version 1.386, 2022/04/30 18:51:36 |
version 1.394, 2022/06/07 09:41:59 |
Line 207 static int roff_evalpar(struct roff *, int, |
|
Line 207 static int roff_evalpar(struct roff *, int, |
|
static int roff_evalstrcond(const char *, int *); |
static int roff_evalstrcond(const char *, int *); |
static int roff_expand(struct roff *, struct buf *, |
static int roff_expand(struct roff *, struct buf *, |
int, int, char); |
int, int, char); |
|
static void roff_expand_patch(struct buf *, int, |
|
const char *, int); |
static void roff_free1(struct roff *); |
static void roff_free1(struct roff *); |
static void roff_freereg(struct roffreg *); |
static void roff_freereg(struct roffreg *); |
static void roff_freestr(struct roffkv *); |
static void roff_freestr(struct roffkv *); |
Line 234 static int roff_nr(ROFF_ARGS); |
|
Line 236 static int roff_nr(ROFF_ARGS); |
|
static int roff_onearg(ROFF_ARGS); |
static int roff_onearg(ROFF_ARGS); |
static enum roff_tok roff_parse(struct roff *, char *, int *, |
static enum roff_tok roff_parse(struct roff *, char *, int *, |
int, int); |
int, int); |
|
static int roff_parse_comment(struct roff *, struct buf *, |
|
int, int, char); |
static int roff_parsetext(struct roff *, struct buf *, |
static int roff_parsetext(struct roff *, struct buf *, |
int, int *); |
int, int *); |
static int roff_renamed(ROFF_ARGS); |
static int roff_renamed(ROFF_ARGS); |
Line 1232 deroff(char **dest, const struct roff_node *n) |
|
Line 1236 deroff(char **dest, const struct roff_node *n) |
|
/* --- main functions of the roff parser ---------------------------------- */ |
/* --- main functions of the roff parser ---------------------------------- */ |
|
|
/* |
/* |
* In the current line, expand escape sequences that produce parsable |
* Save comments preceding the title macro, for example in order to |
* input text. Also check the syntax of the remaining escape sequences, |
* preserve Copyright and license headers in HTML output, |
* which typically produce output glyphs or change formatter state. |
* provide diagnostics about RCS ids and trailing whitespace in comments, |
|
* then discard comments including preceding whitespace. |
|
* This function also handles input line continuation. |
*/ |
*/ |
static int |
static int |
roff_expand(struct roff *r, struct buf *buf, int ln, int pos, char newesc) |
roff_parse_comment(struct roff *r, struct buf *buf, int ln, int pos, char ec) |
{ |
{ |
struct mctx *ctx; /* current macro call context */ |
|
char ubuf[24]; /* buffer to print the number */ |
|
struct roff_node *n; /* used for header comments */ |
struct roff_node *n; /* used for header comments */ |
const char *start; /* start of the string to process */ |
const char *start; /* start of the string to process */ |
|
const char *cp; /* for RCS id parsing */ |
char *stesc; /* start of an escape sequence ('\\') */ |
char *stesc; /* start of an escape sequence ('\\') */ |
const char *esct; /* type of esccape sequence */ |
|
char *ep; /* end of comment string */ |
char *ep; /* end of comment string */ |
const char *stnam; /* start of the name, after "[(*" */ |
|
const char *cp; /* end of the name, e.g. before ']' */ |
|
const char *res; /* the string to be substituted */ |
|
char *nbuf; /* new buffer to copy buf->buf to */ |
|
size_t maxl; /* expected length of the escape name */ |
|
size_t naml; /* actual length of the escape name */ |
|
size_t asz; /* length of the replacement */ |
|
size_t rsz; /* length of the rest of the string */ |
|
int inaml; /* length returned from mandoc_escape() */ |
|
int expand_count; /* to avoid infinite loops */ |
|
int npos; /* position in numeric expression */ |
|
int arg_complete; /* argument not interrupted by eol */ |
|
int quote_args; /* true for \\$@, false for \\$* */ |
|
int done; /* no more input available */ |
|
int deftype; /* type of definition to paste */ |
|
int rcsid; /* kind of RCS id seen */ |
int rcsid; /* kind of RCS id seen */ |
enum mandocerr err; /* for escape sequence problems */ |
|
char sign; /* increment number register */ |
|
char term; /* character terminating the escape */ |
|
|
|
/* Search forward for comments. */ |
for (start = stesc = buf->buf + pos;; stesc++) { |
|
/* |
done = 0; |
* XXX Ugly hack: Remove the newline character that |
start = buf->buf + pos; |
* mparse_buf_r() appended to mark the end of input |
for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) { |
* if it is not preceded by an escape character. |
if (stesc[0] != newesc || stesc[1] == '\0') |
*/ |
continue; |
if (stesc[0] == '\n') { |
stesc++; |
assert(stesc[1] == '\0'); |
if (*stesc != '"' && *stesc != '#') |
stesc[0] = '\0'; |
continue; |
|
|
|
/* Comment found, look for RCS id. */ |
|
|
|
rcsid = 0; |
|
if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) { |
|
rcsid = 1 << MANDOC_OS_OPENBSD; |
|
cp += 8; |
|
} else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) { |
|
rcsid = 1 << MANDOC_OS_NETBSD; |
|
cp += 7; |
|
} |
} |
if (cp != NULL && |
|
isalnum((unsigned char)*cp) == 0 && |
|
strchr(cp, '$') != NULL) { |
|
if (r->man->meta.rcsids & rcsid) |
|
mandoc_msg(MANDOCERR_RCS_REP, ln, |
|
(int)(stesc - buf->buf) + 1, |
|
"%s", stesc + 1); |
|
r->man->meta.rcsids |= rcsid; |
|
} |
|
|
|
/* Handle trailing whitespace. */ |
/* The line ends without continuation or comment. */ |
|
if (stesc[0] == '\0') |
|
return ROFF_CONT; |
|
|
ep = strchr(stesc--, '\0') - 1; |
/* Unescaped byte: skip it. */ |
if (*ep == '\n') { |
if (stesc[0] != ec) |
done = 1; |
continue; |
ep--; |
|
} |
|
if (*ep == ' ' || *ep == '\t') |
|
mandoc_msg(MANDOCERR_SPACE_EOL, |
|
ln, (int)(ep - buf->buf), NULL); |
|
|
|
/* |
/* |
* Save comments preceding the title macro |
* XXX Ugly hack: Do not attempt to append another line |
* in the syntax tree. |
* if the function mparse_buf_r() appended a newline |
|
* character to indicate the end of input. |
*/ |
*/ |
|
if (stesc[1] == '\n') { |
if (newesc != ASCII_ESC && r->options & MPARSE_COMMENT) { |
assert(stesc[2] == '\0'); |
while (*ep == ' ' || *ep == '\t') |
stesc[0] = '\0'; |
ep--; |
return ROFF_CONT; |
ep[1] = '\0'; |
|
n = roff_node_alloc(r->man, |
|
ln, stesc + 1 - buf->buf, |
|
ROFFT_COMMENT, TOKEN_NONE); |
|
n->string = mandoc_strdup(stesc + 2); |
|
roff_node_append(r->man, n); |
|
n->flags |= NODE_VALID | NODE_ENDED; |
|
r->man->next = ROFF_NEXT_SIBLING; |
|
} |
} |
|
|
/* Line continuation with comment. */ |
/* |
|
* An escape character at the end of an input line |
if (stesc[1] == '#') { |
* requests line continuation. |
*stesc = '\0'; |
*/ |
|
if (stesc[1] == '\0') { |
|
stesc[0] = '\0'; |
return ROFF_IGN | ROFF_APPEND; |
return ROFF_IGN | ROFF_APPEND; |
} |
} |
|
|
/* Discard normal comments. */ |
/* Found a comment: process it. */ |
|
if (stesc[1] == '"' || stesc[1] == '#') |
|
break; |
|
|
while (stesc > start && stesc[-1] == ' ' && |
/* Escaped escape character: skip them both. */ |
(stesc == start + 1 || stesc[-2] != '\\')) |
if (stesc[1] == ec) |
stesc--; |
stesc++; |
*stesc = '\0'; |
|
break; |
|
} |
} |
if (stesc == start) |
|
return ROFF_CONT; |
|
stesc--; |
|
|
|
/* Notice the end of the input. */ |
/* Look for an RCS id in the comment. */ |
|
|
if (*stesc == '\n') { |
rcsid = 0; |
*stesc-- = '\0'; |
if ((cp = strstr(stesc + 2, "$" "OpenBSD")) != NULL) { |
done = 1; |
rcsid = 1 << MANDOC_OS_OPENBSD; |
|
cp += 8; |
|
} else if ((cp = strstr(stesc + 2, "$" "NetBSD")) != NULL) { |
|
rcsid = 1 << MANDOC_OS_NETBSD; |
|
cp += 7; |
} |
} |
|
if (cp != NULL && isalnum((unsigned char)*cp) == 0 && |
|
strchr(cp, '$') != NULL) { |
|
if (r->man->meta.rcsids & rcsid) |
|
mandoc_msg(MANDOCERR_RCS_REP, ln, |
|
(int)(stesc - buf->buf) + 2, "%s", stesc + 1); |
|
r->man->meta.rcsids |= rcsid; |
|
} |
|
|
expand_count = 0; |
/* Warn about trailing whitespace at the end of the comment. */ |
while (stesc >= start) { |
|
if (*stesc != newesc) { |
|
|
|
/* |
ep = strchr(stesc + 2, '\0') - 1; |
* If we have a non-standard escape character, |
if (*ep == '\n') |
* escape literal backslashes because all |
*ep-- = '\0'; |
* processing in subsequent functions uses |
if (*ep == ' ' || *ep == '\t') |
* the standard escaping rules. |
mandoc_msg(MANDOCERR_SPACE_EOL, |
*/ |
ln, (int)(ep - buf->buf), NULL); |
|
|
if (newesc != ASCII_ESC && *stesc == '\\') { |
/* Save comments preceding the title macro in the syntax tree. */ |
*stesc = '\0'; |
|
buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s", |
|
buf->buf, stesc + 1) + 1; |
|
start = nbuf + pos; |
|
stesc = nbuf + (stesc - buf->buf); |
|
free(buf->buf); |
|
buf->buf = nbuf; |
|
} |
|
|
|
/* Search backwards for the next escape. */ |
if (r->options & MPARSE_COMMENT) { |
|
while (*ep == ' ' || *ep == '\t') |
|
ep--; |
|
ep[1] = '\0'; |
|
n = roff_node_alloc(r->man, ln, stesc + 1 - buf->buf, |
|
ROFFT_COMMENT, TOKEN_NONE); |
|
n->string = mandoc_strdup(stesc + 2); |
|
roff_node_append(r->man, n); |
|
n->flags |= NODE_VALID | NODE_ENDED; |
|
r->man->next = ROFF_NEXT_SIBLING; |
|
} |
|
|
stesc--; |
/* The comment requests line continuation. */ |
continue; |
|
} |
|
|
|
/* If it is escaped, skip it. */ |
if (stesc[1] == '#') { |
|
*stesc = '\0'; |
|
return ROFF_IGN | ROFF_APPEND; |
|
} |
|
|
for (cp = stesc - 1; cp >= start; cp--) |
/* Discard the comment including preceding whitespace. */ |
if (*cp != r->escape) |
|
break; |
|
|
|
if ((stesc - cp) % 2 == 0) { |
while (stesc > start && stesc[-1] == ' ' && |
while (stesc > cp) |
(stesc == start + 1 || stesc[-2] != '\\')) |
*stesc-- = '\\'; |
stesc--; |
continue; |
*stesc = '\0'; |
} else if (stesc[1] != '\0') { |
return ROFF_CONT; |
*stesc = '\\'; |
} |
} else { |
|
*stesc-- = '\0'; |
|
if (done) |
|
continue; |
|
else |
|
return ROFF_IGN | ROFF_APPEND; |
|
} |
|
|
|
/* Decide whether to expand or to check only. */ |
/* |
|
* In the current line, expand escape sequences that produce parsable |
|
* input text. Also check the syntax of the remaining escape sequences, |
|
* which typically produce output glyphs or change formatter state. |
|
*/ |
|
static int |
|
roff_expand(struct roff *r, struct buf *buf, int ln, int pos, char ec) |
|
{ |
|
char ubuf[24]; /* buffer to print a number */ |
|
struct mctx *ctx; /* current macro call context */ |
|
const char *res; /* the string to be pasted */ |
|
const char *src; /* source for copying */ |
|
char *dst; /* destination for copying */ |
|
int iesc; /* index of leading escape char */ |
|
int inam; /* index of the escape name */ |
|
int iarg; /* index beginning the argument */ |
|
int iendarg; /* index right after the argument */ |
|
int iend; /* index right after the sequence */ |
|
int isrc, idst; /* to reduce \\ and \. in names */ |
|
int deftype; /* type of definition to paste */ |
|
int argi; /* macro argument index */ |
|
int quote_args; /* true for \\$@, false for \\$* */ |
|
int asz; /* length of the replacement */ |
|
int rsz; /* length of the rest of the string */ |
|
int npos; /* position in numeric expression */ |
|
int expand_count; /* to avoid infinite loops */ |
|
|
term = '\0'; |
expand_count = 0; |
cp = stesc + 1; |
while (buf->buf[pos] != '\0') { |
while (*cp == 'E') |
|
cp++; |
/* |
esct = cp; |
* Skip plain ASCII characters. |
switch (*esct) { |
* If we have a non-standard escape character, |
case '*': |
* escape literal backslashes because all processing in |
case '$': |
* subsequent functions uses the standard escaping rules. |
res = NULL; |
*/ |
break; |
|
case 'B': |
if (buf->buf[pos] != ec) { |
case 'w': |
if (ec != ASCII_ESC && buf->buf[pos] == '\\') { |
term = cp[1]; |
roff_expand_patch(buf, pos, "\\e", pos + 1); |
/* FALLTHROUGH */ |
pos++; |
case 'n': |
|
sign = cp[1]; |
|
if (sign == '+' || sign == '-') |
|
cp++; |
|
res = ubuf; |
|
break; |
|
default: |
|
err = MANDOCERR_OK; |
|
switch(mandoc_escape(&cp, &stnam, &inaml)) { |
|
case ESCAPE_SPECIAL: |
|
if (mchars_spec2cp(stnam, inaml) >= 0) |
|
break; |
|
/* FALLTHROUGH */ |
|
case ESCAPE_ERROR: |
|
err = MANDOCERR_ESC_BAD; |
|
break; |
|
case ESCAPE_UNDEF: |
|
err = MANDOCERR_ESC_UNDEF; |
|
break; |
|
case ESCAPE_UNSUPP: |
|
err = MANDOCERR_ESC_UNSUPP; |
|
break; |
|
default: |
|
break; |
|
} |
} |
if (err != MANDOCERR_OK) |
pos++; |
mandoc_msg(err, ln, (int)(stesc - buf->buf), |
|
"%.*s", (int)(cp - stesc), stesc); |
|
stesc--; |
|
continue; |
continue; |
} |
} |
|
|
if (EXPAND_LIMIT < ++expand_count) { |
|
mandoc_msg(MANDOCERR_ROFFLOOP, |
|
ln, (int)(stesc - buf->buf), NULL); |
|
return ROFF_IGN; |
|
} |
|
|
|
/* |
/* |
* The third character decides the length |
* Parse escape sequences, |
* of the name of the string or register. |
* issue diagnostic messages when appropriate, |
* Save a pointer to the name. |
* and skip sequences that do not need expansion. |
|
* If we have a non-standard escape character, translate |
|
* it to backslashes and translate backslashes to \e. |
*/ |
*/ |
|
|
if (term == '\0') { |
if (roff_escape(buf->buf, ln, pos, &iesc, &inam, |
switch (*++cp) { |
&iarg, &iendarg, &iend) != ESCAPE_EXPAND) { |
case '\0': |
while (pos < iend) { |
maxl = 0; |
if (buf->buf[pos] == ec) { |
break; |
buf->buf[pos] = '\\'; |
case '(': |
if (pos + 1 < iend) |
cp++; |
pos++; |
maxl = 2; |
} else if (buf->buf[pos] == '\\') { |
break; |
roff_expand_patch(buf, |
case '[': |
pos, "\\e", pos + 1); |
cp++; |
pos++; |
term = ']'; |
iend++; |
maxl = 0; |
} |
break; |
pos++; |
default: |
|
maxl = 1; |
|
break; |
|
} |
} |
} else { |
continue; |
cp += 2; |
|
maxl = 0; |
|
} |
} |
stnam = cp; |
|
|
|
/* Advance to the end of the name. */ |
/* Reduce \\ and \. in names. */ |
|
|
naml = 0; |
if (buf->buf[inam] == '*' || buf->buf[inam] == 'n') { |
arg_complete = 1; |
isrc = idst = iarg; |
while (maxl == 0 || naml < maxl) { |
while (isrc < iendarg) { |
if (*cp == '\0') { |
if (isrc + 1 < iendarg && |
mandoc_msg(MANDOCERR_ESC_BAD, ln, |
buf->buf[isrc] == '\\' && |
(int)(stesc - buf->buf), "%s", stesc); |
(buf->buf[isrc + 1] == '\\' || |
arg_complete = 0; |
buf->buf[isrc + 1] == '.')) |
break; |
isrc++; |
|
buf->buf[idst++] = buf->buf[isrc++]; |
} |
} |
if (maxl == 0 && *cp == term) { |
iendarg -= isrc - idst; |
cp++; |
|
break; |
|
} |
|
if (*cp++ != '\\' || *esct != 'w') { |
|
naml++; |
|
continue; |
|
} |
|
switch (mandoc_escape(&cp, NULL, NULL)) { |
|
case ESCAPE_SPECIAL: |
|
case ESCAPE_UNICODE: |
|
case ESCAPE_NUMBERED: |
|
case ESCAPE_UNDEF: |
|
case ESCAPE_OVERSTRIKE: |
|
naml++; |
|
break; |
|
default: |
|
break; |
|
} |
|
} |
} |
|
|
/* |
/* Handle expansion. */ |
* Retrieve the replacement string; if it is |
|
* undefined, resume searching for escapes. |
|
*/ |
|
|
|
switch (*esct) { |
res = NULL; |
|
switch (buf->buf[inam]) { |
case '*': |
case '*': |
if (arg_complete) { |
if (iendarg == iarg) |
deftype = ROFFDEF_USER | ROFFDEF_PRE; |
break; |
res = roff_getstrn(r, stnam, naml, &deftype); |
deftype = ROFFDEF_USER | ROFFDEF_PRE; |
|
if ((res = roff_getstrn(r, buf->buf + iarg, |
|
iendarg - iarg, &deftype)) != NULL) |
|
break; |
|
|
/* |
/* |
* If not overriden, let \*(.T |
* If not overriden, |
* through to the formatters. |
* let \*(.T through to the formatters. |
*/ |
*/ |
|
|
if (res == NULL && naml == 2 && |
if (iendarg - iarg == 2 && |
stnam[0] == '.' && stnam[1] == 'T') { |
buf->buf[iarg] == '.' && |
roff_setstrn(&r->strtab, |
buf->buf[iarg + 1] == 'T') { |
".T", 2, NULL, 0, 0); |
roff_setstrn(&r->strtab, ".T", 2, NULL, 0, 0); |
stesc--; |
pos = iend; |
continue; |
continue; |
} |
|
} |
} |
|
|
|
mandoc_msg(MANDOCERR_STR_UNDEF, ln, iesc, |
|
"%.*s", iendarg - iarg, buf->buf + iarg); |
break; |
break; |
|
|
case '$': |
case '$': |
if (r->mstackpos < 0) { |
if (r->mstackpos < 0) { |
mandoc_msg(MANDOCERR_ARG_UNDEF, ln, |
mandoc_msg(MANDOCERR_ARG_UNDEF, ln, iesc, |
(int)(stesc - buf->buf), "%.3s", stesc); |
"%.*s", iend - iesc, buf->buf + iesc); |
break; |
break; |
} |
} |
ctx = r->mstack + r->mstackpos; |
ctx = r->mstack + r->mstackpos; |
npos = esct[1] - '1'; |
argi = buf->buf[iarg] - '1'; |
if (npos >= 0 && npos <= 8) { |
if (argi >= 0 && argi <= 8) { |
res = npos < ctx->argc ? |
if (argi < ctx->argc) |
ctx->argv[npos] : ""; |
res = ctx->argv[argi]; |
break; |
break; |
} |
} |
if (esct[1] == '*') |
if (buf->buf[iarg] == '*') |
quote_args = 0; |
quote_args = 0; |
else if (esct[1] == '@') |
else if (buf->buf[iarg] == '@') |
quote_args = 1; |
quote_args = 1; |
else { |
else { |
mandoc_msg(MANDOCERR_ARG_NONUM, ln, |
mandoc_msg(MANDOCERR_ARG_NONUM, ln, iesc, |
(int)(stesc - buf->buf), "%.3s", stesc); |
"%.*s", iend - iesc, buf->buf + iesc); |
break; |
break; |
} |
} |
asz = 0; |
asz = 0; |
for (npos = 0; npos < ctx->argc; npos++) { |
for (argi = 0; argi < ctx->argc; argi++) { |
if (npos) |
if (argi) |
asz++; /* blank */ |
asz++; /* blank */ |
if (quote_args) |
if (quote_args) |
asz += 2; /* quotes */ |
asz += 2; /* quotes */ |
asz += strlen(ctx->argv[npos]); |
asz += strlen(ctx->argv[argi]); |
} |
} |
if (asz != 3) { |
if (asz != iend - iesc) { |
rsz = buf->sz - (stesc - buf->buf) - 3; |
rsz = buf->sz - iend; |
if (asz < 3) |
if (asz < iend - iesc) |
memmove(stesc + asz, stesc + 3, rsz); |
memmove(buf->buf + iesc + asz, |
buf->sz += asz - 3; |
buf->buf + iend, rsz); |
nbuf = mandoc_realloc(buf->buf, buf->sz); |
buf->sz = iesc + asz + rsz; |
start = nbuf + pos; |
buf->buf = mandoc_realloc(buf->buf, buf->sz); |
stesc = nbuf + (stesc - buf->buf); |
if (asz > iend - iesc) |
buf->buf = nbuf; |
memmove(buf->buf + iesc + asz, |
if (asz > 3) |
buf->buf + iend, rsz); |
memmove(stesc + asz, stesc + 3, rsz); |
|
} |
} |
for (npos = 0; npos < ctx->argc; npos++) { |
dst = buf->buf + iesc; |
if (npos) |
for (argi = 0; argi < ctx->argc; argi++) { |
*stesc++ = ' '; |
if (argi) |
|
*dst++ = ' '; |
if (quote_args) |
if (quote_args) |
*stesc++ = '"'; |
*dst++ = '"'; |
cp = ctx->argv[npos]; |
src = ctx->argv[argi]; |
while (*cp != '\0') |
while (*src != '\0') |
*stesc++ = *cp++; |
*dst++ = *src++; |
if (quote_args) |
if (quote_args) |
*stesc++ = '"'; |
*dst++ = '"'; |
} |
} |
continue; |
continue; |
|
case 'A': |
|
ubuf[0] = iendarg > iarg ? '1' : '0'; |
|
ubuf[1] = '\0'; |
|
res = ubuf; |
|
break; |
case 'B': |
case 'B': |
npos = 0; |
npos = 0; |
ubuf[0] = arg_complete && |
ubuf[0] = iendarg > iarg && iend > iendarg && |
roff_evalnum(r, ln, stnam, &npos, |
roff_evalnum(r, ln, buf->buf + iarg, &npos, |
NULL, ROFFNUM_SCALE) && |
NULL, ROFFNUM_SCALE) && |
stnam + npos + 1 == cp ? '1' : '0'; |
npos == iendarg - iarg ? '1' : '0'; |
ubuf[1] = '\0'; |
ubuf[1] = '\0'; |
|
res = ubuf; |
break; |
break; |
|
case 'V': |
|
mandoc_msg(MANDOCERR_UNSUPP, ln, iesc, |
|
"%.*s", iend - iesc, buf->buf + iesc); |
|
roff_expand_patch(buf, iendarg, "}", iend); |
|
roff_expand_patch(buf, iesc, "${", iarg); |
|
continue; |
|
case 'g': |
|
break; |
case 'n': |
case 'n': |
if (arg_complete) |
if (iendarg > iarg) |
(void)snprintf(ubuf, sizeof(ubuf), "%d", |
(void)snprintf(ubuf, sizeof(ubuf), "%d", |
roff_getregn(r, stnam, naml, sign)); |
roff_getregn(r, buf->buf + iarg, |
|
iendarg - iarg, buf->buf[inam + 1])); |
else |
else |
ubuf[0] = '\0'; |
ubuf[0] = '\0'; |
|
res = ubuf; |
break; |
break; |
case 'w': |
case 'w': |
/* use even incomplete args */ |
(void)snprintf(ubuf, sizeof(ubuf), |
(void)snprintf(ubuf, sizeof(ubuf), "%d", |
"%d", (iendarg - iarg) * 24); |
24 * (int)naml); |
res = ubuf; |
break; |
break; |
|
default: |
|
break; |
} |
} |
|
if (res == NULL) |
if (res == NULL) { |
|
if (*esct == '*') |
|
mandoc_msg(MANDOCERR_STR_UNDEF, |
|
ln, (int)(stesc - buf->buf), |
|
"%.*s", (int)naml, stnam); |
|
res = ""; |
res = ""; |
} else if (buf->sz + strlen(res) > SHRT_MAX) { |
if (++expand_count > EXPAND_LIMIT || |
mandoc_msg(MANDOCERR_ROFFLOOP, |
buf->sz + strlen(res) > SHRT_MAX) { |
ln, (int)(stesc - buf->buf), NULL); |
mandoc_msg(MANDOCERR_ROFFLOOP, ln, iesc, NULL); |
return ROFF_IGN; |
return ROFF_IGN; |
} |
} |
|
roff_expand_patch(buf, iesc, res, iend); |
/* Replace the escape sequence by the string. */ |
|
|
|
*stesc = '\0'; |
|
buf->sz = mandoc_asprintf(&nbuf, "%s%s%s", |
|
buf->buf, res, cp) + 1; |
|
|
|
/* Prepare for the next replacement. */ |
|
|
|
start = nbuf + pos; |
|
stesc = nbuf + (stesc - buf->buf) + strlen(res); |
|
free(buf->buf); |
|
buf->buf = nbuf; |
|
} |
} |
return ROFF_CONT; |
return ROFF_CONT; |
} |
} |
|
|
/* |
/* |
|
* Replace the substring from the start position (inclusive) |
|
* to end position (exclusive) with the repl(acement) string. |
|
*/ |
|
static void |
|
roff_expand_patch(struct buf *buf, int start, const char *repl, int end) |
|
{ |
|
char *nbuf; |
|
|
|
buf->sz = mandoc_asprintf(&nbuf, "%.*s%s%s", start, buf->buf, |
|
repl, buf->buf + end) + 1; |
|
free(buf->buf); |
|
buf->buf = nbuf; |
|
} |
|
|
|
/* |
* Parse a quoted or unquoted roff-style request or macro argument. |
* Parse a quoted or unquoted roff-style request or macro argument. |
* Return a pointer to the parsed argument, which is either the original |
* Return a pointer to the parsed argument, which is either the original |
* pointer or advanced by one byte in case the argument is quoted. |
* pointer or advanced by one byte in case the argument is quoted. |
Line 1856 roff_parseln(struct roff *r, int ln, struct buf *buf, |
|
Line 1805 roff_parseln(struct roff *r, int ln, struct buf *buf, |
|
assert(e == ROFF_CONT); |
assert(e == ROFF_CONT); |
} |
} |
|
|
/* Expand some escape sequences. */ |
/* Handle comments and escape sequences. */ |
|
|
|
e = roff_parse_comment(r, buf, ln, pos, r->escape); |
|
if ((e & ROFF_MASK) == ROFF_IGN) |
|
return e; |
|
assert(e == ROFF_CONT); |
|
|
e = roff_expand(r, buf, ln, pos, r->escape); |
e = roff_expand(r, buf, ln, pos, r->escape); |
if ((e & ROFF_MASK) == ROFF_IGN) |
if ((e & ROFF_MASK) == ROFF_IGN) |
return e; |
return e; |
Line 3786 roff_tr(ROFF_ARGS) |
|
Line 3740 roff_tr(ROFF_ARGS) |
|
{ |
{ |
const char *p, *first, *second; |
const char *p, *first, *second; |
size_t fsz, ssz; |
size_t fsz, ssz; |
enum mandoc_esc esc; |
|
|
|
p = buf->buf + pos; |
p = buf->buf + pos; |
|
|
Line 3800 roff_tr(ROFF_ARGS) |
|
Line 3753 roff_tr(ROFF_ARGS) |
|
|
|
first = p++; |
first = p++; |
if (*first == '\\') { |
if (*first == '\\') { |
esc = mandoc_escape(&p, NULL, NULL); |
if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR) |
if (esc == ESCAPE_ERROR) { |
|
mandoc_msg(MANDOCERR_ESC_BAD, ln, |
|
(int)(p - buf->buf), "%s", first); |
|
return ROFF_IGN; |
return ROFF_IGN; |
} |
|
fsz = (size_t)(p - first); |
fsz = (size_t)(p - first); |
} |
} |
|
|
second = p++; |
second = p++; |
if (*second == '\\') { |
if (*second == '\\') { |
esc = mandoc_escape(&p, NULL, NULL); |
if (mandoc_escape(&p, NULL, NULL) == ESCAPE_ERROR) |
if (esc == ESCAPE_ERROR) { |
|
mandoc_msg(MANDOCERR_ESC_BAD, ln, |
|
(int)(p - buf->buf), "%s", second); |
|
return ROFF_IGN; |
return ROFF_IGN; |
} |
|
ssz = (size_t)(p - second); |
ssz = (size_t)(p - second); |
} else if (*second == '\0') { |
} else if (*second == '\0') { |
mandoc_msg(MANDOCERR_TR_ODD, ln, |
mandoc_msg(MANDOCERR_TR_ODD, ln, |
|
|
roff_getname(struct roff *r, char **cpp, int ln, int pos) |
roff_getname(struct roff *r, char **cpp, int ln, int pos) |
{ |
{ |
char *name, *cp; |
char *name, *cp; |
size_t namesz; |
int namesz, inam, iend; |
|
|
name = *cpp; |
name = *cpp; |
if (*name == '\0') |
if (*name == '\0') |
Line 4072 roff_getname(struct roff *r, char **cpp, int ln, int p |
|
Line 4017 roff_getname(struct roff *r, char **cpp, int ln, int p |
|
|
|
/* Advance cp to the byte after the end of the name. */ |
/* Advance cp to the byte after the end of the name. */ |
|
|
for (cp = name; 1; cp++) { |
cp = name; |
namesz = cp - name; |
namesz = 0; |
|
for (;;) { |
if (*cp == '\0') |
if (*cp == '\0') |
break; |
break; |
if (*cp == ' ' || *cp == '\t') { |
if (*cp == ' ' || *cp == '\t') { |
cp++; |
cp++; |
break; |
break; |
} |
} |
if (*cp != '\\') |
if (*cp != '\\') { |
|
if (name + namesz < cp) { |
|
name[namesz] = *cp; |
|
*cp = ' '; |
|
} |
|
namesz++; |
|
cp++; |
continue; |
continue; |
|
} |
if (cp[1] == '{' || cp[1] == '}') |
if (cp[1] == '{' || cp[1] == '}') |
break; |
break; |
if (*++cp == '\\') |
if (roff_escape(cp, 0, 0, NULL, &inam, |
continue; |
NULL, NULL, &iend) != ESCAPE_UNDEF) { |
mandoc_msg(MANDOCERR_NAMESC, ln, pos, |
mandoc_msg(MANDOCERR_NAMESC, ln, pos, |
"%.*s", (int)(cp - name + 1), name); |
"%.*s%.*s", namesz, name, iend, cp); |
mandoc_escape((const char **)&cp, NULL, NULL); |
cp += iend; |
break; |
break; |
|
} |
|
|
|
/* |
|
* In an identifier, \\, \., \G and so on |
|
* are reduced to \, ., G and so on, |
|
* vaguely similar to copy mode. |
|
*/ |
|
|
|
name[namesz++] = cp[inam]; |
|
while (iend--) { |
|
if (cp >= name + namesz) |
|
*cp = ' '; |
|
cp++; |
|
} |
} |
} |
|
|
/* Read past spaces. */ |
/* Read past spaces. */ |