* FALLTHROUGH */ - case (MDOC_Item): - /* FALLTHROUGH */ - case (MDOC_Inset): - if (MDOC_BODY != node->type) - break; - flushln(p); - break; - default: - flushln(p); - break; + if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) { + p->mc = NULL; + p->flags &= ~TERMP_ENDMC; } - - p->offset = pair->offset; - p->rmargin = pair->rmargin; - p->flags = pair->flag; + if (p->mc != NULL) { + if (p->viscol && p->maxrmargin >= p->viscol) + (*p->advance)(p, p->maxrmargin - p->viscol + 1); + p->flags |= TERMP_NOBUF | TERMP_NOSPACE; + term_word(p, p->mc); + p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC); + } + p->viscol = 0; + p->minbl = 0; + (*p->endline)(p); } - -/* ARGSUSED */ -static int -termp_nm_pre(DECL_ARGS) +/* + * A newline only breaks an existing line; it won't assert vertical + * space. All data in the output buffer is flushed prior to the newline + * assertion. + */ +void +term_newln(struct termp *p) { - if (SEC_SYNOPSIS == node->sec) - newln(p); - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_PROG]); - if (NULL == node->child) - word(p, meta->name); - - return(1); -} - - -/* ARGSUSED */ -static int -termp_fl_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_CMD_FLAG]); - word(p, "\\-"); p->flags |= TERMP_NOSPACE; - return(1); + if (p->lastcol || p->viscol) + term_flushln(p); } - -/* ARGSUSED */ -static int -termp_ar_pre(DECL_ARGS) +/* + * Asserts a vertical space (a full, empty line-break between lines). + * Note that if used twice, this will cause two blank spaces and so on. + * All data in the output buffer is flushed prior to the newline + * assertion. + */ +void +term_vspace(struct termp *p) { - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_CMD_ARG]); - return(1); + term_newln(p); + p->viscol = 0; + p->minbl = 0; + if (0 < p->skipvsp) + p->skipvsp--; + else + (*p->endline)(p); } - -/* ARGSUSED */ -static int -termp_ns_pre(DECL_ARGS) +/* Swap current and previous font; for \fP and .ft P */ +void +term_fontlast(struct termp *p) { + enum termfont f; - p->flags |= TERMP_NOSPACE; - return(1); + f = p->fontl; + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } - -/* ARGSUSED */ -static int -termp_pp_pre(DECL_ARGS) +/* Set font, save current, discard previous; for \f, .ft, .B etc. */ +void +term_fontrepl(struct termp *p, enum termfont f) { - vspace(p); - return(1); + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } - -/* ARGSUSED */ -static int -termp_st_pre(DECL_ARGS) +/* Set font, save previous. */ +void +term_fontpush(struct termp *p, enum termfont f) { - const char *cp; - if (node->child) { - if (MDOC_TEXT != node->child->type) - errx(1, "expected text line arguments"); - if ((cp = mdoc_a2st(node->child->string))) - word(p, cp); + p->fontl = p->fontq[p->fonti]; + if (++p->fonti == p->fontsz) { + p->fontsz += 8; + p->fontq = mandoc_reallocarray(p->fontq, + p->fontsz, sizeof(*p->fontq)); } - return(0); + p->fontq[p->fonti] = f; } - -/* ARGSUSED */ -static int -termp_rs_pre(DECL_ARGS) +/* Flush to make the saved pointer current again. */ +void +term_fontpopq(struct termp *p, int i) { - if (MDOC_BLOCK == node->type && node->prev) - vspace(p); - return(1); + assert(i >= 0); + if (p->fonti > i) + p->fonti = i; } - -/* ARGSUSED */ -static int -termp_rv_pre(DECL_ARGS) +/* Pop one font off the stack. */ +void +term_fontpop(struct termp *p) { - int i; - if (-1 == (i = arg_getattr(MDOC_Std, node))) - errx(1, "expected -std argument"); - if (1 != node->args->argv[i].sz) - errx(1, "expected -std argument"); - - newln(p); - word(p, "The"); - - p->flags |= ttypes[TTYPE_FUNC_NAME]; - word(p, *node->args->argv[i].value); - p->flags &= ~ttypes[TTYPE_FUNC_NAME]; - - word(p, "() function returns the value 0 if successful;"); - word(p, "otherwise the value -1 is returned and the"); - word(p, "global variable"); - - p->flags |= ttypes[TTYPE_VAR_DECL]; - word(p, "errno"); - p->flags &= ~ttypes[TTYPE_VAR_DECL]; - - word(p, "is set to indicate the error."); - - return(1); + assert(p->fonti); + p->fonti--; } - -/* ARGSUSED */ -static int -termp_ex_pre(DECL_ARGS) +/* + * Handle pwords, partial words, which may be either a single word or a + * phrase that cannot be broken down (such as a literal string). This + * handles word styling. + */ +void +term_word(struct termp *p, const char *word) { - int i; + struct roffsu su; + const char nbrsp[2] = { ASCII_NBRSP, 0 }; + const char *seq, *cp; + int sz, uc; + size_t csz, lsz, ssz; + enum mandoc_esc esc; - if (-1 == (i = arg_getattr(MDOC_Std, node))) - errx(1, "expected -std argument"); - if (1 != node->args->argv[i].sz) - errx(1, "expected -std argument"); + if ((p->flags & TERMP_NOBUF) == 0) { + if ((p->flags & TERMP_NOSPACE) == 0) { + if ((p->flags & TERMP_KEEP) == 0) { + bufferc(p, ' '); + if (p->flags & TERMP_SENTENCE) + bufferc(p, ' '); + } else + bufferc(p, ASCII_NBRSP); + } + if (p->flags & TERMP_PREKEEP) + p->flags |= TERMP_KEEP; + if (p->flags & TERMP_NONOSPACE) + p->flags |= TERMP_NOSPACE; + else + p->flags &= ~TERMP_NOSPACE; + p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE); + p->skipvsp = 0; + } - word(p, "The"); - p->flags |= ttypes[TTYPE_PROG]; - word(p, *node->args->argv[i].value); - p->flags &= ~ttypes[TTYPE_PROG]; - word(p, "utility exits 0 on success, and >0 if an error occurs."); + while ('\0' != *word) { + if ('\\' != *word) { + if (TERMP_NBRWORD & p->flags) { + if (' ' == *word) { + encode(p, nbrsp, 1); + word++; + continue; + } + ssz = strcspn(word, "\\ "); + } else + ssz = strcspn(word, "\\"); + encode(p, word, ssz); + word += (int)ssz; + continue; + } - return(1); -} + word++; + esc = mandoc_escape(&word, &seq, &sz); + if (ESCAPE_ERROR == esc) + continue; + switch (esc) { + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, sz - 1); + break; + case ESCAPE_NUMBERED: + uc = mchars_num2char(seq, sz); + if (uc < 0) + continue; + break; + case ESCAPE_SPECIAL: + if (p->enc == TERMENC_ASCII) { + cp = mchars_spec2str(seq, sz, &ssz); + if (cp != NULL) + encode(p, cp, ssz); + } else { + uc = mchars_spec2cp(seq, sz); + if (uc > 0) + encode1(p, uc); + } + continue; + case ESCAPE_FONTBOLD: + term_fontrepl(p, TERMFONT_BOLD); + continue; + case ESCAPE_FONTITALIC: + term_fontrepl(p, TERMFONT_UNDER); + continue; + case ESCAPE_FONTBI: + term_fontrepl(p, TERMFONT_BI); + continue; + case ESCAPE_FONT: + case ESCAPE_FONTROMAN: + term_fontrepl(p, TERMFONT_NONE); + continue; + case ESCAPE_FONTPREV: + term_fontlast(p); + continue; + case ESCAPE_NOSPACE: + if (p->flags & TERMP_BACKAFTER) + p->flags &= ~TERMP_BACKAFTER; + else if (*word == '\0') + p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE); + continue; + case ESCAPE_HORIZ: + if (a2roffsu(seq, &su, SCALE_EM) == 0) + continue; + uc = term_hspan(p, &su) / 24; + if (uc > 0) + while (uc-- > 0) + bufferc(p, ASCII_NBRSP); + else if (p->col > (size_t)(-uc)) + p->col += uc; + else { + uc += p->col; + p->col = 0; + if (p->tcol->offset > (size_t)(-uc)) { + p->ti += uc; + p->tcol->offset += uc; + } else { + p->ti -= p->tcol->offset; + p->tcol->offset = 0; + } + } + continue; + case ESCAPE_HLINE: + if (a2roffsu(seq, &su, SCALE_EM) == 0) + continue; + uc = term_hspan(p, &su) / 24; + if (uc <= 0) { + if (p->tcol->rmargin <= p->tcol->offset) + continue; + lsz = p->tcol->rmargin - p->tcol->offset; + } else + lsz = uc; + while (sz && + strchr(" %&()*+-./0123456789:<=>", *seq)) { + seq++; + sz--; + } + if (sz && strchr("cifMmnPpuv", *seq)) { + seq++; + sz--; + } + if (sz == 0) + uc = -1; + else if (*seq == '\\') { + seq++; + esc = mandoc_escape(&seq, &cp, &sz); + switch (esc) { + case ESCAPE_UNICODE: + uc = mchars_num2uc(cp + 1, sz - 1); + break; + case ESCAPE_NUMBERED: + uc = mchars_num2char(cp, sz); + break; + case ESCAPE_SPECIAL: + uc = mchars_spec2cp(cp, sz); + break; + default: + uc = -1; + break; + } + } else + uc = *seq; + if (uc < 0x20 || (uc > 0x7E && uc < 0xA0)) + uc = '_'; + if (p->enc == TERMENC_ASCII) { + cp = ascii_uc2str(uc); + csz = term_strlen(p, cp); + ssz = strlen(cp); + } else + csz = (*p->width)(p, uc); + while (lsz >= csz) { + if (p->enc == TERMENC_ASCII) + encode(p, cp, ssz); + else + encode1(p, uc); + lsz -= csz; + } + continue; + case ESCAPE_SKIPCHAR: + p->flags |= TERMP_BACKAFTER; + continue; + case ESCAPE_OVERSTRIKE: + cp = seq + sz; + while (seq < cp) { + if (*seq == '\\') { + mandoc_escape(&seq, NULL, NULL); + continue; + } + encode1(p, *seq++); + if (seq < cp) { + if (p->flags & TERMP_BACKBEFORE) + p->flags |= TERMP_BACKAFTER; + else + p->flags |= TERMP_BACKBEFORE; + } + } + /* Trim trailing backspace/blank pair. */ + if (p->lastcol > 2 && + (p->tcol->buf[p->lastcol - 1] == ' ' || + p->tcol->buf[p->lastcol - 1] == '\t')) + p->lastcol -= 2; + if (p->col > p->lastcol) + p->col = p->lastcol; + continue; + default: + continue; + } -/* ARGSUSED */ -static int -termp_nd_pre(DECL_ARGS) -{ + /* + * Common handling for Unicode and numbered + * character escape sequences. + */ - word(p, "\\-"); - return(1); + if (p->enc == TERMENC_ASCII) { + cp = ascii_uc2str(uc); + encode(p, cp, strlen(cp)); + } else { + if ((uc < 0x20 && uc != 0x09) || + (uc > 0x7E && uc < 0xA0)) + uc = 0xFFFD; + encode1(p, uc); + } + } + p->flags &= ~TERMP_NBRWORD; } - -/* ARGSUSED */ static void -termp_bl_post(DECL_ARGS) +adjbuf(struct termp_col *c, size_t sz) { - - if (MDOC_BLOCK == node->type) - newln(p); + if (c->maxcols == 0) + c->maxcols = 1024; + while (c->maxcols <= sz) + c->maxcols <<= 2; + c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf)); } - -/* ARGSUSED */ static void -termp_op_post(DECL_ARGS) +bufferc(struct termp *p, char c) { - - if (MDOC_BODY != node->type) + if (p->flags & TERMP_NOBUF) { + (*p->letter)(p, c); return; - p->flags |= TERMP_NOSPACE; - word(p, "\\(rB"); + } + if (p->col + 1 >= p->tcol->maxcols) + adjbuf(p->tcol, p->col + 1); + if (p->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP)) + p->tcol->buf[p->col] = c; + if (p->lastcol < ++p->col) + p->lastcol = p->col; } - -/* ARGSUSED */ -static int -termp_xr_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - - if (NULL == (n = node->child)) - errx(1, "expected text line argument"); - if (MDOC_TEXT != n->type) - errx(1, "expected text line argument"); - - word(p, n->string); - - if (NULL == (n = n->next)) - return(0); - if (MDOC_TEXT != n->type) - errx(1, "expected text line argument"); - - p->flags |= TERMP_NOSPACE; - word(p, "("); - p->flags |= TERMP_NOSPACE; - word(p, n->string); - p->flags |= TERMP_NOSPACE; - word(p, ")"); - - return(0); -} - - -/* ARGSUSED */ -static int -termp_vt_pre(DECL_ARGS) -{ - - /* FIXME: this can be "type name". */ - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_VAR_DECL]); - return(1); -} - - -/* ARGSUSED */ +/* + * See encode(). + * Do this for a single (probably unicode) value. + * Does not check for non-decorated glyphs. + */ static void -termp_vt_post(DECL_ARGS) +encode1(struct termp *p, int c) { + enum termfont f; - if (node->sec == SEC_SYNOPSIS) - vspace(p); -} - - -/* ARGSUSED */ -static int -termp_fd_pre(DECL_ARGS) -{ - - /* - * FIXME: this naming is bad. This value is used, in general, - * for the #include header or other preprocessor statement. - */ - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_FUNC_DECL]); - return(1); -} - - -/* ARGSUSED */ -static void -termp_fd_post(DECL_ARGS) -{ - - if (node->sec != SEC_SYNOPSIS) + if (p->flags & TERMP_NOBUF) { + (*p->letter)(p, c); return; - newln(p); - if (node->next && MDOC_Fd != node->next->tok) - vspace(p); -} - - -/* ARGSUSED */ -static int -termp_sh_pre(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_HEAD): - vspace(p); - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_SECTION]); - break; - case (MDOC_BODY): - p->offset = INDENT; - break; - default: - break; } - return(1); -} + if (p->col + 7 >= p->tcol->maxcols) + adjbuf(p->tcol, p->col + 7); -/* ARGSUSED */ -static void -termp_sh_post(DECL_ARGS) -{ + f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ? + p->fontq[p->fonti] : TERMFONT_NONE; - switch (node->type) { - case (MDOC_HEAD): - newln(p); - break; - case (MDOC_BODY): - newln(p); - p->offset = 0; - break; - default: - break; + if (p->flags & TERMP_BACKBEFORE) { + if (p->tcol->buf[p->col - 1] == ' ' || + p->tcol->buf[p->col - 1] == '\t') + p->col--; + else + p->tcol->buf[p->col++] = '\b'; + p->flags &= ~TERMP_BACKBEFORE; } -} - - -/* ARGSUSED */ -static int -termp_op_pre(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_BODY): - word(p, "\\(lB"); - p->flags |= TERMP_NOSPACE; - break; - default: - break; + if (f == TERMFONT_UNDER || f == TERMFONT_BI) { + p->tcol->buf[p->col++] = '_'; + p->tcol->buf[p->col++] = '\b'; } - return(1); + if (f == TERMFONT_BOLD || f == TERMFONT_BI) { + if (c == ASCII_HYPH) + p->tcol->buf[p->col++] = '-'; + else + p->tcol->buf[p->col++] = c; + p->tcol->buf[p->col++] = '\b'; + } + if (p->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP)) + p->tcol->buf[p->col] = c; + if (p->lastcol < ++p->col) + p->lastcol = p->col; + if (p->flags & TERMP_BACKAFTER) { + p->flags |= TERMP_BACKBEFORE; + p->flags &= ~TERMP_BACKAFTER; + } } - -/* ARGSUSED */ -static int -termp_bt_pre(DECL_ARGS) -{ - - word(p, "is currently in beta test."); - return(1); -} - - -/* ARGSUSED */ static void -termp_lb_post(DECL_ARGS) +encode(struct termp *p, const char *word, size_t sz) { + size_t i; - newln(p); -} - - -/* ARGSUSED */ -static int -termp_ud_pre(DECL_ARGS) -{ - - word(p, "currently under development."); - return(1); -} - - -/* ARGSUSED */ -static int -termp_d1_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - newln(p); - p->offset += (pair->offset = INDENT); - return(1); -} - - -/* ARGSUSED */ -static void -termp_d1_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) + if (p->flags & TERMP_NOBUF) { + for (i = 0; i < sz; i++) + (*p->letter)(p, word[i]); return; - newln(p); - p->offset -= pair->offset; -} - - -/* ARGSUSED */ -static int -termp_aq_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - word(p, "\\(la"); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp_aq_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - p->flags |= TERMP_NOSPACE; - word(p, "\\(ra"); -} - - -/* ARGSUSED */ -static int -termp_ft_pre(DECL_ARGS) -{ - - if (SEC_SYNOPSIS == node->sec) - if (node->prev && MDOC_Fo == node->prev->tok) - vspace(p); - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_FUNC_TYPE]); - return(1); -} - - -/* ARGSUSED */ -static void -termp_ft_post(DECL_ARGS) -{ - - if (SEC_SYNOPSIS == node->sec) - newln(p); -} - - -/* ARGSUSED */ -static int -termp_fn_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - - if (NULL == node->child) - errx(1, "expected text line arguments"); - if (MDOC_TEXT != node->child->type) - errx(1, "expected text line arguments"); - - /* FIXME: can be "type funcname" "type varname"... */ - - p->flags |= ttypes[TTYPE_FUNC_NAME]; - word(p, node->child->string); - p->flags &= ~ttypes[TTYPE_FUNC_NAME]; - - word(p, "("); - - p->flags |= TERMP_NOSPACE; - for (n = node->child->next; n; n = n->next) { - if (MDOC_TEXT != n->type) - errx(1, "expected text line arguments"); - p->flags |= ttypes[TTYPE_FUNC_ARG]; - word(p, n->string); - p->flags &= ~ttypes[TTYPE_FUNC_ARG]; - if (n->next) - word(p, ","); } - word(p, ")"); + if (p->col + 2 + (sz * 5) >= p->tcol->maxcols) + adjbuf(p->tcol, p->col + 2 + (sz * 5)); - if (SEC_SYNOPSIS == node->sec) - word(p, ";"); + for (i = 0; i < sz; i++) { + if (ASCII_HYPH == word[i] || + isgraph((unsigned char)word[i])) + encode1(p, word[i]); + else { + if (p->lastcol <= p->col || + (word[i] != ' ' && word[i] != ASCII_NBRSP)) + p->tcol->buf[p->col] = word[i]; + p->col++; - return(0); -} + /* + * Postpone the effect of \z while handling + * an overstrike sequence from ascii_uc2str(). + */ - -/* ARGSUSED */ -static void -termp_fn_post(DECL_ARGS) -{ - - if (node->sec == SEC_SYNOPSIS && node->next) - vspace(p); - -} - - -/* ARGSUSED */ -static int -termp_sx_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_LINK]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_fa_pre(DECL_ARGS) -{ - struct mdoc_node *n; - - if (node->parent->tok != MDOC_Fo) { - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_FUNC_ARG]); - return(1); + if (word[i] == '\b' && + (p->flags & TERMP_BACKBEFORE)) { + p->flags &= ~TERMP_BACKBEFORE; + p->flags |= TERMP_BACKAFTER; + } + } } - - for (n = node->child; n; n = n->next) { - if (MDOC_TEXT != n->type) - errx(1, "expected text line arguments"); - - p->flags |= ttypes[TTYPE_FUNC_ARG]; - word(p, n->string); - p->flags &= ~ttypes[TTYPE_FUNC_ARG]; - - if (n->next) - word(p, ","); - } - - if (node->next && node->next->tok == MDOC_Fa) - word(p, ","); - - return(0); + if (p->lastcol < p->col) + p->lastcol = p->col; } - -/* ARGSUSED */ -static int -termp_va_pre(DECL_ARGS) +void +term_setwidth(struct termp *p, const char *wstr) { + struct roffsu su; + int iop, width; - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_VAR_DECL]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_bd_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - int i, type; - - if (MDOC_BLOCK == node->type) { - if (node->prev) - vspace(p); - return(1); - } else if (MDOC_BODY != node->type) - return(1); - - if (NULL == node->parent->args) - errx(1, "missing display type"); - - pair->offset = p->offset; - - for (type = -1, i = 0; - i < (int)node->parent->args->argc; i++) { - switch (node->parent->args->argv[i].arg) { - case (MDOC_Ragged): - /* FALLTHROUGH */ - case (MDOC_Filled): - /* FALLTHROUGH */ - case (MDOC_Unfilled): - /* FALLTHROUGH */ - case (MDOC_Literal): - type = node->parent->args->argv[i].arg; - i = (int)node->parent->args->argc; + iop = 0; + width = 0; + if (NULL != wstr) { + switch (*wstr) { + case '+': + iop = 1; + wstr++; break; + case '-': + iop = -1; + wstr++; + break; default: break; } + if (a2roffsu(wstr, &su, SCALE_MAX)) + width = term_hspan(p, &su); + else + iop = 0; } - - if (NULL == node->parent->args) - errx(1, "missing display type"); - - i = arg_getattr(MDOC_Offset, node->parent); - if (-1 != i) { - if (1 != node->parent->args->argv[i].sz) - errx(1, "expected single value"); - p->offset += arg_offset(&node->parent->args->argv[i]); - } - - switch (type) { - case (MDOC_Literal): - /* FALLTHROUGH */ - case (MDOC_Unfilled): - break; - default: - return(1); - } - - p->flags |= TERMP_LITERAL; - - for (n = node->child; n; n = n->next) { - if (MDOC_TEXT != n->type) { - warnx("non-text children not yet allowed"); - continue; - } - word(p, n->string); - flushln(p); - } - - return(0); + (*p->setwidth)(p, iop, width); } - -/* ARGSUSED */ -static void -termp_bd_post(DECL_ARGS) +size_t +term_len(const struct termp *p, size_t sz) { - if (MDOC_BODY != node->type) - return; - - if ( ! (p->flags & TERMP_LITERAL)) - flushln(p); - - p->flags &= ~TERMP_LITERAL; - p->offset = pair->offset; + return (*p->width)(p, ' ') * sz; } - -/* ARGSUSED */ -static int -termp_qq_pre(DECL_ARGS) +static size_t +cond_width(const struct termp *p, int c, int *skip) { - if (MDOC_BODY != node->type) - return(1); - word(p, "\""); - p->flags |= TERMP_NOSPACE; - return(1); + if (*skip) { + (*skip) = 0; + return 0; + } else + return (*p->width)(p, c); } - -/* ARGSUSED */ -static void -termp_qq_post(DECL_ARGS) +size_t +term_strlen(const struct termp *p, const char *cp) { + size_t sz, rsz, i; + int ssz, skip, uc; + const char *seq, *rhs; + enum mandoc_esc esc; + static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH, + ASCII_BREAK, '\0' }; - if (MDOC_BODY != node->type) - return; - p->flags |= TERMP_NOSPACE; - word(p, "\""); -} + /* + * Account for escaped sequences within string length + * calculations. This follows the logic in term_word() as we + * must calculate the width of produced strings. + */ + sz = 0; + skip = 0; + while ('\0' != *cp) { + rsz = strcspn(cp, rej); + for (i = 0; i < rsz; i++) + sz += cond_width(p, *cp++, &skip); -/* ARGSUSED */ -static int -termp_bsx_pre(DECL_ARGS) -{ + switch (*cp) { + case '\\': + cp++; + esc = mandoc_escape(&cp, &seq, &ssz); + if (ESCAPE_ERROR == esc) + continue; - word(p, "BSDI BSD/OS"); - return(1); -} + rhs = NULL; + switch (esc) { + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, ssz - 1); + break; + case ESCAPE_NUMBERED: + uc = mchars_num2char(seq, ssz); + if (uc < 0) + continue; + break; + case ESCAPE_SPECIAL: + if (p->enc == TERMENC_ASCII) { + rhs = mchars_spec2str(seq, ssz, &rsz); + if (rhs != NULL) + break; + } else { + uc = mchars_spec2cp(seq, ssz); + if (uc > 0) + sz += cond_width(p, uc, &skip); + } + continue; + case ESCAPE_SKIPCHAR: + skip = 1; + continue; + case ESCAPE_OVERSTRIKE: + rsz = 0; + rhs = seq + ssz; + while (seq < rhs) { + if (*seq == '\\') { + mandoc_escape(&seq, NULL, NULL); + continue; + } + i = (*p->width)(p, *seq++); + if (rsz < i) + rsz = i; + } + sz += rsz; + continue; + default: + continue; + } -/* ARGSUSED */ -static void -termp_bx_post(DECL_ARGS) -{ + /* + * Common handling for Unicode and numbered + * character escape sequences. + */ - if (node->child) - p->flags |= TERMP_NOSPACE; - word(p, "BSD"); -} + if (rhs == NULL) { + if (p->enc == TERMENC_ASCII) { + rhs = ascii_uc2str(uc); + rsz = strlen(rhs); + } else { + if ((uc < 0x20 && uc != 0x09) || + (uc > 0x7E && uc < 0xA0)) + uc = 0xFFFD; + sz += cond_width(p, uc, &skip); + continue; + } + } + if (skip) { + skip = 0; + break; + } -/* ARGSUSED */ -static int -termp_ox_pre(DECL_ARGS) -{ + /* + * Common handling for all escape sequences + * printing more than one character. + */ - word(p, "OpenBSD"); - return(1); -} + for (i = 0; i < rsz; i++) + sz += (*p->width)(p, *rhs++); + break; + case ASCII_NBRSP: + sz += cond_width(p, ' ', &skip); + cp++; + break; + case ASCII_HYPH: + sz += cond_width(p, '-', &skip); + cp++; + break; + default: + break; + } + } - -/* ARGSUSED */ -static int -termp_ux_pre(DECL_ARGS) -{ - - word(p, "UNIX"); - return(1); + return sz; } - -/* ARGSUSED */ -static int -termp_fx_pre(DECL_ARGS) +int +term_vspan(const struct termp *p, const struct roffsu *su) { + double r; + int ri; - word(p, "FreeBSD"); - return(1); -} - - -/* ARGSUSED */ -static int -termp_nx_pre(DECL_ARGS) -{ - - word(p, "NetBSD"); - return(1); -} - - -/* ARGSUSED */ -static int -termp_sq_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - word(p, "`"); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp_sq_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - p->flags |= TERMP_NOSPACE; - word(p, "\'"); -} - - -/* ARGSUSED */ -static int -termp_pf_pre(DECL_ARGS) -{ - - p->flags |= TERMP_IGNDELIM; - return(1); -} - - -/* ARGSUSED */ -static void -termp_pf_post(DECL_ARGS) -{ - - p->flags &= ~TERMP_IGNDELIM; - p->flags |= TERMP_NOSPACE; -} - - -/* ARGSUSED */ -static int -termp_ss_pre(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_HEAD): - vspace(p); - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_SSECTION]); - p->offset = INDENT / 2; + switch (su->unit) { + case SCALE_BU: + r = su->scale / 40.0; break; - default: + case SCALE_CM: + r = su->scale * 6.0 / 2.54; break; - } - - return(1); -} - - -/* ARGSUSED */ -static void -termp_ss_post(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_HEAD): - newln(p); - p->offset = INDENT; + case SCALE_FS: + r = su->scale * 65536.0 / 40.0; break; - default: + case SCALE_IN: + r = su->scale * 6.0; break; + case SCALE_MM: + r = su->scale * 0.006; + break; + case SCALE_PC: + r = su->scale; + break; + case SCALE_PT: + r = su->scale / 12.0; + break; + case SCALE_EN: + case SCALE_EM: + r = su->scale * 0.6; + break; + case SCALE_VS: + r = su->scale; + break; + default: + abort(); } + ri = r > 0.0 ? r + 0.4995 : r - 0.4995; + return ri < 66 ? ri : 1; } - -/* ARGSUSED */ -static int -termp_pa_pre(DECL_ARGS) +/* + * Convert a scaling width to basic units, rounding down. + */ +int +term_hspan(const struct termp *p, const struct roffsu *su) { - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_FILE]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_em_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_EMPH]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_cd_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_CONFIG]); - newln(p); - return(1); -} - - -/* ARGSUSED */ -static int -termp_cm_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_CMD_FLAG]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_ic_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_CMD]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_in_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_INCLUDE]); - word(p, "#include"); - word(p, "<"); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp_in_post(DECL_ARGS) -{ - - p->flags |= TERMP_NOSPACE; - word(p, ">"); - - newln(p); - if (SEC_SYNOPSIS != node->sec) - return; - if (node->next && MDOC_In != node->next->tok) - vspace(p); -} - - -/* ARGSUSED */ -static int -termp_at_pre(DECL_ARGS) -{ - const char *att; - - att = NULL; - - if (node->child) { - if (MDOC_TEXT != node->child->type) - errx(1, "expected text line argument"); - att = mdoc_a2att(node->child->string); - } - - if (NULL == att) - att = "AT&T UNIX"; - - word(p, att); - return(0); -} - - -/* ARGSUSED */ -static int -termp_bq_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - word(p, "["); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp_bq_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - word(p, "]"); -} - - -/* ARGSUSED */ -static int -termp_pq_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - word(p, "\\&("); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp_pq_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - word(p, ")"); -} - - -/* ARGSUSED */ -static int -termp_fo_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - - if (MDOC_BODY == node->type) { - word(p, "("); - p->flags |= TERMP_NOSPACE; - return(1); - } else if (MDOC_HEAD != node->type) - return(1); - - /* XXX - groff shows only first parameter */ - - p->flags |= ttypes[TTYPE_FUNC_NAME]; - for (n = node->child; n; n = n->next) { - if (MDOC_TEXT != n->type) - errx(1, "expected text line argument"); - word(p, n->string); - } - p->flags &= ~ttypes[TTYPE_FUNC_NAME]; - - return(0); -} - - -/* ARGSUSED */ -static void -termp_fo_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - word(p, ")"); - word(p, ";"); - newln(p); -} - - -/* ARGSUSED */ -static int -termp_bf_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - - if (MDOC_HEAD == node->type) { - return(0); - } else if (MDOC_BLOCK != node->type) - return(1); - - if (NULL == (n = node->head->child)) { - if (arg_hasattr(MDOC_Emphasis, node)) - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_EMPH]); - else if (arg_hasattr(MDOC_Symbolic, node)) - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_SYMB]); - - return(1); - } - - if (MDOC_TEXT != n->type) - errx(1, "expected text line arguments"); - - if (0 == strcmp("Em", n->string)) - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_EMPH]); - else if (0 == strcmp("Sy", n->string)) - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_EMPH]); - - return(1); -} - - -/* ARGSUSED */ -static int -termp_sy_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_SYMB]); - return(1); -} - - -/* ARGSUSED */ -static int -termp_ms_pre(DECL_ARGS) -{ - - TERMPAIR_SETFLAG(p, pair, ttypes[TTYPE_SYMBOL]); - return(1); -} - - - -/* ARGSUSED */ -static int -termp_sm_pre(DECL_ARGS) -{ - -#if notyet - assert(node->child); - if (0 == strcmp("off", node->child->data.text.string)) { - p->flags &= ~TERMP_NONOSPACE; - p->flags &= ~TERMP_NOSPACE; - } else { - p->flags |= TERMP_NONOSPACE; - p->flags |= TERMP_NOSPACE; - } -#endif - - return(0); -} - - -/* ARGSUSED */ -static int -termp__t_pre(DECL_ARGS) -{ - - /* FIXME: titles are underlined. */ - word(p, "\""); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp__t_post(DECL_ARGS) -{ - - p->flags |= TERMP_NOSPACE; - /* FIXME: titles are underlined. */ - word(p, "\""); - word(p, node->next ? "," : "."); -} - - -/* ARGSUSED */ -static void -termp____post(DECL_ARGS) -{ - - p->flags |= TERMP_NOSPACE; - word(p, node->next ? "," : "."); + return (*p->hspan)(p, su); }