=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.129 retrieving revision 1.145 diff -u -p -r1.129 -r1.145 --- mandoc/term.c 2010/03/23 12:42:22 1.129 +++ mandoc/term.c 2010/06/08 13:22:37 1.145 @@ -1,4 +1,4 @@ -/* $Id: term.c,v 1.129 2010/03/23 12:42:22 kristaps Exp $ */ +/* $Id: term.c,v 1.145 2010/06/08 13:22:37 kristaps Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -22,11 +22,14 @@ #include #include +#include +#include #include #include #include #include +#include "mandoc.h" #include "chars.h" #include "out.h" #include "term.h" @@ -34,33 +37,67 @@ #include "mdoc.h" #include "main.h" -static struct termp *term_alloc(enum termenc); -static void term_free(struct termp *); +#define PS_CHAR_WIDTH 6 +#define PS_CHAR_HEIGHT 12 +#define PS_CHAR_TOPMARG (792 - 24) +#define PS_CHAR_TOP (PS_CHAR_TOPMARG - 36) +#define PS_CHAR_LEFT 36 +#define PS_CHAR_BOTMARG 24 +#define PS_CHAR_BOT (PS_CHAR_BOTMARG + 36) + static void spec(struct termp *, const char *, size_t); static void res(struct termp *, const char *, size_t); static void buffera(struct termp *, const char *, size_t); static void bufferc(struct termp *, char); static void adjbuf(struct termp *p, size_t); static void encode(struct termp *, const char *, size_t); +static void advance(struct termp *, size_t); +static void endline(struct termp *); +static void letter(struct termp *, char); +static void pageopen(struct termp *); void * -ascii_alloc(void) +ascii_alloc(char *outopts) { + struct termp *p; + const char *toks[2]; + char *v; - return(term_alloc(TERMENC_ASCII)); + if (NULL == (p = term_alloc(TERMENC_ASCII))) + return(NULL); + + p->type = TERMTYPE_CHAR; + + toks[0] = "width"; + toks[1] = NULL; + + while (outopts && *outopts) + switch (getsubopt(&outopts, UNCONST(toks), &v)) { + case (0): + p->defrmargin = (size_t)atoi(v); + break; + default: + break; + } + + /* Enforce a lower boundary. */ + if (p->defrmargin < 58) + p->defrmargin = 58; + + return(p); } void -terminal_free(void *arg) +ascii_free(void *arg) { term_free((struct termp *)arg); } -static void +void term_free(struct termp *p) { @@ -73,17 +110,215 @@ term_free(struct termp *p) } -static struct termp * +/* + * Push a single letter into our output engine. + */ +static void +letter(struct termp *p, char c) +{ + + if (TERMTYPE_CHAR == p->type) { + /* + * If using the terminal device, just push the letter + * out into the screen. + */ + putchar(c); + return; + } + + if ( ! (PS_INLINE & p->psstate)) { + /* + * If we're not in a PostScript "word" context, then + * open one now at the current cursor. + */ + printf("%zu %zu moveto\n", p->pscol, p->psrow); + putchar('('); + p->psstate |= PS_INLINE; + } + + /* + * We need to escape these characters as per the PostScript + * specification. We would also escape non-graphable characters + * (like tabs), but none of them would get to this point and + * it's superfluous to abort() on them. + */ + + switch (c) { + case ('('): + /* FALLTHROUGH */ + case (')'): + /* FALLTHROUGH */ + case ('\\'): + putchar('\\'); + break; + default: + break; + } + + /* Write the character and adjust where we are on the page. */ + putchar(c); + p->pscol += PS_CHAR_WIDTH; +} + + +/* + * Begin a "terminal" context. Since terminal encompasses PostScript, + * the actual terminal, etc., there are a few things we can do here. + */ +void +term_begin(struct termp *p, term_margin head, + term_margin foot, const void *arg) +{ + + p->headf = head; + p->footf = foot; + p->argf = arg; + + if (TERMTYPE_CHAR == p->type) { + /* Emit the header and be done. */ + (*p->headf)(p, p->argf); + return; + } + + /* + * Emit the standard PostScript prologue, set our initial page + * position, then run pageopen() on the initial page. + */ + + printf("%s\n", "%!PS"); + printf("%s\n", "/Courier"); + printf("%s\n", "10 selectfont"); + + p->pspage = 1; + p->psstate = 0; + pageopen(p); +} + + +/* + * Open a page. This is only used for -Tps at the moment. It opens a + * page context, printing the header and the footer. THE OUTPUT BUFFER + * MUST BE EMPTY. If it is not, output will ghost on the next line and + * we'll be all gross and out of state. + */ +static void +pageopen(struct termp *p) +{ + + assert(TERMTYPE_PS == p->type); + assert(0 == p->psstate); + + p->pscol = PS_CHAR_LEFT; + p->psrow = PS_CHAR_TOPMARG; + p->psstate |= PS_MARGINS; + + (*p->headf)(p, p->argf); + endline(p); + + p->psstate &= ~PS_MARGINS; + assert(0 == p->psstate); + + p->pscol = PS_CHAR_LEFT; + p->psrow = PS_CHAR_BOTMARG; + p->psstate |= PS_MARGINS; + + (*p->footf)(p, p->argf); + endline(p); + + p->psstate &= ~PS_MARGINS; + assert(0 == p->psstate); + + p->pscol = PS_CHAR_LEFT; + p->psrow = PS_CHAR_TOP; + +} + + +void +term_end(struct termp *p) +{ + + if (TERMTYPE_CHAR == p->type) { + (*p->footf)(p, p->argf); + return; + } + + printf("%s\n", "%%END"); +} + + +static void +endline(struct termp *p) +{ + + if (TERMTYPE_CHAR == p->type) { + putchar('\n'); + return; + } + + if (PS_INLINE & p->psstate) { + printf(") show\n"); + p->psstate &= ~PS_INLINE; + } + + if (PS_MARGINS & p->psstate) + return; + + p->pscol = PS_CHAR_LEFT; + if (p->psrow >= PS_CHAR_HEIGHT + PS_CHAR_BOT) { + p->psrow -= PS_CHAR_HEIGHT; + return; + } + + /* + * XXX: can't run pageopen() until we're certain a flushln() has + * occured, else the buf will reopen in an awkward state on the + * next line. + */ + printf("showpage\n"); + p->psrow = PS_CHAR_TOP; +} + + +/* + * Advance the output engine by a certain amount of whitespace. + */ +static void +advance(struct termp *p, size_t len) +{ + size_t i; + + if (TERMTYPE_CHAR == p->type) { + /* Just print whitespace on the terminal. */ + for (i = 0; i < len; i++) + putchar(' '); + return; + } + + if (PS_INLINE & p->psstate) { + /* Dump out any existing line scope. */ + printf(") show\n"); + p->psstate &= ~PS_INLINE; + } + + p->pscol += len ? len * PS_CHAR_WIDTH : 0; +} + + +struct termp * term_alloc(enum termenc enc) { - struct termp *p; + struct termp *p; p = calloc(1, sizeof(struct termp)); if (NULL == p) { perror(NULL); exit(EXIT_FAILURE); } + + p->tabwidth = 5; p->enc = enc; + p->defrmargin = 78; return(p); } @@ -92,13 +327,11 @@ term_alloc(enum termenc enc) * 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. + * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does + * not have a trailing newline. * - * Specifically, a line is whatever's in p->buf of length p->col, which - * is zeroed after this function returns. + * The following flags may be specified: * - * 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. @@ -131,9 +364,10 @@ term_flushln(struct termp *p) int i; /* current input position in p->buf */ size_t vis; /* current visual position on output */ size_t vbl; /* number of blanks to prepend to output */ - size_t vsz; /* visual characters to write to output */ + size_t vend; /* end of word visual position on output */ size_t bp; /* visual right border position */ int j; /* temporary loop index */ + int jhy; /* last hyphen before line overflow */ size_t maxvis, mmax; /* @@ -154,25 +388,30 @@ term_flushln(struct termp *p) bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; + /* + * Indent the first line of a paragraph. + */ + vbl = p->flags & TERMP_NOLPAD ? 0 : p->offset; + /* * FIXME: if bp is zero, we still output the first word before * breaking the line. */ - vis = 0; + vis = vend = i = 0; + while (i < (int)p->col) { - /* - * If in the standard case (left-justified), then begin with our - * indentation, otherwise (columns, etc.) just start spitting - * out text. - */ + /* + * Handle literal tab characters. + */ + for (j = i; j < (int)p->col; j++) { + if ('\t' != p->buf[j]) + break; + vend = (vis/p->tabwidth+1)*p->tabwidth; + vbl += vend - vis; + vis = vend; + } - 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 @@ -181,69 +420,93 @@ term_flushln(struct termp *p) */ /* LINTED */ - for (j = i, vsz = 0; j < (int)p->col; j++) { - if (j && ' ' == p->buf[j]) + for (jhy = 0; j < (int)p->col; j++) { + if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j]) break; - else if (8 == p->buf[j]) - vsz--; - else - vsz++; + if (8 != p->buf[j]) { + if (vend > vis && vend < bp && + ASCII_HYPH == p->buf[j]) + jhy = j; + vend++; + } else + vend--; } /* - * 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 so, break to the next line. */ - if (vis && vis + vbl + vsz > bp) { - putchar('\n'); + if (vend > bp && 0 == jhy && vis > 0) { + vend -= vis; + endline(p); if (TERMP_NOBREAK & p->flags) { - for (j = 0; j < (int)p->rmargin; j++) - putchar(' '); - vis = p->rmargin - p->offset; + p->viscol = p->rmargin; + advance(p, p->rmargin); + vend += p->rmargin - p->offset; } else { - for (j = 0; j < (int)p->offset; j++) - putchar(' '); - vis = 0; + p->viscol = 0; + vbl = p->offset; } + /* Remove the p->overstep width. */ + bp += (int)/* LINTED */ p->overstep; p->overstep = 0; - } else { - for (j = 0; j < (int)vbl; j++) - putchar(' '); - vis += vbl; } /* - * Finally, write out the word. + * Skip leading tabs, they were handled above. */ + while (i < (int)p->col && '\t' == p->buf[i]) + i++; + + /* Write out the [remaining] word. */ for ( ; i < (int)p->col; i++) { - if (' ' == p->buf[i]) + if (vend > bp && jhy > 0 && i > jhy) break; + if ('\t' == p->buf[i]) + break; + if (' ' == p->buf[i]) { + while (' ' == p->buf[i]) { + vbl++; + i++; + } + break; + } + if (ASCII_NBRSP == p->buf[i]) { + vbl++; + continue; + } - /* The unit sep. is a non-breaking space. */ - if (31 == p->buf[i]) - putchar(' '); + /* + * Now we definitely know there will be + * printable characters to output, + * so write preceding white space now. + */ + if (vbl) { + advance(p, vbl); + p->viscol += vbl; + vbl = 0; + } + + if (ASCII_HYPH == p->buf[i]) + letter(p, '-'); else - putchar(p->buf[i]); + letter(p, p->buf[i]); + + p->viscol += 1; } - vis += vsz; + vend += vbl; + vis = vend; } p->col = 0; p->overstep = 0; if ( ! (TERMP_NOBREAK & p->flags)) { - putchar('\n'); + p->viscol = 0; + endline(p); return; } @@ -274,13 +537,14 @@ term_flushln(struct termp *p) /* Right-pad. */ if (maxvis > vis + /* LINTED */ - ((TERMP_TWOSPACE & p->flags) ? 1 : 0)) - for ( ; vis < maxvis; vis++) - putchar(' '); - else { /* ...or newline break. */ - putchar('\n'); - for (i = 0; i < (int)p->rmargin; i++) - putchar(' '); + ((TERMP_TWOSPACE & p->flags) ? 1 : 0)) { + p->viscol += maxvis - vis; + advance(p, maxvis - vis); + vis += (maxvis - vis); + } else { /* ...or newline break. */ + endline(p); + p->viscol = p->rmargin; + advance(p, p->rmargin); } } @@ -295,7 +559,7 @@ term_newln(struct termp *p) { p->flags |= TERMP_NOSPACE; - if (0 == p->col) { + if (0 == p->col && 0 == p->viscol) { p->flags &= ~TERMP_NOLPAD; return; } @@ -315,7 +579,8 @@ term_vspace(struct termp *p) { term_newln(p); - putchar('\n'); + p->viscol = 0; + endline(p); } @@ -440,8 +705,6 @@ term_word(struct termp *p, const char *word) case(')'): /* FALLTHROUGH */ case(']'): - /* FALLTHROUGH */ - case('}'): if ( ! (TERMP_IGNDELIM & p->flags)) p->flags |= TERMP_NOSPACE; break; @@ -449,12 +712,17 @@ term_word(struct termp *p, const char *word) break; } - if ( ! (TERMP_NOSPACE & p->flags)) + if ( ! (TERMP_NOSPACE & p->flags)) { bufferc(p, ' '); + if (TERMP_SENTENCE & p->flags) + bufferc(p, ' '); + } if ( ! (p->flags & TERMP_NONOSPACE)) p->flags &= ~TERMP_NOSPACE; + p->flags &= ~TERMP_SENTENCE; + /* FIXME: use strcspn. */ while (*word) { @@ -495,13 +763,15 @@ term_word(struct termp *p, const char *word) p->flags |= TERMP_NOSPACE; } + /* + * Note that we don't process the pipe: the parser sees it as + * punctuation, but we don't in terms of typography. + */ if (sv[0] && 0 == sv[1]) switch (sv[0]) { case('('): /* FALLTHROUGH */ case('['): - /* FALLTHROUGH */ - case('{'): p->flags |= TERMP_NOSPACE; break; default: @@ -562,7 +832,10 @@ encode(struct termp *p, const char *word, size_t sz) * character by character. */ - if (TERMFONT_NONE == (f = term_fonttop(p))) { + if (TERMTYPE_PS == p->type) { + buffera(p, word, sz); + return; + } else if (TERMFONT_NONE == (f = term_fonttop(p))) { buffera(p, word, sz); return; }