=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.16 retrieving revision 1.83 diff -u -p -r1.16 -r1.83 --- mandoc/term.c 2009/02/24 16:16:45 1.16 +++ mandoc/term.c 2009/06/22 13:13:10 1.83 @@ -1,1457 +1,577 @@ -/* $Id: term.c,v 1.16 2009/02/24 16:16:45 kristaps Exp $ */ +/* $Id: term.c,v 1.83 2009/06/22 13:13:10 kristaps Exp $ */ /* - * Copyright (c) 2009 Kristaps Dzonsons + * Copyright (c) 2008, 2009 Kristaps Dzonsons * * 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 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. */ #include +#include +#include #include #include #include "term.h" +#include "man.h" +#include "mdoc.h" -#define INDENT 4 +extern int man_run(struct termp *, + const struct man *); +extern int mdoc_run(struct termp *, + const struct mdoc *); -/* - * Performs actions on nodes of the abstract syntax tree. Both pre- and - * post-fix operations are defined here. - */ +static struct termp *term_alloc(enum termenc); +static void term_free(struct termp *); +static void term_pword(struct termp *, const char *, int); +static void term_pescape(struct termp *, + const char *, int *, int); +static void term_nescape(struct termp *, + const char *, size_t); +static void term_chara(struct termp *, char); +static void term_encodea(struct termp *, char); +static int term_isopendelim(const char *, int); +static int term_isclosedelim(const char *, int); -/* FIXME: indent/tab. */ -/* FIXME: handle nested lists. */ -#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_NMAX 16 - -/* - * 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). - */ - -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 */ -}; - -static int arg_hasattr(int, size_t, - const struct mdoc_arg *); -static int arg_getattr(int, size_t, - const struct mdoc_arg *); -static size_t arg_offset(const struct mdoc_arg *); -static size_t arg_width(const struct mdoc_arg *); - -/* - * What follows describes prefix and postfix operations for the abstract - * syntax tree descent. - */ - -#define DECL_ARGS \ - struct termp *p, \ - 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_aq); -DECL_PREPOST(termp_ar); -DECL_PREPOST(termp_bd); -DECL_PREPOST(termp_bq); -DECL_PREPOST(termp_cd); -DECL_PREPOST(termp_cm); -DECL_PREPOST(termp_d1); -DECL_PREPOST(termp_dq); -DECL_PREPOST(termp_em); -DECL_PREPOST(termp_fa); -DECL_PREPOST(termp_fd); -DECL_PREPOST(termp_fl); -DECL_PREPOST(termp_fn); -DECL_PREPOST(termp_fo); -DECL_PREPOST(termp_ft); -DECL_PREPOST(termp_ic); -DECL_PREPOST(termp_in); -DECL_PREPOST(termp_it); -DECL_PREPOST(termp_nm); -DECL_PREPOST(termp_op); -DECL_PREPOST(termp_pa); -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_sx); -DECL_PREPOST(termp_va); -DECL_PREPOST(termp_vt); - -DECL_PRE(termp_at); -DECL_PRE(termp_bsx); -DECL_PRE(termp_bx); -DECL_PRE(termp_ex); -DECL_PRE(termp_fx); -DECL_PRE(termp_nd); -DECL_PRE(termp_ns); -DECL_PRE(termp_nx); -DECL_PRE(termp_ox); -DECL_PRE(termp_pp); -DECL_PRE(termp_rv); -DECL_PRE(termp_st); -DECL_PRE(termp_ud); -DECL_PRE(termp_ux); -DECL_PRE(termp_xr); - -DECL_POST(termp_bl); - -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 */ - { NULL, NULL }, /* 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, termp_ar_post }, /* Ar */ - { termp_cd_pre, termp_cd_post }, /* Cd */ - { termp_cm_pre, termp_cm_post }, /* Cm */ - { NULL, NULL }, /* Dv */ - { NULL, NULL }, /* Er */ - { NULL, NULL }, /* Ev */ - { termp_ex_pre, NULL }, /* Ex */ - { termp_fa_pre, termp_fa_post }, /* Fa */ - { termp_fd_pre, termp_fd_post }, /* Fd */ - { termp_fl_pre, termp_fl_post }, /* Fl */ - { termp_fn_pre, termp_fn_post }, /* Fn */ - { termp_ft_pre, termp_ft_post }, /* Ft */ - { termp_ic_pre, termp_ic_post }, /* Ic */ - { termp_in_pre, termp_in_post }, /* In */ - { NULL, NULL }, /* Li */ - { termp_nd_pre, NULL }, /* Nd */ - { termp_nm_pre, termp_nm_post }, /* Nm */ - { termp_op_pre, termp_op_post }, /* Op */ - { NULL, NULL }, /* Ot */ - { termp_pa_pre, termp_pa_post }, /* Pa */ - { termp_rv_pre, NULL }, /* Rv */ - { termp_st_pre, NULL }, /* St */ - { termp_va_pre, termp_va_post }, /* Va */ - { termp_vt_pre, termp_vt_post }, /* Vt */ - { termp_xr_pre, NULL }, /* Xr */ - { NULL, NULL }, /* %A */ - { NULL, NULL }, /* %B */ - { NULL, NULL }, /* %D */ - { NULL, NULL }, /* %I */ - { NULL, NULL }, /* %J */ - { NULL, NULL }, /* %N */ - { NULL, NULL }, /* %O */ - { NULL, NULL }, /* %P */ - { NULL, NULL }, /* %R */ - { NULL, NULL }, /* %T */ - { NULL, NULL }, /* %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 */ - { NULL, NULL }, /* Bf */ - { termp_bq_pre, termp_bq_post }, /* Bo */ - { termp_bq_pre, termp_bq_post }, /* Bq */ - { termp_bsx_pre, NULL }, /* Bsx */ - { termp_bx_pre, NULL }, /* 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, termp_em_post }, /* Em */ - { NULL, NULL }, /* Eo */ - { termp_fx_pre, NULL }, /* Fx */ - { NULL, 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 */ - { NULL, NULL }, /* Rs */ - { NULL, NULL }, /* Sc */ - { termp_sq_pre, termp_sq_post }, /* So */ - { termp_sq_pre, termp_sq_post }, /* Sq */ - { NULL, NULL }, /* Sm */ - { termp_sx_pre, termp_sx_post }, /* Sx */ - { NULL, 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 */ - { NULL, NULL }, /* Bt */ - { NULL, NULL }, /* Hf */ - { NULL, NULL }, /* Fr */ - { termp_ud_pre, NULL }, /* Ud */ -}; - -const struct termact *termacts = __termacts; - - -static size_t -arg_width(const struct mdoc_arg *arg) +void * +ascii_alloc(void) { - /* TODO */ - assert(*arg->value); - return(strlen(*arg->value)); + return(term_alloc(TERMENC_ASCII)); } -static size_t -arg_offset(const struct mdoc_arg *arg) +int +terminal_man(void *arg, const struct man *man) { + 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); + p = (struct termp *)arg; + if (NULL == p->symtab) + p->symtab = term_ascii2htab(); - return(strlen(*arg->value)); + return(man_run(p, man)); } -static int -arg_hasattr(int arg, size_t argc, const struct mdoc_arg *argv) +int +terminal_mdoc(void *arg, const struct mdoc *mdoc) { + struct termp *p; - return(-1 != arg_getattr(arg, argc, argv)); -} + p = (struct termp *)arg; + if (NULL == p->symtab) + p->symtab = term_ascii2htab(); - -static int -arg_getattr(int arg, size_t argc, const struct mdoc_arg *argv) -{ - int i; - - for (i = 0; i < (int)argc; i++) - if (argv[i].arg == arg) - return(i); - return(-1); + return(mdoc_run(p, mdoc)); } -/* ARGSUSED */ -static int -termp_dq_pre(DECL_ARGS) +void +terminal_free(void *arg) { - if (MDOC_BODY != node->type) - return(1); - - word(p, "``"); - p->flags |= TERMP_NOSPACE; - return(1); + term_free((struct termp *)arg); } -/* ARGSUSED */ static void -termp_dq_post(DECL_ARGS) +term_free(struct termp *p) { - if (MDOC_BODY != node->type) - return; + if (p->buf) + free(p->buf); + if (TERMENC_ASCII == p->enc && p->symtab) + term_asciifree(p->symtab); - p->flags |= TERMP_NOSPACE; - word(p, "''"); + free(p); } -/* ARGSUSED */ -static void -termp_it_post(DECL_ARGS) +static struct termp * +term_alloc(enum termenc enc) { - const struct mdoc_node *n, *it; - const struct mdoc_block *bl; - int i; - size_t width, offset; + struct termp *p; - /* - * This (and termp_it_pre()) are the most complicated functions - * here. They must account for a considerable number of - * switches that completely change the output behaviour, like - * -tag versus -column. Yech. - */ - - switch (node->type) { - case (MDOC_BODY): - /* FALLTHROUGH */ - case (MDOC_HEAD): - break; - default: - return; - } - - it = node->parent; - assert(MDOC_BLOCK == it->type); - assert(MDOC_It == it->tok); - - n = it->parent; - assert(MDOC_BODY == n->type); - assert(MDOC_Bl == n->tok); - n = n->parent; - bl = &n->data.block; - - /* If `-tag', adjust our margins accordingly. */ - - if (arg_hasattr(MDOC_Tag, bl->argc, bl->argv)) { - flushln(p); - - /* FIXME: this should auto-size. */ - i = arg_getattr(MDOC_Width, bl->argc, bl->argv); - width = i >= 0 ? arg_width(&bl->argv[i]) : 10; - - /* FIXME: nesting! Should happen at block. */ - i = arg_getattr(MDOC_Offset, bl->argc, bl->argv); - offset = i >= 0 ? arg_width(&bl->argv[i]) : 0; - - if (MDOC_HEAD == node->type) { - p->rmargin = p->maxrmargin; - p->offset -= offset; - p->flags &= ~TERMP_NOBREAK; - } else { - p->offset -= width; - p->flags &= ~TERMP_NOLPAD; - } - } - - if (arg_hasattr(MDOC_Ohang, bl->argc, bl->argv)) { - i = arg_getattr(MDOC_Offset, bl->argc, bl->argv); - offset = i >= 0 ? arg_offset(&bl->argv[i]) : 0; - - flushln(p); - p->offset -= offset; - return; - } + if (NULL == (p = malloc(sizeof(struct termp)))) + err(1, "malloc"); + bzero(p, sizeof(struct termp)); + p->maxrmargin = 78; + p->enc = enc; + return(p); } -/* ARGSUSED */ static int -termp_it_pre(DECL_ARGS) +term_isclosedelim(const char *p, int len) { - const struct mdoc_node *n, *it; - const struct mdoc_block *bl; - int i; - size_t width, offset; - /* - * Also see termp_it_post() for general comments. - */ + if (1 != len) + return(0); - switch (node->type) { - case (MDOC_BODY): + switch (*p) { + case('.'): /* FALLTHROUGH */ - case (MDOC_HEAD): - it = node->parent; - break; - case (MDOC_BLOCK): - it = node; - break; - default: + case(','): + /* FALLTHROUGH */ + case(';'): + /* FALLTHROUGH */ + case(':'): + /* FALLTHROUGH */ + case('?'): + /* FALLTHROUGH */ + case('!'): + /* FALLTHROUGH */ + case(')'): + /* FALLTHROUGH */ + case(']'): + /* FALLTHROUGH */ + case('}'): return(1); + default: + break; } - assert(MDOC_BLOCK == it->type); - assert(MDOC_It == it->tok); - - n = it->parent; - assert(MDOC_BODY == n->type); - assert(MDOC_Bl == n->tok); - n = n->parent; - bl = &n->data.block; - - /* If `-compact', don't assert vertical space. */ - - if (MDOC_BLOCK == node->type) { - if (arg_hasattr(MDOC_Compact, bl->argc, bl->argv)) - newln(p); - else - vspace(p); - return(1); - } - - assert(MDOC_HEAD == node->type - || MDOC_BODY == node->type); - - /* FIXME: see termp_it_post(). */ - - /* If `-tag', adjust our margins accordingly. */ - - if (arg_hasattr(MDOC_Tag, bl->argc, bl->argv)) { - p->flags |= TERMP_NOSPACE; - - i = arg_getattr(MDOC_Width, bl->argc, bl->argv); - width = i >= 0 ? arg_width(&bl->argv[i]) : 10; - - i = arg_getattr(MDOC_Offset, bl->argc, bl->argv); - offset = i >= 0 ? arg_offset(&bl->argv[i]) : 0; - - if (MDOC_HEAD == node->type) { - p->flags |= TERMP_NOBREAK; - p->offset += offset; - p->rmargin = p->offset + width; - } else { - p->flags |= TERMP_NOSPACE; - p->flags |= TERMP_NOLPAD; - p->offset += width; - } - return(1); - } - - /* If `-ohang', adjust left-margin. */ - - if (arg_hasattr(MDOC_Ohang, bl->argc, bl->argv)) { - i = arg_getattr(MDOC_Offset, bl->argc, bl->argv); - offset = i >= 0 ? arg_offset(&bl->argv[i]) : 0; - - p->flags |= TERMP_NOSPACE; - p->offset += offset; - return(1); - } - - return(1); + return(0); } -/* ARGSUSED */ -static void -termp_nm_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_PROG]; -} - - -/* ARGSUSED */ -static void -termp_fl_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_CMD_FLAG]; -} - - -/* ARGSUSED */ static int -termp_ar_pre(DECL_ARGS) +term_isopendelim(const char *p, int len) { - p->flags |= ttypes[TTYPE_CMD_ARG]; - if (NULL == node->child) - word(p, "..."); - return(1); -} + if (1 != len) + return(0); - -/* ARGSUSED */ -static int -termp_nm_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_PROG]; - if (NULL == node->child) - word(p, meta->name); - return(1); -} - - -/* ARGSUSED */ -static int -termp_ns_pre(DECL_ARGS) -{ - - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static int -termp_pp_pre(DECL_ARGS) -{ - - vspace(p); - return(1); -} - - -/* ARGSUSED */ -static void -termp_ar_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_CMD_ARG]; -} - - -/* ARGSUSED */ -static int -termp_st_pre(DECL_ARGS) -{ - const char *tp; - - assert(1 == node->data.elem.argc); - - tp = mdoc_st2a(node->data.elem.argv[0].arg); - word(p, tp); - - return(1); -} - - -/* ARGSUSED */ -static int -termp_rv_pre(DECL_ARGS) -{ - int i; - - i = arg_getattr(MDOC_Std, node->data.elem.argc, - node->data.elem.argv); - assert(i >= 0); - - newln(p); - word(p, "The"); - - p->flags |= ttypes[TTYPE_FUNC_NAME]; - word(p, *node->data.elem.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); -} - - -/* ARGSUSED */ -static int -termp_ex_pre(DECL_ARGS) -{ - int i; - - i = arg_getattr(MDOC_Std, node->data.elem.argc, - node->data.elem.argv); - assert(i >= 0); - - word(p, "The"); - p->flags |= ttypes[TTYPE_PROG]; - word(p, *node->data.elem.argv[i].value); - p->flags &= ~ttypes[TTYPE_PROG]; - word(p, "utility exits 0 on success, and >0 if an error occurs."); - - return(1); -} - - -/* ARGSUSED */ -static int -termp_nd_pre(DECL_ARGS) -{ - - word(p, "\\-"); - return(1); -} - - -/* ARGSUSED */ -static void -termp_bl_post(DECL_ARGS) -{ - - if (MDOC_BLOCK == node->type) - newln(p); -} - - -/* ARGSUSED */ -static void -termp_op_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - p->flags |= TERMP_NOSPACE; - word(p, "\\(rB"); -} - - -/* ARGSUSED */ -static void -termp_sh_post(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_HEAD): - p->flags &= ~ttypes[TTYPE_SECTION]; - newln(p); - break; - case (MDOC_BODY): - newln(p); - p->offset = 0; - break; + switch (*p) { + case('('): + /* FALLTHROUGH */ + case('['): + /* FALLTHROUGH */ + case('{'): + return(1); default: break; } -} - -/* ARGSUSED */ -static int -termp_xr_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - - n = node->child; - assert(n); - - assert(MDOC_TEXT == n->type); - word(p, n->data.text.string); - - if (NULL == (n = n->next)) - return(0); - - assert(MDOC_TEXT == n->type); - p->flags |= TERMP_NOSPACE; - word(p, "("); - p->flags |= TERMP_NOSPACE; - word(p, n->data.text.string); - p->flags |= TERMP_NOSPACE; - word(p, ")"); - return(0); } -/* ARGSUSED */ -static int -termp_vt_pre(DECL_ARGS) +/* + * Flush a line of text. A "line" is loosely defined as being something + * that should be followed by a newline, regardless of whether it's + * broken apart by newlines getting there. A line can also be a + * fragment of a columnar list. + * + * Specifically, a line is whatever's in p->buf of length p->col, which + * is zeroed after this function returns. + * + * The variables TERMP_NOLPAD, TERMP_LITERAL and TERMP_NOBREAK are of + * critical importance here. Their behaviour follows: + * + * - TERMP_NOLPAD: when beginning to write the line, don't left-pad the + * offset value. This is useful when doing columnar lists where the + * prior column has right-padded. + * + * - TERMP_NOBREAK: this is the most important and is used when making + * columns. In short: don't print a newline and instead pad to the + * right margin. Used in conjunction with TERMP_NOLPAD. + * + * - TERMP_NONOBREAK: don't newline when TERMP_NOBREAK is specified. + * + * In-line line breaking: + * + * If TERMP_NOBREAK is specified and the line overruns the right + * margin, it will break and pad-right to the right margin after + * writing. If maxrmargin is violated, it will break and continue + * writing from the right-margin, which will lead to the above + * scenario upon exit. + * + * Otherwise, the line will break at the right margin. Extremely long + * lines will cause the system to emit a warning (TODO: hyphenate, if + * possible). + * + * FIXME: newline breaks occur (in groff) also occur when a single + * space follows a NOBREAK! + */ +void +term_flushln(struct termp *p) { + int i, j; + size_t vbl, vsz, vis, maxvis, mmax, bp; - /* FIXME: this can be "type name". */ - p->flags |= ttypes[TTYPE_VAR_DECL]; - return(1); -} + /* + * First, establish the maximum columns of "visible" content. + * This is usually the difference between the right-margin and + * an indentation, but can be, for tagged lists or columns, a + * small set of values. + */ + assert(p->offset < p->rmargin); + maxvis = p->rmargin - p->offset; + mmax = p->maxrmargin - p->offset; + bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; + vis = 0; -/* ARGSUSED */ -static void -termp_vt_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_VAR_DECL]; - 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. + /* + * If in the standard case (left-justified), then begin with our + * indentation, otherwise (columns, etc.) just start spitting + * out text. */ - p->flags |= ttypes[TTYPE_FUNC_DECL]; - return(1); -} + if ( ! (p->flags & TERMP_NOLPAD)) + /* LINTED */ + for (j = 0; j < (int)p->offset; j++) + putchar(' '); -/* ARGSUSED */ -static void -termp_fd_post(DECL_ARGS) -{ + for (i = 0; i < (int)p->col; i++) { + /* + * 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). + */ - p->flags &= ~ttypes[TTYPE_FUNC_DECL]; - if (node->sec == SEC_SYNOPSIS) - vspace(p); + /* LINTED */ + for (j = i, vsz = 0; j < (int)p->col; j++) { + if (' ' == p->buf[j]) + break; + else if (8 == p->buf[j]) + j += 1; + else + vsz++; + } -} + /* + * Choose the number of blanks to prepend: no blank at the + * beginning of a line, one between words -- but do not + * actually write them yet. + */ + vbl = (size_t)(0 == vis ? 0 : 1); + /* + * Find out whether we would exceed the right margin. + * If so, break to the next line. (TODO: hyphenate) + * Otherwise, write the chosen number of blanks now. + */ + if (vis && vis + vbl + vsz > bp) { + putchar('\n'); + if (TERMP_NOBREAK & p->flags) { + for (j = 0; j < (int)p->rmargin; j++) + putchar(' '); + vis = p->rmargin - p->offset; + } else { + for (j = 0; j < (int)p->offset; j++) + putchar(' '); + vis = 0; + } + } else { + for (j = 0; j < (int)vbl; j++) + putchar(' '); + vis += vbl; + } -/* ARGSUSED */ -static int -termp_sh_pre(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_HEAD): - vspace(p); - p->flags |= ttypes[TTYPE_SECTION]; - break; - case (MDOC_BODY): - p->offset = INDENT; - break; - default: - break; + /* + * Finally, write out the word. + */ + for ( ; i < (int)p->col; i++) { + if (' ' == p->buf[i]) + break; + putchar(p->buf[i]); + } + vis += vsz; } - return(1); -} + /* + * If we've overstepped our maximum visible no-break space, then + * cause a newline and offset at the right margin. + */ -/* 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 ((TERMP_NOBREAK & p->flags) && vis >= maxvis) { + if ( ! (TERMP_NONOBREAK & p->flags)) { + putchar('\n'); + for (i = 0; i < (int)p->rmargin; i++) + putchar(' '); + } + p->col = 0; + return; } - return(1); -} + /* + * If we're not to right-marginalise it (newline), then instead + * pad to the right margin and stay off. + */ -/* ARGSUSED */ -static int -termp_ud_pre(DECL_ARGS) -{ + if (p->flags & TERMP_NOBREAK) { + if ( ! (TERMP_NONOBREAK & p->flags)) + for ( ; vis < maxvis; vis++) + putchar(' '); + } else + putchar('\n'); - word(p, "currently under development."); - return(1); + p->col = 0; } -/* ARGSUSED */ -static int -termp_fl_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) { - p->flags |= ttypes[TTYPE_CMD_FLAG]; - word(p, "\\-"); p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static int -termp_d1_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - newln(p); - p->offset += INDENT; - return(1); -} - - -/* ARGSUSED */ -static void -termp_d1_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) + if (0 == p->col) { + p->flags &= ~TERMP_NOLPAD; return; - newln(p); - p->offset -= INDENT; -} - - -/* ARGSUSED */ -static int -termp_aq_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - word(p, "<"); - 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, ">"); -} - - -/* ARGSUSED */ -static int -termp_ft_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_FUNC_TYPE]; - return(1); -} - - -/* ARGSUSED */ -static void -termp_ft_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_FUNC_TYPE]; - if (node->sec == SEC_SYNOPSIS) - newln(p); - -} - - -/* ARGSUSED */ -static int -termp_fn_pre(DECL_ARGS) -{ - const struct mdoc_node *n; - - assert(node->child); - assert(MDOC_TEXT == node->child->type); - - /* FIXME: can be "type funcname" "type varname"... */ - - p->flags |= ttypes[TTYPE_FUNC_NAME]; - word(p, node->child->data.text.string); - p->flags &= ~ttypes[TTYPE_FUNC_NAME]; - - word(p, "("); - - p->flags |= TERMP_NOSPACE; - for (n = node->child->next; n; n = n->next) { - assert(MDOC_TEXT == n->type); - p->flags |= ttypes[TTYPE_FUNC_ARG]; - word(p, n->data.text.string); - p->flags &= ~ttypes[TTYPE_FUNC_ARG]; - if (n->next) - word(p, ","); } - - word(p, ")"); - - if (SEC_SYNOPSIS == node->sec) - word(p, ";"); - - return(0); + term_flushln(p); + p->flags &= ~TERMP_NOLPAD; } -/* ARGSUSED */ -static void -termp_fn_post(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) { - if (node->sec == SEC_SYNOPSIS) - vspace(p); - + term_newln(p); + putchar('\n'); } -/* ARGSUSED */ -static int -termp_sx_pre(DECL_ARGS) +/* + * Break apart a word into "pwords" (partial-words, usually from + * breaking up a phrase into individual words) and, eventually, put them + * into the output buffer. If we're a literal word, then don't break up + * the word and put it verbatim into the output buffer. + */ +void +term_word(struct termp *p, const char *word) { + int i, j, len; - p->flags |= ttypes[TTYPE_LINK]; - return(1); -} + len = (int)strlen(word); - -/* ARGSUSED */ -static void -termp_sx_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_LINK]; -} - - -/* ARGSUSED */ -static int -termp_fa_pre(DECL_ARGS) -{ - struct mdoc_node *n; - - if (node->parent->tok != MDOC_Fo) { - p->flags |= ttypes[TTYPE_FUNC_ARG]; - return(1); + if (p->flags & TERMP_LITERAL) { + term_pword(p, word, len); + return; } - for (n = node->child; n; n = n->next) { - assert(MDOC_TEXT == n->type); + /* LINTED */ + for (j = i = 0; i < len; i++) { + if (' ' != word[i]) { + j++; + continue; + } + + /* Escaped spaces don't delimit... */ + if (i && ' ' == word[i] && '\\' == word[i - 1]) { + j++; + continue; + } - p->flags |= ttypes[TTYPE_FUNC_ARG]; - word(p, n->data.text.string); - p->flags &= ~ttypes[TTYPE_FUNC_ARG]; - - if (n->next) - word(p, ","); + if (0 == j) + continue; + assert(i >= j); + term_pword(p, &word[i - j], j); + j = 0; } - - if (node->next && node->next->tok == MDOC_Fa) - word(p, ","); - - return(0); + if (j > 0) { + assert(i >= j); + term_pword(p, &word[i - j], j); + } } -/* ARGSUSED */ +/* + * Determine the symbol indicated by an escape sequences, that is, one + * starting with a backslash. Once done, we pass this value into the + * output buffer by way of the symbol table. + */ static void -termp_fa_post(DECL_ARGS) +term_nescape(struct termp *p, const char *word, size_t len) { - - p->flags &= ~ttypes[TTYPE_FUNC_ARG]; -} - - -/* ARGSUSED */ -static int -termp_va_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_VAR_DECL]; - return(1); -} - - -/* ARGSUSED */ -static void -termp_va_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_VAR_DECL]; -} - - -/* ARGSUSED */ -static int -termp_bd_pre(DECL_ARGS) -{ - const struct mdoc_block *bl; - const struct mdoc_node *n; + const char *rhs; + size_t sz; int i; - if (MDOC_BLOCK == node->type) { - vspace(p); - return(1); - } else if (MDOC_BODY != node->type) - return(1); - - assert(MDOC_BLOCK == node->parent->type); - - bl = &node->parent->data.block; - - i = arg_getattr(MDOC_Offset, bl->argc, bl->argv); - if (-1 != i) { - assert(1 == bl->argv[i].sz); - p->offset += arg_offset(&bl->argv[i]); - } - - if ( ! arg_hasattr(MDOC_Literal, bl->argc, bl->argv)) - return(1); - - p->flags |= TERMP_LITERAL; - - for (n = node->child; n; n = n->next) { - assert(MDOC_TEXT == n->type); /* FIXME */ - if ((*n->data.text.string)) { - word(p, n->data.text.string); - flushln(p); - } else - vspace(p); - - } - - p->flags &= ~TERMP_LITERAL; - return(0); + rhs = term_a2ascii(p->symtab, word, len, &sz); + if (rhs) + for (i = 0; i < (int)sz; i++) + term_encodea(p, rhs[i]); } -/* ARGSUSED */ +/* + * Handle an escape sequence: determine its length and pass it to the + * escape-symbol look table. Note that we assume mdoc(3) has validated + * the escape sequence (we assert upon badly-formed escape sequences). + */ static void -termp_bd_post(DECL_ARGS) +term_pescape(struct termp *p, const char *word, int *i, int len) { - int i; - const struct mdoc_block *bl; + int j; - if (MDOC_BODY != node->type) + if (++(*i) >= len) return; - assert(MDOC_BLOCK == node->parent->type); - bl = &node->parent->data.block; + if ('(' == word[*i]) { + (*i)++; + if (*i + 1 >= len) + return; - i = arg_getattr(MDOC_Offset, bl->argc, bl->argv); - if (-1 != i) { - assert(1 == bl->argv[i].sz); - p->offset -= arg_offset(&bl->argv[i]); - } -} - - -/* ARGSUSED */ -static int -termp_qq_pre(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return(1); - word(p, "\""); - p->flags |= TERMP_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static void -termp_qq_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) + term_nescape(p, &word[*i], 2); + (*i)++; return; - p->flags |= TERMP_NOSPACE; - word(p, "\""); -} + } else if ('*' == word[*i]) { + (*i)++; + if (*i >= len) + return; -/* ARGSUSED */ -static int -termp_bsx_pre(DECL_ARGS) -{ + switch (word[*i]) { + case ('('): + (*i)++; + if (*i + 1 >= len) + return; - word(p, "BSDI BSD/OS"); - return(1); -} - - -/* ARGSUSED */ -static int -termp_bx_pre(DECL_ARGS) -{ - - word(p, "BSD"); - return(1); -} - - -/* ARGSUSED */ -static int -termp_ox_pre(DECL_ARGS) -{ - - word(p, "OpenBSD"); - return(1); -} - - -/* ARGSUSED */ -static int -termp_ux_pre(DECL_ARGS) -{ - - word(p, "UNIX"); - return(1); -} - - -/* ARGSUSED */ -static int -termp_fx_pre(DECL_ARGS) -{ - - 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) + term_nescape(p, &word[*i], 2); + (*i)++; + return; + case ('['): + break; + default: + term_nescape(p, &word[*i], 1); + return; + } + + } else if ('f' == word[*i]) { + (*i)++; + if (*i >= len) + return; + switch (word[*i]) { + case ('B'): + p->flags |= TERMP_BOLD; + break; + case ('I'): + p->flags |= TERMP_UNDER; + break; + case ('P'): + /* FALLTHROUGH */ + case ('R'): + p->flags &= ~TERMP_STYLE; + break; + default: + break; + } 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); - p->flags |= ttypes[TTYPE_SSECTION]; - p->offset = INDENT / 2; - break; - default: - break; + } else if ('[' != word[*i]) { + term_nescape(p, &word[*i], 1); + return; } - return(1); -} + (*i)++; + for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++) + /* Loop... */ ; + if (0 == word[*i]) + return; -/* ARGSUSED */ -static void -termp_ss_post(DECL_ARGS) -{ - - switch (node->type) { - case (MDOC_HEAD): - p->flags &= ~ttypes[TTYPE_SSECTION]; - newln(p); - p->offset = INDENT; - break; - default: - break; - } + term_nescape(p, &word[*i - j], (size_t)j); } -/* ARGSUSED */ -static int -termp_pa_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_FILE]; - return(1); -} - - -/* ARGSUSED */ +/* + * 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. + */ static void -termp_pa_post(DECL_ARGS) +term_pword(struct termp *p, const char *word, int len) { + int i; - p->flags &= ~ttypes[TTYPE_FILE]; -} + if (term_isclosedelim(word, len)) + if ( ! (TERMP_IGNDELIM & p->flags)) + p->flags |= TERMP_NOSPACE; + if ( ! (TERMP_NOSPACE & p->flags)) + term_chara(p, ' '); -/* ARGSUSED */ -static int -termp_em_pre(DECL_ARGS) -{ + if ( ! (p->flags & TERMP_NONOSPACE)) + p->flags &= ~TERMP_NOSPACE; - p->flags |= ttypes[TTYPE_EMPH]; - return(1); -} + /* + * If ANSI (word-length styling), then apply our style now, + * before the word. + */ + for (i = 0; i < len; i++) + if ('\\' == word[i]) + term_pescape(p, word, &i, len); + else + term_encodea(p, word[i]); -/* ARGSUSED */ -static void -termp_em_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_EMPH]; + if (term_isopendelim(word, len)) + p->flags |= TERMP_NOSPACE; } -/* ARGSUSED */ -static int -termp_cd_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_CONFIG]; - return(1); -} - - -/* ARGSUSED */ +/* + * Insert a single character into the line-buffer. If the buffer's + * space is exceeded, then allocate more space by doubling the buffer + * size. + */ static void -termp_cd_post(DECL_ARGS) +term_chara(struct termp *p, char c) { + size_t s; - p->flags &= ~ttypes[TTYPE_CONFIG]; -} - - -/* ARGSUSED */ -static int -termp_cm_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_CMD_FLAG]; - return(1); -} - - -/* ARGSUSED */ -static void -termp_cm_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_CMD_FLAG]; -} - - -/* ARGSUSED */ -static int -termp_ic_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_CMD]; - return(1); -} - - -/* ARGSUSED */ -static void -termp_ic_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_CMD]; -} - - -/* ARGSUSED */ -static int -termp_in_pre(DECL_ARGS) -{ - - p->flags |= ttypes[TTYPE_INCLUDE]; - return(1); -} - - -/* ARGSUSED */ -static void -termp_in_post(DECL_ARGS) -{ - - p->flags &= ~ttypes[TTYPE_INCLUDE]; -} - - -/* ARGSUSED */ -static int -termp_at_pre(DECL_ARGS) -{ - enum mdoc_att c; - - c = ATT_DEFAULT; - if (node->child) { - assert(MDOC_TEXT == node->child->type); - c = mdoc_atoatt(node->child->data.text.string); + if (p->col + 1 >= p->maxcols) { + if (0 == p->maxcols) + p->maxcols = 256; + s = p->maxcols * 2; + p->buf = realloc(p->buf, s); + if (NULL == p->buf) + err(1, "realloc"); + p->maxcols = s; } - - word(p, mdoc_att2a(c)); - return(0); + p->buf[(int)(p->col)++] = c; } -/* 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) +term_encodea(struct termp *p, char c) { - 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); - - p->flags |= ttypes[TTYPE_FUNC_NAME]; - for (n = node->child; n; n = n->next) { - assert(MDOC_TEXT == n->type); - word(p, n->data.text.string); + if (TERMP_STYLE & p->flags) { + if (TERMP_BOLD & p->flags) { + term_chara(p, c); + term_chara(p, 8); + } + if (TERMP_UNDER & p->flags) { + term_chara(p, '_'); + term_chara(p, 8); + } } - p->flags &= ~ttypes[TTYPE_FUNC_NAME]; - - return(0); + term_chara(p, c); } - - -/* ARGSUSED */ -static void -termp_fo_post(DECL_ARGS) -{ - - if (MDOC_BODY != node->type) - return; - word(p, ")"); - word(p, ";"); - newln(p); -} - -