src - OpenBSD base system

diff options


context:
space:
mode:

author	Ingo Schwarze <schwarze@cvs.openbsd.org>	2012-05-28 13:00:52 +0000
committer	Ingo Schwarze <schwarze@cvs.openbsd.org>	2012-05-28 13:00:52 +0000
commit	08973cb4d1227bb54783bdeab50035c9a6bfbec0 (patch)
tree	29302e4a7dce3f5a46a365d627f6e4d95bf18e20 /usr.bin/mandoc
parent	83c0700d2695f59ff8403c43a705f01176ecab53 (diff)

Implement the roff \z escape sequence, intended to output the next

character without advancing the cursor position; implement it to simply skip the next character, as it will usually be overwritten. With this change, the pod2man(1) preamble user-defined string \*:, intended to render as a diaeresis or umlaut diacritic above the preceding character, is rendered in a slightly less ugly way, though still not correctly. It was rendered as "z.." and is now rendered as ".". Given that the definition of \*: uses elaborate manual \h positioning, there is little chance for mandoc(1) to ever render it correctly, but at least we can refrain from printing out a spurious "z", and we can make the \z do something semi-reasonable for easier cases.

Diffstat (limited to 'usr.bin/mandoc')

-rw-r--r--

usr.bin/mandoc/html.c

-rw-r--r--

usr.bin/mandoc/html.h

-rw-r--r--

usr.bin/mandoc/mandoc.c

-rw-r--r--

usr.bin/mandoc/mandoc.h

-rw-r--r--

usr.bin/mandoc/term.c

-rw-r--r--

usr.bin/mandoc/term.h

6 files changed, 131 insertions, 45 deletions

diff --git a/usr.bin/mandoc/html.c b/usr.bin/mandoc/html.c
index 80f6acc6956..41e9814955f 100644
--- a/usr.bin/mandoc/html.c
+++ b/usr.bin/mandoc/html.c

@@ -1,7 +1,7 @@

-/* $Id: html.c,v 1.29 2011/10/09 17:59:56 schwarze Exp $ */

+/* $Id: html.c,v 1.30 2012/05/28 13:00:51 schwarze Exp $ */

* Permission to use, copy, modify, and distribute this software for any

* purpose with or without fee is hereby granted, provided that the above

@@ -258,8 +258,8 @@ print_metaf(struct html *h, enum mandoc_esc deco)

int

html_strlen(const char *cp)

{

- int ssz, sz;

- const char *seq, *p;

+ size_t rsz;

+ int skip, sz;

* Account for escaped sequences within string length

@@ -270,10 +270,21 @@ html_strlen(const char *cp)

sz = 0;

- while (NULL != (p = strchr(cp, '\\'))) {

- sz += (int)(p - cp);

- ++cp;

- switch (mandoc_escape(&cp, &seq, &ssz)) {

+ skip = 0;

+ while (1) {

+ rsz = strcspn(cp, "\\");

+ if (rsz) {

+ cp += rsz;

+ if (skip) {

+ skip = 0;

+ rsz--;

+ }

+ sz += rsz;

+ }

+ if ('\0' == *cp)

+ break;

+ cp++;

+ switch (mandoc_escape(&cp, NULL, NULL)) {

case (ESCAPE_ERROR):

return(sz);

case (ESCAPE_UNICODE):

@@ -281,15 +292,19 @@ html_strlen(const char *cp)

case (ESCAPE_NUMBERED):

/* FALLTHROUGH */

case (ESCAPE_SPECIAL):

- sz++;

+ if (skip)

+ skip = 0;

+ else

+ sz++;

+ break;

+ case (ESCAPE_SKIPCHAR):

+ skip = 1;

break;

default:

break;

}

- assert(sz >= 0);

- return(sz + strlen(cp));

+ return(sz);

}

static int

@@ -304,6 +319,12 @@ print_encode(struct html *h, const char *p, int norecurse)

nospace = 0;

while ('\0' != *p) {

+ if (HTML_SKIPCHAR & h->flags && '\\' != *p) {

+ h->flags &= ~HTML_SKIPCHAR;

+ p++;

+ continue;

+ }

sz = strcspn(p, rejs);

fwrite(p, 1, sz, stdout);

@@ -334,6 +355,31 @@ print_encode(struct html *h, const char *p, int norecurse)

break;

switch (esc) {

+ case (ESCAPE_FONT):

+ /* FALLTHROUGH */

+ case (ESCAPE_FONTPREV):

+ /* FALLTHROUGH */

+ case (ESCAPE_FONTBOLD):

+ /* FALLTHROUGH */

+ case (ESCAPE_FONTITALIC):

+ /* FALLTHROUGH */

+ case (ESCAPE_FONTROMAN):

+ if (0 == norecurse)

+ print_metaf(h, esc);

+ continue;

+ case (ESCAPE_SKIPCHAR):

+ h->flags |= HTML_SKIPCHAR;

+ continue;

+ default:

+ break;

+ }

+ if (h->flags & HTML_SKIPCHAR) {

+ h->flags &= ~HTML_SKIPCHAR;

+ continue;

+ }

+ switch (esc) {

case (ESCAPE_UNICODE):

/* Skip passed "u" header. */

c = mchars_num2uc(seq + 1, len - 1);

@@ -352,19 +398,6 @@ print_encode(struct html *h, const char *p, int norecurse)

else if (-1 == c && 1 == len)

putchar((int)*seq);

break;

- case (ESCAPE_FONT):

- /* FALLTHROUGH */

- case (ESCAPE_FONTPREV):

- /* FALLTHROUGH */

- case (ESCAPE_FONTBOLD):

- /* FALLTHROUGH */

- case (ESCAPE_FONTITALIC):

- /* FALLTHROUGH */

- case (ESCAPE_FONTROMAN):

- if (norecurse)

- break;

- print_metaf(h, esc);

- break;

case (ESCAPE_NOSPACE):

if ('\0' == *p)

nospace = 1;

diff --git a/usr.bin/mandoc/html.h b/usr.bin/mandoc/html.h
index bcc6d6f7c39..d3a4307c4c7 100644
--- a/usr.bin/mandoc/html.h
+++ b/usr.bin/mandoc/html.h

@@ -1,4 +1,4 @@

-/* $Id: html.h,v 1.18 2011/10/09 17:59:56 schwarze Exp $ */

+/* $Id: html.h,v 1.19 2012/05/28 13:00:51 schwarze Exp $ */

@@ -117,6 +117,7 @@ struct html {

#define HTML_PREKEEP (1 << 3)

#define HTML_NONOSPACE (1 << 4) /* never add spaces */

#define HTML_LITERAL (1 << 5) /* literal (e.g., <PRE>) context */

+#define HTML_SKIPCHAR (1 << 6) /* skip the next character */

struct tagq tags; /* stack of open tags */

struct rofftbl tbl; /* current table */

struct tag *tblt; /* current open table scope */

diff --git a/usr.bin/mandoc/mandoc.c b/usr.bin/mandoc/mandoc.c
index 842cea4b3b2..2b21522948c 100644
--- a/usr.bin/mandoc/mandoc.c
+++ b/usr.bin/mandoc/mandoc.c

@@ -1,7 +1,7 @@

-/* $Id: mandoc.c,v 1.31 2011/11/17 11:58:11 schwarze Exp $ */

+/* $Id: mandoc.c,v 1.32 2012/05/28 13:00:51 schwarze Exp $ */

* Permission to use, copy, modify, and distribute this software for any

* purpose with or without fee is hereby granted, provided that the above

@@ -139,6 +139,16 @@ mandoc_escape(const char **end, const char **start, int *sz)

break;

+ * The \z escape is supposed to output the following

+ * character without advancing the cursor position.

+ * Since we are mostly dealing with terminal mode,

+ * let us just skip the next character.

+ */

+ case ('z'):

+ (*end)++;

+ return(ESCAPE_SKIPCHAR);

+ /*

* Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where

* 'X' is the trigger. These have opaque sub-strings.

diff --git a/usr.bin/mandoc/mandoc.h b/usr.bin/mandoc/mandoc.h
index a66bb0d74ad..e91f2e7bf2e 100644
--- a/usr.bin/mandoc/mandoc.h
+++ b/usr.bin/mandoc/mandoc.h

@@ -1,4 +1,4 @@

-/* $Id: mandoc.h,v 1.45 2012/05/26 20:03:34 schwarze Exp $ */

+/* $Id: mandoc.h,v 1.46 2012/05/28 13:00:51 schwarze Exp $ */

@@ -379,7 +379,8 @@ enum mandoc_esc {

ESCAPE_FONTPREV, /* previous font mode */

ESCAPE_NUMBERED, /* a numbered glyph */

ESCAPE_UNICODE, /* a unicode codepoint */

- ESCAPE_NOSPACE /* suppress space if the last on a line */

+ ESCAPE_NOSPACE, /* suppress space if the last on a line */

+ ESCAPE_SKIPCHAR /* skip the next character */

};

typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel,

diff --git a/usr.bin/mandoc/term.c b/usr.bin/mandoc/term.c
index acbe2c802ef..a30c2440d39 100644
--- a/usr.bin/mandoc/term.c
+++ b/usr.bin/mandoc/term.c

@@ -1,4 +1,4 @@

-/* $Id: term.c,v 1.63 2012/05/27 01:01:24 schwarze Exp $ */

+/* $Id: term.c,v 1.64 2012/05/28 13:00:51 schwarze Exp $ */

@@ -29,6 +29,7 @@

#include "term.h"

#include "main.h"

+static size_t cond_width(const struct termp *, int, int *);

static void adjbuf(struct termp *p, int);

static void bufferc(struct termp *, char);

static void encode(struct termp *, const char *, size_t);

@@ -415,12 +416,17 @@ term_word(struct termp *p, const char *word)

p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM);

while ('\0' != *word) {

- if ((ssz = strcspn(word, "\\")) > 0)

+ if ('\\' != *word) {

+ if (TERMP_SKIPCHAR & p->flags) {

+ p->flags &= ~TERMP_SKIPCHAR;

+ word++;

+ continue;

+ }

+ ssz = strcspn(word, "\\");

encode(p, word, ssz);

- word += (int)ssz;

- if ('\\' != *word)

+ word += (int)ssz;

continue;

+ }

word++;

esc = mandoc_escape(&word, &seq, &sz);

@@ -476,9 +482,14 @@ term_word(struct termp *p, const char *word)

term_fontlast(p);

break;

case (ESCAPE_NOSPACE):

- if ('\0' == *word)

+ if (TERMP_SKIPCHAR & p->flags)

+ p->flags &= ~TERMP_SKIPCHAR;

+ else if ('\0' == *word)

p->flags |= TERMP_NOSPACE;

break;

+ case (ESCAPE_SKIPCHAR):

+ p->flags |= TERMP_SKIPCHAR;

+ break;

default:

break;

}

@@ -518,6 +529,11 @@ encode1(struct termp *p, int c)

{

enum termfont f;

+ if (TERMP_SKIPCHAR & p->flags) {

+ p->flags &= ~TERMP_SKIPCHAR;

+ return;

+ }

if (p->col + 4 >= p->maxcols)

adjbuf(p, p->col + 4);

@@ -541,6 +557,11 @@ encode(struct termp *p, const char *word, size_t sz)

enum termfont f;

int i, len;

+ if (TERMP_SKIPCHAR & p->flags) {

+ p->flags &= ~TERMP_SKIPCHAR;

+ return;

+ }

/* LINTED */

len = sz;

@@ -589,12 +610,22 @@ term_len(const struct termp *p, size_t sz)

return((*p->width)(p, ' ') * sz);

}

+static size_t

+cond_width(const struct termp *p, int c, int *skip)

+ if (*skip) {

+ (*skip) = 0;

+ return(0);

+ } else

+ return((*p->width)(p, c));

size_t

term_strlen(const struct termp *p, const char *cp)

{

size_t sz, rsz, i;

- int ssz, c;

+ int ssz, skip, c;

const char *seq, *rhs;

enum mandoc_esc esc;

static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' };

@@ -606,10 +637,11 @@ term_strlen(const struct termp *p, const char *cp)

sz = 0;

+ skip = 0;

while ('\0' != *cp) {

rsz = strcspn(cp, rej);

for (i = 0; i < rsz; i++)

- sz += (*p->width)(p, *cp++);

+ sz += cond_width(p, *cp++, &skip);

c = 0;

switch (*cp) {

@@ -626,14 +658,14 @@ term_strlen(const struct termp *p, const char *cp)

(seq + 1, ssz - 1);

if ('\0' == c)

break;

- sz += (*p->width)(p, c);

+ sz += cond_width(p, c, &skip);

continue;

case (ESCAPE_SPECIAL):

c = mchars_spec2cp

(p->symtab, seq, ssz);

if (c <= 0)

break;

- sz += (*p->width)(p, c);

+ sz += cond_width(p, c, &skip);

continue;

default:

break;

@@ -643,12 +675,12 @@ term_strlen(const struct termp *p, const char *cp)

switch (esc) {

case (ESCAPE_UNICODE):

- sz += (*p->width)(p, '?');

+ sz += cond_width(p, '?', &skip);

break;

case (ESCAPE_NUMBERED):

c = mchars_num2char(seq, ssz);

if ('\0' != c)

- sz += (*p->width)(p, c);

+ sz += cond_width(p, c, &skip);

break;

case (ESCAPE_SPECIAL):

rhs = mchars_spec2str

@@ -660,6 +692,9 @@ term_strlen(const struct termp *p, const char *cp)

rhs = seq;

rsz = ssz;

break;

+ case (ESCAPE_SKIPCHAR):

+ skip = 1;

+ break;

default:

break;

}

@@ -667,15 +702,20 @@ term_strlen(const struct termp *p, const char *cp)

if (NULL == rhs)

break;

+ if (skip) {

+ skip = 0;

+ break;

+ }

for (i = 0; i < rsz; i++)

sz += (*p->width)(p, *rhs++);

break;

case (ASCII_NBRSP):

- sz += (*p->width)(p, ' ');

+ sz += cond_width(p, ' ', &skip);

cp++;

break;

case (ASCII_HYPH):

- sz += (*p->width)(p, '-');

+ sz += cond_width(p, '-', &skip);

cp++;

break;

default:

diff --git a/usr.bin/mandoc/term.h b/usr.bin/mandoc/term.h
index 92a74e535a8..7317725e56f 100644
--- a/usr.bin/mandoc/term.h
+++ b/usr.bin/mandoc/term.h

@@ -1,4 +1,4 @@

-/* $Id: term.h,v 1.35 2012/05/27 01:01:24 schwarze Exp $ */

+/* $Id: term.h,v 1.36 2012/05/28 13:00:51 schwarze Exp $ */

@@ -78,6 +78,7 @@ struct termp {

#define TERMP_ANPREC (1 << 13) /* See termp_an_pre(). */

#define TERMP_KEEP (1 << 14) /* Keep words together. */

#define TERMP_PREKEEP (1 << 15) /* ...starting with the next one. */

+#define TERMP_SKIPCHAR (1 << 16) /* Skip the next character. */

int *buf; /* Output buffer. */

enum termenc enc; /* Type of encoding. */

struct mchars *symtab; /* Encoded-symbol table. */