=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.7 retrieving revision 1.74 diff -u -p -r1.7 -r1.74 --- mandoc/term.c 2009/02/22 14:31:08 1.7 +++ mandoc/term.c 2009/04/12 19:45:26 1.74 @@ -1,65 +1,213 @@ -/* $Id: term.c,v 1.7 2009/02/22 14:31:08 kristaps Exp $ */ +/* $Id: term.c,v 1.74 2009/04/12 19:45:26 kristaps Exp $ */ /* - * Copyright (c) 2008, 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 #include -#include -#ifdef __linux__ -#include -#endif - #include "term.h" +#include "man.h" +#include "mdoc.h" -enum termstyle { - STYLE_CLEAR, - STYLE_BOLD, - STYLE_UNDERLINE -}; +extern int man_run(struct termp *, + const struct man *); +extern int mdoc_run(struct termp *, + const struct mdoc *); -static void termprint_r(struct termp *, - const struct mdoc_meta *, - const struct mdoc_node *); -static void termprint_header(struct termp *, - const struct mdoc_meta *); -static void termprint_footer(struct termp *, - 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_stringa(struct termp *, + const char *, size_t); +static int term_isopendelim(const char *, int); +static int term_isclosedelim(const char *, int); -static void pword(struct termp *, const char *, size_t); -static void pescape(struct termp *, - const char *, size_t *, size_t); -static void chara(struct termp *, char); -static void style(struct termp *, enum termstyle); -#ifdef __linux__ -extern size_t strlcat(char *, const char *, size_t); -extern size_t strlcpy(char *, const char *, size_t); -#endif +void * +ascii_alloc(void) +{ + return(term_alloc(TERMENC_ASCII)); +} + + +int +terminal_man(void *arg, const struct man *man) +{ + struct termp *p; + + p = (struct termp *)arg; + if (NULL == p->symtab) + p->symtab = term_ascii2htab(); + + return(man_run(p, man)); +} + + +int +terminal_mdoc(void *arg, const struct mdoc *mdoc) +{ + struct termp *p; + + p = (struct termp *)arg; + if (NULL == p->symtab) + p->symtab = term_ascii2htab(); + + return(mdoc_run(p, mdoc)); +} + + void -flushln(struct termp *p) +terminal_free(void *arg) { - size_t i, j, vsz, vis, maxvis; + 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(0); +} + + +static int +term_isopendelim(const char *p, int len) +{ + + if (1 != len) + return(0); + + switch (*p) { + case('('): + /* FALLTHROUGH */ + case('['): + /* FALLTHROUGH */ + case('{'): + return(1); + default: + break; + } + + return(0); +} + + +/* + * 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). + */ +void +term_flushln(struct termp *p) +{ + int i, j; + size_t vsz, vis, maxvis, mmax, bp; + /* * First, establish the maximum columns of "visible" content. * This is usually the difference between the right-margin and @@ -69,6 +217,8 @@ flushln(struct termp *p) assert(p->offset < p->rmargin); maxvis = p->rmargin - p->offset; + mmax = p->maxrmargin - p->offset; + bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; vis = 0; /* @@ -79,74 +229,91 @@ flushln(struct termp *p) if ( ! (p->flags & TERMP_NOLPAD)) /* LINTED */ - for (j = 0; j < p->offset; j++) + for (j = 0; j < (int)p->offset; j++) putchar(' '); - for (i = 0; i < p->col; i++) { + for (i = 0; i < (int)p->col; i++) { /* * Count up visible word characters. Control sequences - * (starting with the CSI) aren't counted. + * (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). */ - assert( ! isspace(p->buf[i])); /* LINTED */ - for (j = i, vsz = 0; j < p->col; j++) { - if (isspace(p->buf[j])) + for (j = i, vsz = 0; j < (int)p->col; j++) { + if (' ' == p->buf[j]) break; - else if (27 == p->buf[j]) { - assert(j + 4 <= p->col); - j += 3; - } else + else if (8 == p->buf[j]) + j += 1; + else vsz++; } - assert(vsz > 0); /* - * If a word is too long and we're within a line, put it - * on the next line. Puke if we're being asked to write - * something that will exceed the right margin (i.e., - * from a fresh line or when we're not allowed to break - * the line with TERMP_NOBREAK). + * Do line-breaking. If we're greater than our + * break-point and already in-line, break to the next + * line and start writing. If we're at the line start, + * then write out the word (TODO: hyphenate) and break + * in a subsequent loop invocation. */ - if (vis && vis + vsz >= maxvis) { - /* FIXME */ - if (p->flags & TERMP_NOBREAK) - errx(1, "word breaks right margin"); + if ( ! (TERMP_NOBREAK & p->flags)) { + if (vis && vis + vsz > bp) { + putchar('\n'); + for (j = 0; j < (int)p->offset; j++) + putchar(' '); + vis = 0; + } + } else if (vis && vis + vsz > bp) { putchar('\n'); - for (j = 0; j < p->offset; j++) + for (j = 0; j < (int)p->rmargin; j++) putchar(' '); - vis = 0; - } else if (vis + vsz >= maxvis) { - /* FIXME */ - errx(1, "word breaks right margin"); + vis = p->rmargin - p->offset; } /* * Write out the word and a trailing space. Omit the - * space if we're the last word in the line. + * space if we're the last word in the line or beyond + * our breakpoint. */ - for ( ; i < p->col; i++) { - if (isspace(p->buf[i])) + for ( ; i < (int)p->col; i++) { + if (' ' == p->buf[i]) break; putchar(p->buf[i]); } vis += vsz; - if (i < p->col) { + if (i < (int)p->col && vis <= bp) { putchar(' '); vis++; } } /* + * 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_NONOBREAK & p->flags)) { + putchar('\n'); + for (i = 0; i < (int)p->rmargin; i++) + putchar(' '); + } + 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) { - for ( ; vis <= maxvis; vis++) - putchar(' '); + if ( ! (TERMP_NONOBREAK & p->flags)) + for ( ; vis < maxvis; vis++) + putchar(' '); } else putchar('\n'); @@ -154,369 +321,279 @@ flushln(struct termp *p) } +/* + * 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 -newln(struct termp *p) +term_newln(struct termp *p) { - /* - * A newline only breaks an existing line; it won't assert - * vertical space. - */ p->flags |= TERMP_NOSPACE; - if (0 == p->col) + if (0 == p->col) { + p->flags &= ~TERMP_NOLPAD; return; - flushln(p); + } + term_flushln(p); + p->flags &= ~TERMP_NOLPAD; } +/* + * 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 -vspace(struct termp *p) +term_vspace(struct termp *p) { - /* - * Asserts a vertical space (a full, empty line-break between - * lines). - */ - newln(p); + term_newln(p); putchar('\n'); } -static void -chara(struct termp *p, char c) +/* + * 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; - /* TODO: dynamically expand the buffer. */ - if (p->col + 1 >= p->maxcols) - errx(1, "line overrun"); - p->buf[(p->col)++] = c; -} + len = (int)strlen(word); - -static void -style(struct termp *p, enum termstyle esc) -{ - - if (p->col + 4 >= p->maxcols) - errx(1, "line overrun"); - - p->buf[(p->col)++] = 27; - p->buf[(p->col)++] = '['; - switch (esc) { - case (STYLE_CLEAR): - p->buf[(p->col)++] = '0'; - break; - case (STYLE_BOLD): - p->buf[(p->col)++] = '1'; - break; - case (STYLE_UNDERLINE): - p->buf[(p->col)++] = '4'; - break; - default: - abort(); - /* NOTREACHED */ - } - p->buf[(p->col)++] = 'm'; -} - - -static void -pescape(struct termp *p, const char *word, size_t *i, size_t len) -{ - - (*i)++; - assert(*i < len); - - if ('(' == word[*i]) { - /* Two-character escapes. */ - (*i)++; - assert(*i + 1 < len); - - if ('r' == word[*i] && 'B' == word[*i + 1]) - chara(p, ']'); - else if ('l' == word[*i] && 'B' == word[*i + 1]) - chara(p, '['); - - (*i)++; + if (p->flags & TERMP_LITERAL) { + term_pword(p, word, len); return; - - } else if ('[' != word[*i]) { - /* One-character escapes. */ - switch (word[*i]) { - case ('\\'): - /* FALLTHROUGH */ - case ('\''): - /* FALLTHROUGH */ - case ('`'): - /* FALLTHROUGH */ - case ('-'): - /* FALLTHROUGH */ - case ('.'): - chara(p, word[*i]); - default: - break; - } - return; } - /* n-character escapes. */ -} - -static void -pword(struct termp *p, const char *word, size_t len) -{ - size_t i; - - assert(len > 0); - - if ( ! (p->flags & TERMP_NOSPACE)) - chara(p, ' '); - - p->flags &= ~TERMP_NOSPACE; - - if (p->flags & TERMP_BOLD) - style(p, STYLE_BOLD); - if (p->flags & TERMP_UNDERLINE) - style(p, STYLE_UNDERLINE); - - for (i = 0; i < len; i++) { - if ('\\' == word[i]) { - pescape(p, word, &i, len); - continue; - } - chara(p, word[i]); - } - - if (p->flags & TERMP_BOLD || - p->flags & TERMP_UNDERLINE) - style(p, STYLE_CLEAR); -} - - -void -word(struct termp *p, const char *word) -{ - size_t i, j, len; - - if (mdoc_isdelim(word)) - p->flags |= TERMP_NOSPACE; - - len = strlen(word); - assert(len > 0); - /* LINTED */ for (j = i = 0; i < len; i++) { - if ( ! isspace(word[i])) { + if (' ' != word[i]) { j++; continue; + } + + /* Escaped spaces don't delimit... */ + if (i && ' ' == word[i] && '\\' == word[i - 1]) { + j++; + continue; } + if (0 == j) continue; assert(i >= j); - pword(p, &word[i - j], j); + term_pword(p, &word[i - j], j); j = 0; } if (j > 0) { assert(i >= j); - pword(p, &word[i - j], 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_r(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +term_nescape(struct termp *p, const char *word, size_t len) { - int dochild; + const char *rhs; + size_t sz; - /* Pre-processing. */ - - dochild = 1; - - if (MDOC_TEXT != node->type) { - if (termacts[node->tok].pre) - if ( ! (*termacts[node->tok].pre)(p, meta, node)) - dochild = 0; - } else /* MDOC_TEXT == node->type */ - word(p, node->data.text.string); - - /* Children. */ - - if (dochild && node->child) - termprint_r(p, meta, node->child); - - /* Post-processing. */ - - if (MDOC_TEXT != node->type) - if (termacts[node->tok].post) - (*termacts[node->tok].post)(p, meta, node); - - /* Siblings. */ - - if (node->next) - termprint_r(p, meta, node->next); + if (NULL == (rhs = term_a2ascii(p->symtab, word, len, &sz))) + return; + term_stringa(p, rhs, sz); } +/* + * 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 -termprint_footer(struct termp *p, const struct mdoc_meta *meta) +term_pescape(struct termp *p, const char *word, int *i, int len) { - struct tm *tm; - char *buf, *os; - size_t sz, osz, ssz, i; + int j; - if (NULL == (buf = malloc(p->rmargin))) - err(1, "malloc"); - if (NULL == (os = malloc(p->rmargin))) - err(1, "malloc"); + if (++(*i) >= len) + return; - tm = localtime(&meta->date); + if ('(' == word[*i]) { + (*i)++; + if (*i + 1 >= len) + return; -#ifdef __linux__ - if (0 == strftime(buf, p->rmargin, "%B %d, %Y", tm)) -#else - if (NULL == strftime(buf, p->rmargin, "%B %d, %Y", tm)) -#endif - err(1, "strftime"); + term_nescape(p, &word[*i], 2); + (*i)++; + return; - osz = strlcpy(os, meta->os, p->rmargin); + } else if ('*' == word[*i]) { + (*i)++; + if (*i >= len) + return; - sz = strlen(buf); - ssz = sz + osz + 1; + switch (word[*i]) { + case ('('): + (*i)++; + if (*i + 1 >= len) + return; - if (ssz > p->rmargin) { - ssz -= p->rmargin; - assert(ssz <= osz); - os[osz - ssz] = 0; - ssz = 1; - } else - ssz = p->rmargin - ssz + 1; - - printf("\n"); - printf("%s", os); - for (i = 0; i < ssz; i++) - printf(" "); - - printf("%s\n", buf); - fflush(stdout); - - free(buf); - free(os); -} - - -static void -termprint_header(struct termp *p, const struct mdoc_meta *meta) -{ - char *buf, *title; - const char *pp, *msec; - size_t ssz, tsz, ttsz, i;; - - if (NULL == (buf = malloc(p->rmargin))) - err(1, "malloc"); - if (NULL == (title = malloc(p->rmargin))) - err(1, "malloc"); - - if (NULL == (pp = mdoc_vol2a(meta->vol))) - switch (meta->msec) { - case (MSEC_1): - /* FALLTHROUGH */ - case (MSEC_6): - /* FALLTHROUGH */ - case (MSEC_7): - pp = mdoc_vol2a(VOL_URM); + term_nescape(p, &word[*i], 2); + (*i)++; + return; + case ('['): break; - case (MSEC_8): - pp = mdoc_vol2a(VOL_SMM); + 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 (MSEC_2): + case ('I'): + p->flags |= TERMP_UNDER; + break; + case ('P'): /* FALLTHROUGH */ - case (MSEC_3): - /* FALLTHROUGH */ - case (MSEC_4): - /* FALLTHROUGH */ - case (MSEC_5): - pp = mdoc_vol2a(VOL_PRM); + case ('R'): + p->flags &= ~TERMP_STYLE; break; - case (MSEC_9): - pp = mdoc_vol2a(VOL_KM); - break; default: - /* FIXME: capitalise. */ - if (NULL == (pp = mdoc_msec2a(meta->msec))) - pp = mdoc_msec2a(MSEC_local); break; } - assert(pp); + return; - tsz = strlcpy(buf, pp, p->rmargin); - assert(tsz < p->rmargin); - - if ((pp = mdoc_arch2a(meta->arch))) { - tsz = strlcat(buf, " (", p->rmargin); - assert(tsz < p->rmargin); - tsz = strlcat(buf, pp, p->rmargin); - assert(tsz < p->rmargin); - tsz = strlcat(buf, ")", p->rmargin); - assert(tsz < p->rmargin); + } else if ('[' != word[*i]) { + term_nescape(p, &word[*i], 1); + return; } - ttsz = strlcpy(title, meta->title, p->rmargin); + (*i)++; + for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++) + /* Loop... */ ; - if (NULL == (msec = mdoc_msec2a(meta->msec))) - msec = ""; + if (0 == word[*i]) + return; - ssz = (2 * (ttsz + 2 + strlen(msec))) + tsz + 2; + term_nescape(p, &word[*i - j], (size_t)j); +} - if (ssz > p->rmargin) { - if ((ssz -= p->rmargin) % 2) - ssz++; - ssz /= 2; - - assert(ssz <= ttsz); - title[ttsz - ssz] = 0; - ssz = 1; - } else - ssz = ((p->rmargin - ssz) / 2) + 1; - printf("%s(%s)", title, msec); +/* + * 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) +{ + int i; - for (i = 0; i < ssz; i++) - printf(" "); + if (term_isclosedelim(word, len)) + if ( ! (TERMP_IGNDELIM & p->flags)) + p->flags |= TERMP_NOSPACE; - printf("%s", buf); + if ( ! (TERMP_NOSPACE & p->flags)) + term_chara(p, ' '); - for (i = 0; i < ssz; i++) - printf(" "); + if ( ! (p->flags & TERMP_NONOSPACE)) + p->flags &= ~TERMP_NOSPACE; - printf("%s(%s)\n", title, msec); - fflush(stdout); + /* + * If ANSI (word-length styling), then apply our style now, + * before the word. + */ - free(title); - free(buf); + for (i = 0; i < len; i++) { + if ('\\' == word[i]) { + term_pescape(p, word, &i, len); + continue; + } + + if (TERMP_STYLE & p->flags) { + if (TERMP_BOLD & p->flags) { + term_chara(p, word[i]); + term_chara(p, 8); + } + if (TERMP_UNDER & p->flags) { + term_chara(p, '_'); + term_chara(p, 8); + } + } + + term_chara(p, word[i]); + } + + if (term_isopendelim(word, len)) + p->flags |= TERMP_NOSPACE; } -void -termprint(const struct mdoc_node *node, - const struct mdoc_meta *meta) +/* + * Like term_chara() but for arbitrary-length buffers. Resize the + * buffer by a factor of two (if the buffer is less than that) or the + * buffer's size. + */ +static void +term_stringa(struct termp *p, const char *c, size_t sz) { - struct termp p; + size_t s; - p.maxrmargin = 80; /* XXX */ - p.rmargin = p.maxrmargin; - p.maxcols = 1024; - p.offset = p.col = 0; - p.flags = TERMP_NOSPACE; + if (0 == sz) + return; - if (NULL == (p.buf = malloc(p.maxcols))) - err(1, "malloc"); + assert(c); + if (p->col + sz >= p->maxcols) { + if (0 == p->maxcols) + p->maxcols = 256; + s = sz > p->maxcols * 2 ? sz : p->maxcols * 2; + p->buf = realloc(p->buf, s); + if (NULL == p->buf) + err(1, "realloc"); + p->maxcols = s; + } - termprint_header(&p, meta); - termprint_r(&p, meta, node); - termprint_footer(&p, meta); - - free(p.buf); + (void)memcpy(&p->buf[(int)p->col], c, sz); + p->col += sz; } + +/* + * 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; +}