=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.1 retrieving revision 1.85 diff -u -p -r1.1 -r1.85 --- mandoc/term.c 2009/02/20 11:04:23 1.1 +++ mandoc/term.c 2009/07/14 15:49:44 1.85 @@ -1,360 +1,590 @@ -/* $Id: term.c,v 1.1 2009/02/20 11:04:23 kristaps Exp $ */ +/* $Id: term.c,v 1.85 2009/07/14 15:49:44 kristaps Exp $ */ /* - * Copyright (c) 2008 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 #include -#include -#include +#include "term.h" +#include "man.h" #include "mdoc.h" +extern int man_run(struct termp *, + const struct man *); +extern int mdoc_run(struct termp *, + const struct mdoc *); -static int termprint_r(size_t, size_t, - const struct mdoc_node *); -static void termprint_head(size_t, - const struct mdoc_meta *); -static void termprint_tail(size_t, - const struct mdoc_meta *); +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); -static char *arch2a(enum mdoc_arch); -static char *vol2a(enum mdoc_vol); -static char *msec2a(enum mdoc_msec); -static size_t ttitle2a(char *, enum mdoc_vol, enum mdoc_msec, - enum mdoc_arch, size_t); +void * +ascii_alloc(void) +{ + return(term_alloc(TERMENC_ASCII)); +} -static char * -arch2a(enum mdoc_arch arch) + +int +terminal_man(void *arg, const struct man *man) { + struct termp *p; - switch (arch) { - case (ARCH_alpha): - return("Alpha"); - case (ARCH_amd64): - return("AMD64"); - case (ARCH_amiga): - return("Amiga"); - case (ARCH_arc): - return("ARC"); - case (ARCH_arm): - return("ARM"); - case (ARCH_armish): - return("ARMISH"); - case (ARCH_aviion): - return("AViion"); - case (ARCH_hp300): - return("HP300"); - case (ARCH_hppa): - return("HPPA"); - case (ARCH_hppa64): - return("HPPA64"); - case (ARCH_i386): - return("i386"); - case (ARCH_landisk): - return("LANDISK"); - case (ARCH_luna88k): - return("Luna88k"); - case (ARCH_mac68k): - return("Mac68k"); - case (ARCH_macppc): - return("MacPPC"); - case (ARCH_mvme68k): - return("MVME68k"); - case (ARCH_mvme88k): - return("MVME88k"); - case (ARCH_mvmeppc): - return("MVMEPPC"); - case (ARCH_pmax): - return("PMAX"); - case (ARCH_sgi): - return("SGI"); - case (ARCH_socppc): - return("SOCPPC"); - case (ARCH_sparc): - return("SPARC"); - case (ARCH_sparc64): - return("SPARC64"); - case (ARCH_sun3): - return("Sun3"); - case (ARCH_vax): - return("VAX"); - case (ARCH_zaurus): - return("Zaurus"); - default: - break; - } + p = (struct termp *)arg; + if (NULL == p->symtab) + p->symtab = term_ascii2htab(); - return(NULL); + return(man_run(p, man)); } -static char * -vol2a(enum mdoc_vol vol) +int +terminal_mdoc(void *arg, const struct mdoc *mdoc) { + struct termp *p; - switch (vol) { - case (VOL_AMD): - return("OpenBSD Ancestral Manual Documents"); - case (VOL_IND): - return("OpenBSD Manual Master Index"); - case (VOL_KM): - return("OpenBSD Kernel Manual"); - case (VOL_LOCAL): - return("OpenBSD Local Manual"); - case (VOL_PRM): - return("OpenBSD Programmer's Manual"); - case (VOL_PS1): - return("OpenBSD Programmer's Supplementary Documents"); - case (VOL_SMM): - return("OpenBSD System Manager's Manual"); - case (VOL_URM): - return("OpenBSD Reference Manual"); - case (VOL_USD): - return("OpenBSD User's Supplementary Documents"); + p = (struct termp *)arg; + if (NULL == p->symtab) + p->symtab = term_ascii2htab(); + + return(mdoc_run(p, mdoc)); +} + + +void +terminal_free(void *arg) +{ + + term_free((struct termp *)arg); +} + + +static void +term_free(struct termp *p) +{ + + if (p->buf) + free(p->buf); + if (TERMENC_ASCII == p->enc && p->symtab) + term_asciifree(p->symtab); + + free(p); +} + + +static struct termp * +term_alloc(enum termenc enc) +{ + struct termp *p; + + if (NULL == (p = malloc(sizeof(struct termp)))) + err(1, "malloc"); + bzero(p, sizeof(struct termp)); + p->maxrmargin = 78; + p->enc = enc; + return(p); +} + + +static int +term_isclosedelim(const char *p, int len) +{ + + if (1 != len) + return(0); + + switch (*p) { + case('.'): + /* FALLTHROUGH */ + case(','): + /* FALLTHROUGH */ + case(';'): + /* FALLTHROUGH */ + case(':'): + /* FALLTHROUGH */ + case('?'): + /* FALLTHROUGH */ + case('!'): + /* FALLTHROUGH */ + case(')'): + /* FALLTHROUGH */ + case(']'): + /* FALLTHROUGH */ + case('}'): + return(1); default: break; } - return(NULL); + return(0); } -static char * -msec2a(enum mdoc_msec msec) +static int +term_isopendelim(const char *p, int len) { - switch (msec) { - case(MSEC_1): - return("1"); - case(MSEC_2): - return("2"); - case(MSEC_3): - return("3"); - case(MSEC_3f): - return("3f"); - case(MSEC_3p): - return("3p"); - case(MSEC_4): - return("4"); - case(MSEC_5): - return("5"); - case(MSEC_6): - return("6"); - case(MSEC_7): - return("7"); - case(MSEC_8): - return("8"); - case(MSEC_9): - return("9"); - case(MSEC_X11): - return("X11"); - case(MSEC_X11R6): - return("X11R6"); - case(MSEC_local): - return("local"); - case(MSEC_n): - return("n"); - case(MSEC_unass): + if (1 != len) + return(0); + + switch (*p) { + case('('): /* FALLTHROUGH */ - case(MSEC_draft): - return("draft"); - case(MSEC_paper): - return("paper"); + case('['): + /* FALLTHROUGH */ + case('{'): + return(1); default: break; } - return(NULL); + + return(0); } -static size_t -ttitle2a(char *dst, enum mdoc_vol vol, enum mdoc_msec msec, - enum mdoc_arch arch, size_t sz) +/* + * 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 usage of termp:flags is as 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_DANGLE: don't newline when TERMP_NOBREAK is specified and + * the line is overrun, and don't pad-right if it's underrun. + * + * - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when + * overruning, instead save the position and continue at that point + * when the next invocation. + * + * 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) { - char *p; - size_t ssz; + int i, j; + size_t vbl, vsz, vis, maxvis, mmax, bp; + static int sv = -1; - if (NULL == (p = vol2a(vol))) - switch (msec) { - case (MSEC_1): - /* FALLTHROUGH */ - case (MSEC_6): - /* FALLTHROUGH */ - case (MSEC_7): - p = vol2a(VOL_URM); - break; - case (MSEC_8): - p = vol2a(VOL_SMM); - break; - case (MSEC_2): - /* FALLTHROUGH */ - case (MSEC_3): - /* FALLTHROUGH */ - case (MSEC_4): - /* FALLTHROUGH */ - case (MSEC_5): - p = vol2a(VOL_PRM); - break; - case (MSEC_9): - p = vol2a(VOL_KM); - break; - default: - /* FIXME: capitalise. */ - if (NULL == (p = msec2a(msec))) - p = msec2a(MSEC_local); - break; + /* + * 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; + + if (sv >= 0) { + vis = (size_t)sv; + sv = -1; + } + + /* + * If in the standard case (left-justified), then begin with our + * indentation, otherwise (columns, etc.) just start spitting + * out text. + */ + + if ( ! (p->flags & TERMP_NOLPAD)) + /* LINTED */ + for (j = 0; j < (int)p->offset; j++) + putchar(' '); + + 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). + */ + + /* 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++; } - assert(p); - if ((ssz = strlcpy(dst, p, sz)) >= sz) - return(ssz); + /* + * 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); - if ((p = arch2a(arch))) { - if ((ssz = strlcat(dst, " (", sz)) >= sz) - return(ssz); - if ((ssz = strlcat(dst, p, sz)) >= sz) - return(ssz); - if ((ssz = strlcat(dst, ")", sz)) >= sz) - return(ssz); + /* + * 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; + } + + /* + * Finally, write out the word. + */ + for ( ; i < (int)p->col; i++) { + if (' ' == p->buf[i]) + break; + putchar(p->buf[i]); + } + vis += vsz; } - return(ssz); + /* + * If we've overstepped our maximum visible no-break space, then + * cause a newline and offset at the right margin. + */ + + if ((TERMP_NOBREAK & p->flags) && vis >= maxvis) { + if ( ! (TERMP_DANGLE & p->flags) && + ! (TERMP_HANG & p->flags)) { + putchar('\n'); + for (i = 0; i < (int)p->rmargin; i++) + putchar(' '); + } + if (TERMP_HANG & p->flags) + sv = (int)(vis - maxvis); + p->col = 0; + return; + } + + /* + * If we're not to right-marginalise it (newline), then instead + * pad to the right margin and stay off. + */ + + if (p->flags & TERMP_NOBREAK) { + if ( ! (TERMP_DANGLE & p->flags)) + for ( ; vis < maxvis; vis++) + putchar(' '); + } else + putchar('\n'); + + p->col = 0; } -static int -termprint_r(size_t cols, size_t indent, const struct mdoc_node *node) +/* + * 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) { - return(1); + p->flags |= TERMP_NOSPACE; + if (0 == p->col) { + p->flags &= ~TERMP_NOLPAD; + return; + } + term_flushln(p); + p->flags &= ~TERMP_NOLPAD; } -static void -termprint_tail(size_t cols, const struct mdoc_meta *meta) +/* + * 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) { - struct tm *tm; - char *buf, *os; - size_t sz, osz, ssz, i; - if (NULL == (buf = malloc(cols))) - err(1, "malloc"); - if (NULL == (os = malloc(cols))) - err(1, "malloc"); + term_newln(p); + putchar('\n'); +} - tm = localtime(&meta->date); - if (NULL == strftime(buf, cols, "%B %d, %Y", tm)) - err(1, "strftime"); - osz = strlcpy(os, meta->os, cols); +/* + * 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; - sz = strlen(buf); - ssz = sz + osz + 1; + len = (int)strlen(word); - if (ssz > cols) { - ssz -= cols; - assert(ssz <= osz); - os[osz - ssz] = 0; - ssz = 1; - } else - ssz = cols - ssz + 1; + if (p->flags & TERMP_LITERAL) { + term_pword(p, word, len); + return; + } - printf("%s", os); - for (i = 0; i < ssz; i++) - printf(" "); + /* 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; + } - printf("%s\n", buf); - - free(buf); - free(os); + if (0 == j) + continue; + assert(i >= j); + term_pword(p, &word[i - j], j); + j = 0; + } + if (j > 0) { + assert(i >= j); + term_pword(p, &word[i - j], j); + } } +/* + * 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 -termprint_head(size_t cols, const struct mdoc_meta *meta) +term_nescape(struct termp *p, const char *word, size_t len) { - char *msec, *buf, *title; - size_t ssz, tsz, ttsz, i; + const char *rhs; + size_t sz; + int i; - if (NULL == (buf = malloc(cols))) - err(1, "malloc"); - if (NULL == (title = malloc(cols))) - err(1, "malloc"); + rhs = term_a2ascii(p->symtab, word, len, &sz); + if (rhs) + for (i = 0; i < (int)sz; i++) + term_encodea(p, rhs[i]); +} - /* Format the manual page header. */ - tsz = ttitle2a(buf, meta->vol, meta->msec, meta->arch, cols); - ttsz = strlcpy(title, meta->title, cols); +/* + * 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 +term_pescape(struct termp *p, const char *word, int *i, int len) +{ + int j; - if (NULL == (msec = msec2a(meta->msec))) - msec = ""; + if (++(*i) >= len) + return; - ssz = (2 * (ttsz + 2 + strlen(msec))) + tsz + 2; + if ('(' == word[*i]) { + (*i)++; + if (*i + 1 >= len) + return; - if (ssz > cols) { - if ((ssz -= cols) % 2) - ssz++; - ssz /= 2; - - assert(ssz <= ttsz); - title[ttsz - ssz] = 0; - ssz = 1; - } else - ssz = ((cols - ssz) / 2) + 1; + term_nescape(p, &word[*i], 2); + (*i)++; + return; - printf("%s(%s)", title, msec); + } else if ('*' == word[*i]) { + (*i)++; + if (*i >= len) + return; - for (i = 0; i < ssz; i++) - printf(" "); + switch (word[*i]) { + case ('('): + (*i)++; + if (*i + 1 >= len) + return; - printf("%s", buf); + 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; - for (i = 0; i < ssz; i++) - printf(" "); + } else if ('[' != word[*i]) { + term_nescape(p, &word[*i], 1); + return; + } - printf("%s(%s)\n\n", title, msec); + (*i)++; + for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++) + /* Loop... */ ; - free(title); - free(buf); + if (0 == word[*i]) + return; + + term_nescape(p, &word[*i - j], (size_t)j); } -int -termprint(const struct mdoc_node *node, - const struct mdoc_meta *meta) +/* + * 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 +term_pword(struct termp *p, const char *word, int len) { - size_t cols; + int i; - if (ERR == setupterm(NULL, STDOUT_FILENO, NULL)) - return(0); + if (term_isclosedelim(word, len)) + if ( ! (TERMP_IGNDELIM & p->flags)) + p->flags |= TERMP_NOSPACE; - cols = columns < 60 ? 60 : (size_t)columns; + if ( ! (TERMP_NOSPACE & p->flags)) + term_chara(p, ' '); - termprint_head(cols, meta); - if ( ! termprint_r(cols, 0, node)) - return(0); - termprint_tail(cols, meta); - return(1); + if ( ! (p->flags & TERMP_NONOSPACE)) + p->flags &= ~TERMP_NOSPACE; + + /* + * 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]); + + if (term_isopendelim(word, len)) + p->flags |= TERMP_NOSPACE; } +/* + * 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 +term_chara(struct termp *p, char c) +{ + size_t s; + + 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; + } + p->buf[(int)(p->col)++] = c; +} + + +static void +term_encodea(struct termp *p, char c) +{ + + 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); + } + } + term_chara(p, c); +}