1/*	$Id: term_ascii.c,v 1.58 2017/06/14 14:24:20 schwarze Exp $ */
2/*
3 * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2014, 2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
5 *
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
9 *
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 */
18#include "config.h"
19
20#include <sys/types.h>
21
22#include <assert.h>
23#if HAVE_WCHAR
24#include <locale.h>
25#endif
26#include <stdint.h>
27#include <stdio.h>
28#include <stdlib.h>
29#include <unistd.h>
30#if HAVE_WCHAR
31#include <wchar.h>
32#endif
33
34#include "mandoc.h"
35#include "mandoc_aux.h"
36#include "out.h"
37#include "term.h"
38#include "manconf.h"
39#include "main.h"
40
41static	struct termp	 *ascii_init(enum termenc, const struct manoutput *);
42static	int		  ascii_hspan(const struct termp *,
43				const struct roffsu *);
44static	size_t		  ascii_width(const struct termp *, int);
45static	void		  ascii_advance(struct termp *, size_t);
46static	void		  ascii_begin(struct termp *);
47static	void		  ascii_end(struct termp *);
48static	void		  ascii_endline(struct termp *);
49static	void		  ascii_letter(struct termp *, int);
50static	void		  ascii_setwidth(struct termp *, int, int);
51
52#if HAVE_WCHAR
53static	void		  locale_advance(struct termp *, size_t);
54static	void		  locale_endline(struct termp *);
55static	void		  locale_letter(struct termp *, int);
56static	size_t		  locale_width(const struct termp *, int);
57#endif
58
59
60static struct termp *
61ascii_init(enum termenc enc, const struct manoutput *outopts)
62{
63#if HAVE_WCHAR
64	char		*v;
65#endif
66	struct termp	*p;
67
68	p = mandoc_calloc(1, sizeof(*p));
69	p->tcol = p->tcols = mandoc_calloc(1, sizeof(*p->tcol));
70	p->maxtcol = 1;
71
72	p->line = 1;
73	p->defrmargin = p->lastrmargin = 78;
74	p->fontq = mandoc_reallocarray(NULL,
75	     (p->fontsz = 8), sizeof(*p->fontq));
76	p->fontq[0] = p->fontl = TERMFONT_NONE;
77
78	p->begin = ascii_begin;
79	p->end = ascii_end;
80	p->hspan = ascii_hspan;
81	p->type = TERMTYPE_CHAR;
82
83	p->enc = TERMENC_ASCII;
84	p->advance = ascii_advance;
85	p->endline = ascii_endline;
86	p->letter = ascii_letter;
87	p->setwidth = ascii_setwidth;
88	p->width = ascii_width;
89
90#if HAVE_WCHAR
91	if (TERMENC_ASCII != enc) {
92
93		/*
94		 * Do not change any of this to LC_ALL.  It might break
95		 * the formatting by subtly changing the behaviour of
96		 * various functions, for example strftime(3).  As a
97		 * worst case, it might even cause buffer overflows.
98		 */
99
100		v = TERMENC_LOCALE == enc ?
101		    setlocale(LC_CTYPE, "") :
102		    setlocale(LC_CTYPE, UTF8_LOCALE);
103		if (NULL != v && MB_CUR_MAX > 1) {
104			p->enc = enc;
105			p->advance = locale_advance;
106			p->endline = locale_endline;
107			p->letter = locale_letter;
108			p->width = locale_width;
109		}
110	}
111#endif
112
113	if (outopts->mdoc) {
114		p->mdocstyle = 1;
115		p->defindent = 5;
116	}
117	if (outopts->indent)
118		p->defindent = outopts->indent;
119	if (outopts->width)
120		p->defrmargin = outopts->width;
121	if (outopts->synopsisonly)
122		p->synopsisonly = 1;
123
124	return p;
125}
126
127void *
128ascii_alloc(const struct manoutput *outopts)
129{
130
131	return ascii_init(TERMENC_ASCII, outopts);
132}
133
134void *
135utf8_alloc(const struct manoutput *outopts)
136{
137
138	return ascii_init(TERMENC_UTF8, outopts);
139}
140
141void *
142locale_alloc(const struct manoutput *outopts)
143{
144
145	return ascii_init(TERMENC_LOCALE, outopts);
146}
147
148static void
149ascii_setwidth(struct termp *p, int iop, int width)
150{
151
152	width /= 24;
153	p->tcol->rmargin = p->defrmargin;
154	if (iop > 0)
155		p->defrmargin += width;
156	else if (iop == 0)
157		p->defrmargin = width ? (size_t)width : p->lastrmargin;
158	else if (p->defrmargin > (size_t)width)
159		p->defrmargin -= width;
160	else
161		p->defrmargin = 0;
162	p->lastrmargin = p->tcol->rmargin;
163	p->tcol->rmargin = p->maxrmargin = p->defrmargin;
164}
165
166void
167terminal_sepline(void *arg)
168{
169	struct termp	*p;
170	size_t		 i;
171
172	p = (struct termp *)arg;
173	(*p->endline)(p);
174	for (i = 0; i < p->defrmargin; i++)
175		(*p->letter)(p, '-');
176	(*p->endline)(p);
177	(*p->endline)(p);
178}
179
180static size_t
181ascii_width(const struct termp *p, int c)
182{
183
184	return 1;
185}
186
187void
188ascii_free(void *arg)
189{
190
191	term_free((struct termp *)arg);
192}
193
194static void
195ascii_letter(struct termp *p, int c)
196{
197
198	putchar(c);
199}
200
201static void
202ascii_begin(struct termp *p)
203{
204
205	(*p->headf)(p, p->argf);
206}
207
208static void
209ascii_end(struct termp *p)
210{
211
212	(*p->footf)(p, p->argf);
213}
214
215static void
216ascii_endline(struct termp *p)
217{
218
219	p->line++;
220	p->tcol->offset -= p->ti;
221	p->ti = 0;
222	putchar('\n');
223}
224
225static void
226ascii_advance(struct termp *p, size_t len)
227{
228	size_t		i;
229
230	for (i = 0; i < len; i++)
231		putchar(' ');
232}
233
234static int
235ascii_hspan(const struct termp *p, const struct roffsu *su)
236{
237	double		 r;
238
239	switch (su->unit) {
240	case SCALE_BU:
241		r = su->scale;
242		break;
243	case SCALE_CM:
244		r = su->scale * 240.0 / 2.54;
245		break;
246	case SCALE_FS:
247		r = su->scale * 65536.0;
248		break;
249	case SCALE_IN:
250		r = su->scale * 240.0;
251		break;
252	case SCALE_MM:
253		r = su->scale * 0.24;
254		break;
255	case SCALE_VS:
256	case SCALE_PC:
257		r = su->scale * 40.0;
258		break;
259	case SCALE_PT:
260		r = su->scale * 10.0 / 3.0;
261		break;
262	case SCALE_EN:
263	case SCALE_EM:
264		r = su->scale * 24.0;
265		break;
266	default:
267		abort();
268	}
269	return r > 0.0 ? r + 0.01 : r - 0.01;
270}
271
272const char *
273ascii_uc2str(int uc)
274{
275	static const char nbrsp[2] = { ASCII_NBRSP, '\0' };
276	static const char *tab[] = {
277	"<NUL>","<SOH>","<STX>","<ETX>","<EOT>","<ENQ>","<ACK>","<BEL>",
278	"<BS>",	"\t",	"<LF>",	"<VT>",	"<FF>",	"<CR>",	"<SO>",	"<SI>",
279	"<DLE>","<DC1>","<DC2>","<DC3>","<DC4>","<NAK>","<SYN>","<ETB>",
280	"<CAN>","<EM>",	"<SUB>","<ESC>","<FS>",	"<GS>",	"<RS>",	"<US>",
281	" ",	"!",	"\"",	"#",	"$",	"%",	"&",	"'",
282	"(",	")",	"*",	"+",	",",	"-",	".",	"/",
283	"0",	"1",	"2",	"3",	"4",	"5",	"6",	"7",
284	"8",	"9",	":",	";",	"<",	"=",	">",	"?",
285	"@",	"A",	"B",	"C",	"D",	"E",	"F",	"G",
286	"H",	"I",	"J",	"K",	"L",	"M",	"N",	"O",
287	"P",	"Q",	"R",	"S",	"T",	"U",	"V",	"W",
288	"X",	"Y",	"Z",	"[",	"\\",	"]",	"^",	"_",
289	"`",	"a",	"b",	"c",	"d",	"e",	"f",	"g",
290	"h",	"i",	"j",	"k",	"l",	"m",	"n",	"o",
291	"p",	"q",	"r",	"s",	"t",	"u",	"v",	"w",
292	"x",	"y",	"z",	"{",	"|",	"}",	"~",	"<DEL>",
293	"<80>",	"<81>",	"<82>",	"<83>",	"<84>",	"<85>",	"<86>",	"<87>",
294	"<88>",	"<89>",	"<8A>",	"<8B>",	"<8C>",	"<8D>",	"<8E>",	"<8F>",
295	"<90>",	"<91>",	"<92>",	"<93>",	"<94>",	"<95>",	"<96>",	"<97>",
296	"<98>",	"<99>",	"<9A>",	"<9B>",	"<9C>",	"<9D>",	"<9E>",	"<9F>",
297	nbrsp,	"!",	"/\bc",	"GBP",	"o\bx",	"=\bY",	"|",	"<sec>",
298	"\"",	"(C)",	"_\ba",	"<<",	"~",	"",	"(R)",	"-",
299	"<deg>","+-",	"2",	"3",	"'",	",\bu",	"<par>",".",
300	",",	"1",	"_\bo",	">>",	"1/4",	"1/2",	"3/4",	"?",
301	"`\bA",	"'\bA",	"^\bA",	"~\bA",	"\"\bA","o\bA",	"AE",	",\bC",
302	"`\bE",	"'\bE",	"^\bE",	"\"\bE","`\bI",	"'\bI",	"^\bI",	"\"\bI",
303	"-\bD",	"~\bN",	"`\bO",	"'\bO",	"^\bO",	"~\bO",	"\"\bO","x",
304	"/\bO",	"`\bU",	"'\bU",	"^\bU",	"\"\bU","'\bY",	"Th",	"ss",
305	"`\ba",	"'\ba",	"^\ba",	"~\ba",	"\"\ba","o\ba",	"ae",	",\bc",
306	"`\be",	"'\be",	"^\be",	"\"\be","`\bi",	"'\bi",	"^\bi",	"\"\bi",
307	"d",	"~\bn",	"`\bo",	"'\bo",	"^\bo",	"~\bo",	"\"\bo","-:-",
308	"/\bo",	"`\bu",	"'\bu",	"^\bu",	"\"\bu","'\by",	"th",	"\"\by",
309	"A",	"a",	"A",	"a",	"A",	"a",	"'\bC",	"'\bc",
310	"^\bC",	"^\bc",	"C",	"c",	"C",	"c",	"D",	"d",
311	"/\bD",	"/\bd",	"E",	"e",	"E",	"e",	"E",	"e",
312	"E",	"e",	"E",	"e",	"^\bG",	"^\bg",	"G",	"g",
313	"G",	"g",	",\bG",	",\bg",	"^\bH",	"^\bh",	"/\bH",	"/\bh",
314	"~\bI",	"~\bi",	"I",	"i",	"I",	"i",	"I",	"i",
315	"I",	"i",	"IJ",	"ij",	"^\bJ",	"^\bj",	",\bK",	",\bk",
316	"q",	"'\bL",	"'\bl",	",\bL",	",\bl",	"L",	"l",	"L",
317	"l",	"/\bL",	"/\bl",	"'\bN",	"'\bn",	",\bN",	",\bn",	"N",
318	"n",	"'n",	"Ng",	"ng",	"O",	"o",	"O",	"o",
319	"O",	"o",	"OE",	"oe",	"'\bR",	"'\br",	",\bR",	",\br",
320	"R",	"r",	"'\bS",	"'\bs",	"^\bS",	"^\bs",	",\bS",	",\bs",
321	"S",	"s",	",\bT",	",\bt",	"T",	"t",	"/\bT",	"/\bt",
322	"~\bU",	"~\bu",	"U",	"u",	"U",	"u",	"U",	"u",
323	"U",	"u",	"U",	"u",	"^\bW",	"^\bw",	"^\bY",	"^\by",
324	"\"\bY","'\bZ",	"'\bz",	"Z",	"z",	"Z",	"z",	"s",
325	"b",	"B",	"B",	"b",	"6",	"6",	"O",	"C",
326	"c",	"D",	"D",	"D",	"d",	"d",	"3",	"@",
327	"E",	"F",	",\bf",	"G",	"G",	"hv",	"I",	"/\bI",
328	"K",	"k",	"/\bl",	"l",	"W",	"N",	"n",	"~\bO",
329	"O",	"o",	"OI",	"oi",	"P",	"p",	"YR",	"2",
330	"2",	"SH",	"sh",	"t",	"T",	"t",	"T",	"U",
331	"u",	"Y",	"V",	"Y",	"y",	"/\bZ",	"/\bz",	"ZH",
332	"ZH",	"zh",	"zh",	"/\b2",	"5",	"5",	"ts",	"w",
333	"|",	"||",	"|=",	"!",	"DZ",	"Dz",	"dz",	"LJ",
334	"Lj",	"lj",	"NJ",	"Nj",	"nj",	"A",	"a",	"I",
335	"i",	"O",	"o",	"U",	"u",	"U",	"u",	"U",
336	"u",	"U",	"u",	"U",	"u",	"@",	"A",	"a",
337	"A",	"a",	"AE",	"ae",	"/\bG",	"/\bg",	"G",	"g",
338	"K",	"k",	"O",	"o",	"O",	"o",	"ZH",	"zh",
339	"j",	"DZ",	"Dz",	"dz",	"'\bG",	"'\bg",	"HV",	"W",
340	"`\bN",	"`\bn",	"A",	"a",	"'\bAE","'\bae","O",	"o"};
341
342	assert(uc >= 0);
343	if ((size_t)uc < sizeof(tab)/sizeof(tab[0]))
344		return tab[uc];
345	return mchars_uc2str(uc);
346}
347
348#if HAVE_WCHAR
349static size_t
350locale_width(const struct termp *p, int c)
351{
352	int		rc;
353
354	if (c == ASCII_NBRSP)
355		c = ' ';
356	rc = wcwidth(c);
357	if (rc < 0)
358		rc = 0;
359	return rc;
360}
361
362static void
363locale_advance(struct termp *p, size_t len)
364{
365	size_t		i;
366
367	for (i = 0; i < len; i++)
368		putwchar(L' ');
369}
370
371static void
372locale_endline(struct termp *p)
373{
374
375	p->line++;
376	p->tcol->offset -= p->ti;
377	p->ti = 0;
378	putwchar(L'\n');
379}
380
381static void
382locale_letter(struct termp *p, int c)
383{
384
385	putwchar(c);
386}
387#endif
388