version 1.75, 2009/06/10 20:18:44 |
version 1.96, 2009/07/27 13:10:08 |
Line 31 extern int mdoc_run(struct termp *, |
|
Line 31 extern int mdoc_run(struct termp *, |
|
|
|
static struct termp *term_alloc(enum termenc); |
static struct termp *term_alloc(enum termenc); |
static void term_free(struct termp *); |
static void term_free(struct termp *); |
static void term_pword(struct termp *, const char *, int); |
|
static void term_pescape(struct termp *, |
static void do_escaped(struct termp *, const char **); |
const char *, int *, int); |
static void do_special(struct termp *, |
static void term_nescape(struct termp *, |
|
const char *, size_t); |
const char *, size_t); |
static void term_chara(struct termp *, char); |
static void do_reserved(struct termp *, |
static void term_stringa(struct termp *, |
|
const char *, size_t); |
const char *, size_t); |
static int term_isopendelim(const char *, int); |
static void buffer(struct termp *, char); |
static int term_isclosedelim(const char *, int); |
static void encode(struct termp *, char); |
|
static int isopendelim(const char *); |
|
static int isclosedelim(const char *); |
|
|
|
|
void * |
void * |
Line 113 term_alloc(enum termenc enc) |
|
Line 113 term_alloc(enum termenc enc) |
|
|
|
|
|
static int |
static int |
term_isclosedelim(const char *p, int len) |
isclosedelim(const char *p) |
{ |
{ |
|
|
if (1 != len) |
if ( ! (*p && 0 == *(p + 1))) |
return(0); |
return(0); |
|
|
switch (*p) { |
switch (*p) { |
Line 147 term_isclosedelim(const char *p, int len) |
|
Line 147 term_isclosedelim(const char *p, int len) |
|
|
|
|
|
static int |
static int |
term_isopendelim(const char *p, int len) |
isopendelim(const char *p) |
{ |
{ |
|
|
if (1 != len) |
if ( ! (*p && 0 == *(p + 1))) |
return(0); |
return(0); |
|
|
switch (*p) { |
switch (*p) { |
Line 177 term_isopendelim(const char *p, int len) |
|
Line 177 term_isopendelim(const char *p, int len) |
|
* Specifically, a line is whatever's in p->buf of length p->col, which |
* Specifically, a line is whatever's in p->buf of length p->col, which |
* is zeroed after this function returns. |
* is zeroed after this function returns. |
* |
* |
* The variables TERMP_NOLPAD, TERMP_LITERAL and TERMP_NOBREAK are of |
* The usage of termp:flags is as follows: |
* critical importance here. Their behaviour follows: |
|
* |
* |
* - TERMP_NOLPAD: when beginning to write the line, don't left-pad the |
* - TERMP_NOLPAD: when beginning to write the line, don't left-pad the |
* offset value. This is useful when doing columnar lists where the |
* offset value. This is useful when doing columnar lists where the |
Line 188 term_isopendelim(const char *p, int len) |
|
Line 187 term_isopendelim(const char *p, int len) |
|
* columns. In short: don't print a newline and instead pad to the |
* columns. In short: don't print a newline and instead pad to the |
* right margin. Used in conjunction with TERMP_NOLPAD. |
* right margin. Used in conjunction with TERMP_NOLPAD. |
* |
* |
* - TERMP_NONOBREAK: don't newline when TERMP_NOBREAK is specified. |
* - TERMP_TWOSPACE: when padding, make sure there are at least two |
|
* space characters of padding. Otherwise, rather break the line. |
* |
* |
|
* - 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: |
* In-line line breaking: |
* |
* |
* If TERMP_NOBREAK is specified and the line overruns the right |
* If TERMP_NOBREAK is specified and the line overruns the right |
|
|
term_flushln(struct termp *p) |
term_flushln(struct termp *p) |
{ |
{ |
int i, j; |
int i, j; |
size_t vsz, vis, maxvis, mmax, bp; |
size_t vbl, vsz, vis, maxvis, mmax, bp; |
|
static int overstep = 0; |
|
|
/* |
/* |
* First, establish the maximum columns of "visible" content. |
* First, establish the maximum columns of "visible" content. |
Line 216 term_flushln(struct termp *p) |
|
Line 224 term_flushln(struct termp *p) |
|
*/ |
*/ |
|
|
assert(p->offset < p->rmargin); |
assert(p->offset < p->rmargin); |
maxvis = p->rmargin - p->offset; |
assert((int)(p->rmargin - p->offset) - overstep > 0); |
mmax = p->maxrmargin - p->offset; |
|
|
maxvis = /* LINTED */ |
|
p->rmargin - p->offset - overstep; |
|
mmax = /* LINTED */ |
|
p->maxrmargin - p->offset - overstep; |
|
|
bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; |
bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; |
vis = 0; |
vis = 0; |
|
overstep = 0; |
|
|
/* |
/* |
* If in the standard case (left-justified), then begin with our |
* If in the standard case (left-justified), then begin with our |
Line 242 term_flushln(struct termp *p) |
|
Line 256 term_flushln(struct termp *p) |
|
|
|
/* LINTED */ |
/* LINTED */ |
for (j = i, vsz = 0; j < (int)p->col; j++) { |
for (j = i, vsz = 0; j < (int)p->col; j++) { |
if (' ' == p->buf[j]) |
if (j && ' ' == p->buf[j]) |
break; |
break; |
else if (8 == p->buf[j]) |
else if (8 == p->buf[j]) |
j += 1; |
vsz--; |
else |
else |
vsz++; |
vsz++; |
} |
} |
|
|
/* |
/* |
* Do line-breaking. If we're greater than our |
* Choose the number of blanks to prepend: no blank at the |
* break-point and already in-line, break to the next |
* beginning of a line, one between words -- but do not |
* line and start writing. If we're at the line start, |
* actually write them yet. |
* then write out the word (TODO: hyphenate) and break |
|
* in a subsequent loop invocation. |
|
*/ |
*/ |
|
vbl = (size_t)(0 == vis ? 0 : 1); |
|
|
if ( ! (TERMP_NOBREAK & p->flags)) { |
/* |
if (vis && vis + vsz > bp) { |
* Find out whether we would exceed the right margin. |
putchar('\n'); |
* 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++) |
for (j = 0; j < (int)p->offset; j++) |
putchar(' '); |
putchar(' '); |
vis = 0; |
vis = 0; |
} |
} |
} else if (vis && vis + vsz > bp) { |
} else { |
putchar('\n'); |
for (j = 0; j < (int)vbl; j++) |
for (j = 0; j < (int)p->rmargin; j++) |
|
putchar(' '); |
putchar(' '); |
vis = p->rmargin - p->offset; |
vis += vbl; |
} |
} |
|
|
/* |
/* |
* Write out the word and a trailing space. Omit the |
* Finally, write out the word. |
* space if we're the last word in the line or beyond |
|
* our breakpoint. |
|
*/ |
*/ |
|
|
for ( ; i < (int)p->col; i++) { |
for ( ; i < (int)p->col; i++) { |
if (' ' == p->buf[i]) |
if (' ' == p->buf[i]) |
break; |
break; |
putchar(p->buf[i]); |
putchar(p->buf[i]); |
} |
} |
vis += vsz; |
vis += vsz; |
if (i < (int)p->col && vis <= bp) { |
|
putchar(' '); |
|
vis++; |
|
} |
|
} |
} |
|
p->col = 0; |
|
|
/* |
if ( ! (TERMP_NOBREAK & p->flags)) { |
* If we've overstepped our maximum visible no-break space, then |
putchar('\n'); |
* 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; |
return; |
} |
} |
|
|
/* |
if (TERMP_HANG & p->flags) { |
* If we're not to right-marginalise it (newline), then instead |
/* We need one blank after the tag. */ |
* pad to the right margin and stay off. |
overstep = /* LINTED */ |
*/ |
vis - maxvis + 1; |
|
|
if (p->flags & TERMP_NOBREAK) { |
/* |
if ( ! (TERMP_NONOBREAK & p->flags)) |
* Behave exactly the same way as groff: |
for ( ; vis < maxvis; vis++) |
* If we have overstepped the margin, temporarily move |
putchar(' '); |
* it to the right and flag the rest of the line to be |
} else |
* shorter. |
putchar('\n'); |
* If we landed right at the margin, be happy. |
|
* If we are one step before the margin, temporarily |
|
* move it one step LEFT and flag the rest of the line |
|
* to be longer. |
|
*/ |
|
if (overstep >= -1) { |
|
assert((int)maxvis + overstep >= 0); |
|
/* LINTED */ |
|
maxvis += overstep; |
|
} else |
|
overstep = 0; |
|
|
p->col = 0; |
} else if (TERMP_DANGLE & p->flags) |
|
return; |
|
|
|
/* 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(' '); |
|
} |
} |
} |
|
|
|
|
Line 355 term_vspace(struct termp *p) |
|
Line 382 term_vspace(struct termp *p) |
|
} |
} |
|
|
|
|
/* |
static void |
* Break apart a word into "pwords" (partial-words, usually from |
do_special(struct termp *p, const char *word, size_t len) |
* 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; |
const char *rhs; |
|
size_t sz; |
|
int i; |
|
|
len = (int)strlen(word); |
rhs = term_a2ascii(p->symtab, word, len, &sz); |
|
|
if (p->flags & TERMP_LITERAL) { |
if (NULL == rhs) { |
term_pword(p, word, len); |
#if 1 |
|
fputs("Unknown special character: ", stderr); |
|
for (i = 0; i < (int)len; i++) |
|
fputc(word[i], stderr); |
|
fputc('\n', stderr); |
|
#endif |
return; |
return; |
} |
} |
|
for (i = 0; i < (int)sz; i++) |
/* LINTED */ |
encode(p, rhs[i]); |
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; |
|
} |
|
|
|
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 |
static void |
term_nescape(struct termp *p, const char *word, size_t len) |
do_reserved(struct termp *p, const char *word, size_t len) |
{ |
{ |
const char *rhs; |
const char *rhs; |
size_t sz; |
size_t sz; |
|
int i; |
|
|
if (NULL == (rhs = term_a2ascii(p->symtab, word, len, &sz))) |
rhs = term_a2res(p->symtab, word, len, &sz); |
|
|
|
if (NULL == rhs) { |
|
#if 0 |
|
fputs("Unknown reserved word: ", stderr); |
|
for (i = 0; i < (int)len; i++) |
|
fputc(word[i], stderr); |
|
fputc('\n', stderr); |
|
#endif |
return; |
return; |
term_stringa(p, rhs, sz); |
} |
|
for (i = 0; i < (int)sz; i++) |
|
encode(p, rhs[i]); |
} |
} |
|
|
|
|
Line 422 term_nescape(struct termp *p, const char *word, size_t |
|
Line 434 term_nescape(struct termp *p, const char *word, size_t |
|
* the escape sequence (we assert upon badly-formed escape sequences). |
* the escape sequence (we assert upon badly-formed escape sequences). |
*/ |
*/ |
static void |
static void |
term_pescape(struct termp *p, const char *word, int *i, int len) |
do_escaped(struct termp *p, const char **word) |
{ |
{ |
int j; |
int j; |
|
const char *wp; |
|
|
if (++(*i) >= len) |
wp = *word; |
|
|
|
if (0 == *(++wp)) { |
|
*word = wp; |
return; |
return; |
|
} |
|
|
if ('(' == word[*i]) { |
if ('(' == *wp) { |
(*i)++; |
wp++; |
if (*i + 1 >= len) |
if (0 == *wp || 0 == *(wp + 1)) { |
|
*word = 0 == *wp ? wp : wp + 1; |
return; |
return; |
|
} |
|
|
term_nescape(p, &word[*i], 2); |
do_special(p, wp, 2); |
(*i)++; |
*word = ++wp; |
return; |
return; |
|
|
} else if ('*' == word[*i]) { |
} else if ('*' == *wp) { |
(*i)++; |
if (0 == *(++wp)) { |
if (*i >= len) |
*word = wp; |
return; |
return; |
|
} |
|
|
switch (word[*i]) { |
switch (*wp) { |
case ('('): |
case ('('): |
(*i)++; |
wp++; |
if (*i + 1 >= len) |
if (0 == *wp || 0 == *(wp + 1)) { |
|
*word = 0 == *wp ? wp : wp + 1; |
return; |
return; |
|
} |
|
|
term_nescape(p, &word[*i], 2); |
do_reserved(p, wp, 2); |
(*i)++; |
*word = ++wp; |
return; |
return; |
case ('['): |
case ('['): |
break; |
break; |
default: |
default: |
term_nescape(p, &word[*i], 1); |
do_reserved(p, wp, 1); |
|
*word = wp; |
return; |
return; |
} |
} |
|
|
} else if ('f' == word[*i]) { |
} else if ('f' == *wp) { |
(*i)++; |
if (0 == *(++wp)) { |
if (*i >= len) |
*word = wp; |
return; |
return; |
switch (word[*i]) { |
} |
|
|
|
switch (*wp) { |
case ('B'): |
case ('B'): |
p->flags |= TERMP_BOLD; |
p->flags |= TERMP_BOLD; |
break; |
break; |
Line 478 term_pescape(struct termp *p, const char *word, int *i |
|
Line 503 term_pescape(struct termp *p, const char *word, int *i |
|
default: |
default: |
break; |
break; |
} |
} |
|
|
|
*word = wp; |
return; |
return; |
|
|
} else if ('[' != word[*i]) { |
} else if ('[' != *wp) { |
term_nescape(p, &word[*i], 1); |
do_special(p, wp, 1); |
|
*word = wp; |
return; |
return; |
} |
} |
|
|
(*i)++; |
wp++; |
for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++) |
for (j = 0; *wp && ']' != *wp; wp++, j++) |
/* Loop... */ ; |
/* Loop... */ ; |
|
|
if (0 == word[*i]) |
if (0 == *wp) { |
|
*word = wp; |
return; |
return; |
|
} |
|
|
term_nescape(p, &word[*i - j], (size_t)j); |
do_special(p, wp - j, (size_t)j); |
|
*word = wp; |
} |
} |
|
|
|
|
Line 501 term_pescape(struct termp *p, const char *word, int *i |
|
Line 532 term_pescape(struct termp *p, const char *word, int *i |
|
* phrase that cannot be broken down (such as a literal string). This |
* phrase that cannot be broken down (such as a literal string). This |
* handles word styling. |
* handles word styling. |
*/ |
*/ |
static void |
void |
term_pword(struct termp *p, const char *word, int len) |
term_word(struct termp *p, const char *word) |
{ |
{ |
int i; |
const char *sv; |
|
|
if (term_isclosedelim(word, len)) |
if (isclosedelim(word)) |
if ( ! (TERMP_IGNDELIM & p->flags)) |
if ( ! (TERMP_IGNDELIM & p->flags)) |
p->flags |= TERMP_NOSPACE; |
p->flags |= TERMP_NOSPACE; |
|
|
if ( ! (TERMP_NOSPACE & p->flags)) |
if ( ! (TERMP_NOSPACE & p->flags)) |
term_chara(p, ' '); |
buffer(p, ' '); |
|
|
if ( ! (p->flags & TERMP_NONOSPACE)) |
if ( ! (p->flags & TERMP_NONOSPACE)) |
p->flags &= ~TERMP_NOSPACE; |
p->flags &= ~TERMP_NOSPACE; |
|
|
/* |
for (sv = word; *word; word++) |
* If ANSI (word-length styling), then apply our style now, |
if ('\\' != *word) |
* before the word. |
encode(p, *word); |
*/ |
else |
|
do_escaped(p, &word); |
|
|
for (i = 0; i < len; i++) { |
if (isopendelim(sv)) |
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; |
p->flags |= TERMP_NOSPACE; |
} |
} |
|
|
|
|
/* |
/* |
* 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) |
|
{ |
|
size_t s; |
|
|
|
if (0 == sz) |
|
return; |
|
|
|
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; |
|
} |
|
|
|
(void)memcpy(&p->buf[(int)p->col], c, sz); |
|
p->col += sz; |
|
} |
|
|
|
|
|
/* |
|
* Insert a single character into the line-buffer. If the buffer's |
* Insert a single character into the line-buffer. If the buffer's |
* space is exceeded, then allocate more space by doubling the buffer |
* space is exceeded, then allocate more space by doubling the buffer |
* size. |
* size. |
*/ |
*/ |
static void |
static void |
term_chara(struct termp *p, char c) |
buffer(struct termp *p, char c) |
{ |
{ |
size_t s; |
size_t s; |
|
|
Line 597 term_chara(struct termp *p, char c) |
|
Line 580 term_chara(struct termp *p, char c) |
|
p->buf[(int)(p->col)++] = c; |
p->buf[(int)(p->col)++] = c; |
} |
} |
|
|
|
|
|
static void |
|
encode(struct termp *p, char c) |
|
{ |
|
|
|
if (' ' != c && TERMP_STYLE & p->flags) { |
|
if (TERMP_BOLD & p->flags) { |
|
buffer(p, c); |
|
buffer(p, 8); |
|
} |
|
if (TERMP_UNDER & p->flags) { |
|
buffer(p, '_'); |
|
buffer(p, 8); |
|
} |
|
} |
|
buffer(p, c); |
|
} |