-/* $Id: main.c,v 1.111 2010/12/01 15:09:01 kristaps Exp $ */
+/* $Id: main.c,v 1.151 2011/03/16 15:28:35 kristaps Exp $ */
/*
- * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
- * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
+ * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
#define MAP_FILE 0
#endif
-#define UNCONST(a) ((void *)(uintptr_t)(const void *)(a))
+#define REPARSE_LIMIT 1000
/* FIXME: Intel's compiler? LLVM? pcc? */
};
struct curparse {
- const char *file; /* Current parse. */
- int fd; /* Current parse. */
- int line; /* Line number in the file. */
- enum mandoclevel wlevel; /* Ignore messages below this. */
- int wstop; /* Stop after a file with a warning. */
+ enum mandoclevel exit_status; /* status of all file parses */
+ const char *file; /* current file-name */
+ enum mandoclevel file_status; /* error status of current parse */
+ int fd; /* current file-descriptor */
+ int line; /* line number in the file */
+ enum mandoclevel wlevel; /* ignore messages below this */
+ int wstop; /* stop after a file with a warning */
enum intt inttype; /* which parser to use */
+ struct man *pman; /* persistent man parser */
+ struct mdoc *pmdoc; /* persistent mdoc parser */
struct man *man; /* man parser */
struct mdoc *mdoc; /* mdoc parser */
struct roff *roff; /* roff parser (!NULL) */
struct regset regs; /* roff registers */
+ int reparse_count; /* finite interpolation stack */
enum outt outtype; /* which output to use */
out_mdoc outmdoc; /* mdoc output ptr */
out_man outman; /* man output ptr */
"generic warning",
- "text should be uppercase",
- "sections out of conventional order",
- "section name repeats",
- "out of order prologue",
- "repeated prologue entry",
- "list type must come first",
- "tab in non-literal context",
- "bad escape sequence",
- "unterminated quoted string",
- "argument requires the width argument",
- "superfluous width argument",
- "bad date argument",
- "bad width argument",
+ /* related to the prologue */
+ "no title in document",
+ "document title should be all caps",
"unknown manual section",
+ "date missing, using today's date",
+ "cannot parse date, using it verbatim",
+ "prologue macros out of order",
+ "duplicate prologue macro",
+ "macro not allowed in prologue",
+ "macro not allowed in body",
+
+ /* related to document structure */
+ ".so is fragile, better use ln(1)",
+ "NAME section must come first",
+ "bad NAME section contents",
+ "manual name not yet set",
+ "sections out of conventional order",
+ "duplicate section name",
"section not in conventional manual section",
- "end of line whitespace",
- "blocks badly nested",
- "generic error",
+ /* related to macros and nesting */
+ "skipping obsolete macro",
+ "skipping paragraph macro",
+ "skipping no-space macro",
+ "blocks badly nested",
+ "child violates parent syntax",
+ "nested displays are not portable",
+ "already in literal mode",
- "NAME section must come first",
+ /* related to missing macro arguments */
+ "skipping empty macro",
+ "argument count wrong",
+ "missing display type",
+ "list type must come first",
+ "tag lists require a width argument",
+ "missing font type",
+ "skipping end of block that is not open",
+
+ /* related to bad macro arguments */
+ "skipping argument",
+ "duplicate argument",
+ "duplicate display type",
+ "duplicate list type",
+ "unknown AT&T UNIX version",
"bad Boolean value",
- "child violates parent syntax",
- "bad AT&T symbol",
- "bad standard",
- "list type repeated",
- "display type repeated",
- "argument repeated",
- "ignoring argument",
- "manual name not yet set",
- "obsolete macro ignored",
- "empty macro ignored",
- "macro not allowed in body",
- "macro not allowed in prologue",
- "bad character",
- "bad NAME section contents",
- "no blank lines",
- "no text in this context",
+ "unknown font",
+ "unknown standard specifier",
+ "bad width argument",
+
+ /* related to plain text */
+ "blank line in non-literal context",
+ "tab in non-literal context",
+ "end of line whitespace",
"bad comment style",
- "unknown macro will be lost",
- "NOT IMPLEMENTED: skipping request",
+ "unknown escape sequence",
+ "unterminated quoted string",
+
+ "generic error",
+
+ /* related to tables */
+ "bad table syntax",
+ "bad table option",
+ "bad table layout",
+ "no table layout cells specified",
+ "no table data cells specified",
+ "ignore data in cell",
+ "data block still open",
+ "ignoring extra data cells",
+
+ "input stack limit exceeded, infinite loop?",
+ "skipping bad character",
+ "escaped character not allowed in a name",
+ "skipping text before the first section header",
+ "skipping unknown macro",
+ "NOT IMPLEMENTED, please use groff: skipping request",
"line scope broken",
"argument count wrong",
- "request scope close w/none open",
- "scope already open",
+ "skipping end of block that is not open",
+ "missing end of block",
"scope open on exit",
+ "uname(3) system call failed",
"macro requires line argument(s)",
"macro requires body argument(s)",
"macro requires argument(s)",
- "no title in document",
"missing list type",
- "missing display type",
- "missing font type",
"line argument(s) will be lost",
"body argument(s) will be lost",
- "paragraph macro ignored",
"generic fatal error",
"column syntax is inconsistent",
- "displays may not be nested",
- "unsupported display type",
- "blocks badly nested",
- "no such block is open",
+ "NOT IMPLEMENTED: .Bd -file",
"line scope broken, syntax violated",
"argument count wrong, violates syntax",
"child violates parent syntax",
"argument count wrong, violates syntax",
+ "NOT IMPLEMENTED: .so with absolute path or \"..\"",
"no document body",
"no document prologue",
- "utsname system call failed",
"static buffer exhausted",
};
static int moptions(enum intt *, char *);
static int mmsg(enum mandocerr, void *,
int, int, const char *);
-static void pset(const char *, int, struct curparse *,
- struct man **, struct mdoc **);
+static void pset(const char *, int, struct curparse *);
static int toptions(struct curparse *, char *);
static void usage(void) __attribute__((noreturn));
static void version(void) __attribute__((noreturn));
static int woptions(struct curparse *, char *);
static const char *progname;
-static enum mandoclevel exit_status = MANDOCLEVEL_OK;
int
main(int argc, char *argv[])
curp.inttype = INTT_AUTO;
curp.outtype = OUTT_ASCII;
curp.wlevel = MANDOCLEVEL_FATAL;
+ curp.exit_status = MANDOCLEVEL_OK;
/* LINTED */
while (-1 != (c = getopt(argc, argv, "m:O:T:VW:")))
while (*argv) {
ffile(*argv, &curp);
- if (MANDOCLEVEL_OK != exit_status && curp.wstop)
+ if (MANDOCLEVEL_OK != curp.exit_status && curp.wstop)
break;
++argv;
}
if (curp.outfree)
(*curp.outfree)(curp.outdata);
- if (curp.mdoc)
- mdoc_free(curp.mdoc);
- if (curp.man)
- man_free(curp.man);
+ if (curp.pmdoc)
+ mdoc_free(curp.pmdoc);
+ if (curp.pman)
+ man_free(curp.pman);
if (curp.roff)
roff_free(curp.roff);
- return((int)exit_status);
+ return((int)curp.exit_status);
}
usage(void)
{
- (void)fprintf(stderr, "usage: %s [-V] [-foption] "
- "[-mformat] [-Ooption] [-Toutput] "
- "[-Werr] [file...]\n", progname);
+ (void)fprintf(stderr, "usage: %s "
+ "[-V] "
+ "[-foption] "
+ "[-mformat] "
+ "[-Ooption] "
+ "[-Toutput] "
+ "[-Werr] "
+ "[file...]\n",
+ progname);
+
exit((int)MANDOCLEVEL_BADARG);
}
-
static void
ffile(const char *file, struct curparse *curp)
{
+ /*
+ * Called once per input file. Get the file ready for reading,
+ * pass it through to the parser-driver, then close it out.
+ * XXX: don't do anything special as this is only called for
+ * files; stdin goes directly to fdesc().
+ */
+
curp->file = file;
+
if (-1 == (curp->fd = open(curp->file, O_RDONLY, 0))) {
perror(curp->file);
- exit_status = MANDOCLEVEL_SYSERR;
+ curp->exit_status = MANDOCLEVEL_SYSERR;
return;
}
if (-1 == (fd = open(file, O_RDONLY, 0))) {
perror(file);
- exit_status = MANDOCLEVEL_SYSERR;
+ curp->file_status = MANDOCLEVEL_SYSERR;
return(0);
}
if (-1 == close(fd))
perror(file);
- return(MANDOCLEVEL_FATAL > exit_status ? 1 : 0);
+ return(MANDOCLEVEL_FATAL > curp->file_status ? 1 : 0);
}
resize_buf(struct buf *buf, size_t initial)
{
- buf->sz = buf->sz ? 2 * buf->sz : initial;
+ buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial;
buf->buf = realloc(buf->buf, buf->sz);
if (NULL == buf->buf) {
perror(NULL);
static void
fdesc(struct curparse *curp)
{
- struct man *man;
- struct mdoc *mdoc;
- struct roff *roff;
- pdesc(curp);
+ /*
+ * Called once per file with an opened file descriptor. All
+ * pre-file-parse operations (whether stdin or a file) should go
+ * here.
+ *
+ * This calls down into the nested parser, which drills down and
+ * fully parses a file and all its dependences (i.e., `so'). It
+ * then runs the cleanup validators and pushes to output.
+ */
+
+ /* Zero the parse type. */
- man = curp->man;
- mdoc = curp->mdoc;
- roff = curp->roff;
+ curp->mdoc = NULL;
+ curp->man = NULL;
+ curp->file_status = MANDOCLEVEL_OK;
+
+ /* Make sure the mandotory roff parser is initialised. */
+
+ if (NULL == curp->roff) {
+ curp->roff = roff_alloc(&curp->regs, curp, mmsg);
+ assert(curp->roff);
+ }
+
+ /* Fully parse the file. */
+
+ pdesc(curp);
- if (MANDOCLEVEL_FATAL <= exit_status)
+ if (MANDOCLEVEL_FATAL <= curp->file_status)
goto cleanup;
/* NOTE a parser may not have been assigned, yet. */
- if ( ! (man || mdoc)) {
+ if ( ! (curp->man || curp->mdoc)) {
fprintf(stderr, "%s: Not a manual\n", curp->file);
- exit_status = MANDOCLEVEL_FATAL;
+ curp->file_status = MANDOCLEVEL_FATAL;
goto cleanup;
}
/* Clean up the parse routine ASTs. */
- if (mdoc && ! mdoc_endparse(mdoc)) {
- assert(MANDOCLEVEL_FATAL <= exit_status);
+ if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) {
+ assert(MANDOCLEVEL_FATAL <= curp->file_status);
goto cleanup;
}
- if (man && ! man_endparse(man)) {
- assert(MANDOCLEVEL_FATAL <= exit_status);
- goto cleanup;
- }
- if (roff && ! roff_endparse(roff)) {
- assert(MANDOCLEVEL_FATAL <= exit_status);
+
+ if (curp->man && ! man_endparse(curp->man)) {
+ assert(MANDOCLEVEL_FATAL <= curp->file_status);
goto cleanup;
}
+ assert(curp->roff);
+ roff_endparse(curp->roff);
+
/*
* With -Wstop and warnings or errors of at least
* the requested level, do not produce output.
*/
- if (MANDOCLEVEL_OK != exit_status && curp->wstop)
+ if (MANDOCLEVEL_OK != curp->file_status && curp->wstop)
goto cleanup;
/* If unset, allocate output dev now (if applicable). */
/* Execute the out device, if it exists. */
- if (man && curp->outman)
- (*curp->outman)(curp->outdata, man);
- if (mdoc && curp->outmdoc)
- (*curp->outmdoc)(curp->outdata, mdoc);
+ if (curp->man && curp->outman)
+ (*curp->outman)(curp->outdata, curp->man);
+ if (curp->mdoc && curp->outmdoc)
+ (*curp->outmdoc)(curp->outdata, curp->mdoc);
cleanup:
+
memset(&curp->regs, 0, sizeof(struct regset));
- if (mdoc)
- mdoc_reset(mdoc);
- if (man)
- man_reset(man);
- if (roff)
- roff_reset(roff);
+
+ /* Reset the current-parse compilers. */
+
+ if (curp->mdoc)
+ mdoc_reset(curp->mdoc);
+ if (curp->man)
+ man_reset(curp->man);
+
+ assert(curp->roff);
+ roff_reset(curp->roff);
+
+ if (curp->exit_status < curp->file_status)
+ curp->exit_status = curp->file_status;
return;
}
-
static void
pdesc(struct curparse *curp)
{
struct buf blk;
int with_mmap;
+ /*
+ * Run for each opened file; may be called more than once for
+ * each full parse sequence if the opened file is nested (i.e.,
+ * from `so'). Simply sucks in the whole file and moves into
+ * the parse phase for the file.
+ */
+
if ( ! read_whole_file(curp, &blk, &with_mmap)) {
- exit_status = MANDOCLEVEL_SYSERR;
+ curp->file_status = MANDOCLEVEL_SYSERR;
return;
}
- if (NULL == curp->roff)
- curp->roff = roff_alloc(&curp->regs, curp, mmsg);
- assert(curp->roff);
+ /* Line number is per-file. */
curp->line = 1;
+
parsebuf(curp, blk, 1);
if (with_mmap)
free(blk.buf);
}
+/*
+ * Main parse routine for an opened file. This is called for each
+ * opened file and simply loops around the full input file, possibly
+ * nesting (i.e., with `so').
+ */
static void
parsebuf(struct curparse *curp, struct buf blk, int start)
{
+ const struct tbl_span *span;
struct buf ln;
- int i, pos, lnn, of;
+ enum rofferr rr;
+ int i, of, rc;
+ int pos; /* byte number in the ln buffer */
+ int lnn; /* line number in the real file */
unsigned char c;
- struct man *man;
- struct mdoc *mdoc;
- struct roff *roff;
-
- man = curp->man;
- mdoc = curp->mdoc;
- roff = curp->roff;
memset(&ln, 0, sizeof(struct buf));
- lnn = curp->line; /* line number in the real file */
- pos = 0; /* byte number in the ln buffer */
+ lnn = curp->line;
+ pos = 0;
- for (i = 0; i < (int)blk.sz;) {
+ for (i = 0; i < (int)blk.sz; ) {
if (0 == pos && '\0' == blk.buf[i])
break;
- if (start)
+
+ if (start) {
curp->line = lnn;
+ curp->reparse_count = 0;
+ }
while (i < (int)blk.sz && (start || '\0' != blk.buf[i])) {
+
+ /*
+ * When finding an unescaped newline character,
+ * leave the character loop to process the line.
+ * Skip a preceding carriage return, if any.
+ */
+
+ if ('\r' == blk.buf[i] && i + 1 < (int)blk.sz &&
+ '\n' == blk.buf[i + 1])
+ ++i;
if ('\n' == blk.buf[i]) {
++i;
++lnn;
*/
c = (unsigned char) blk.buf[i];
- if ( ! (isascii(c) && (isgraph(c) || isblank(c)))) {
+
+ if ( ! (isascii(c) &&
+ (isgraph(c) || isblank(c)))) {
mmsg(MANDOCERR_BADCHAR, curp,
curp->line, pos, "ignoring byte");
i++;
continue;
}
- /* Trailing backslash is like a plain character. */
+ /* Trailing backslash = a plain char. */
+
if ('\\' != blk.buf[i] || i + 1 == (int)blk.sz) {
if (pos >= (int)ln.sz)
resize_buf(&ln, 256);
ln.buf[pos++] = blk.buf[i++];
continue;
}
- /* Found an escape and at least one other character. */
+
+ /*
+ * Found escape and at least one other character.
+ * When it's a newline character, skip it.
+ * When there is a carriage return in between,
+ * skip that one as well.
+ */
+
+ if ('\r' == blk.buf[i + 1] && i + 2 < (int)blk.sz &&
+ '\n' == blk.buf[i + 2])
+ ++i;
if ('\n' == blk.buf[i + 1]) {
- /* Escaped newlines are skipped over */
i += 2;
++lnn;
continue;
}
+
if ('"' == blk.buf[i + 1]) {
i += 2;
/* Comment, skip to end of line */
break;
}
}
+
/* Backout trailing whitespaces */
for (; pos > 0; --pos) {
if (ln.buf[pos - 1] != ' ')
}
break;
}
- /* Some other escape sequence, copy and continue. */
+
+ /* Some other escape sequence, copy & cont. */
+
if (pos + 1 >= (int)ln.sz)
resize_buf(&ln, 256);
if (pos >= (int)ln.sz)
resize_buf(&ln, 256);
+
ln.buf[pos] = '\0';
/*
*/
of = 0;
+
rerun:
- switch (roff_parseln(roff, curp->line, &ln.buf, &ln.sz,
- of, &of)) {
+ rr = roff_parseln
+ (curp->roff, curp->line,
+ &ln.buf, &ln.sz, of, &of);
+
+ switch (rr) {
case (ROFF_REPARSE):
- parsebuf(curp, ln, 0);
+ if (REPARSE_LIMIT >= ++curp->reparse_count)
+ parsebuf(curp, ln, 0);
+ else
+ mmsg(MANDOCERR_ROFFLOOP, curp,
+ curp->line, pos, NULL);
pos = 0;
continue;
case (ROFF_APPEND):
- pos = strlen(ln.buf);
+ pos = (int)strlen(ln.buf);
continue;
case (ROFF_RERUN):
goto rerun;
pos = 0;
continue;
case (ROFF_ERR):
- assert(MANDOCLEVEL_FATAL <= exit_status);
+ assert(MANDOCLEVEL_FATAL <= curp->file_status);
break;
case (ROFF_SO):
if (pfile(ln.buf + of, curp)) {
continue;
} else
break;
- case (ROFF_CONT):
+ default:
break;
}
+ /*
+ * If we encounter errors in the recursive parsebuf()
+ * call, make sure we don't continue parsing.
+ */
+
+ if (MANDOCLEVEL_FATAL <= curp->file_status)
+ break;
+
/*
* If input parsers have not been allocated, do so now.
* We keep these instanced betwen parsers, but set them
* parsers with each one.
*/
- if ( ! (man || mdoc))
- pset(ln.buf + of, pos - of, curp, &man, &mdoc);
+ if ( ! (curp->man || curp->mdoc))
+ pset(ln.buf + of, pos - of, curp);
+
+ /*
+ * Lastly, push down into the parsers themselves. One
+ * of these will have already been set in the pset()
+ * routine.
+ * If libroff returns ROFF_TBL, then add it to the
+ * currently open parse. Since we only get here if
+ * there does exist data (see tbl_data.c), we're
+ * guaranteed that something's been allocated.
+ * Do the same for ROFF_EQN.
+ */
- /* Lastly, push down into the parsers themselves. */
+ rc = -1;
- if (man && ! man_parseln(man, curp->line, ln.buf, of)) {
- assert(MANDOCLEVEL_FATAL <= exit_status);
- break;
- }
- if (mdoc && ! mdoc_parseln(mdoc, curp->line, ln.buf, of)) {
- assert(MANDOCLEVEL_FATAL <= exit_status);
+ if (ROFF_TBL == rr)
+ while (NULL != (span = roff_span(curp->roff))) {
+ rc = curp->man ?
+ man_addspan(curp->man, span) :
+ mdoc_addspan(curp->mdoc, span);
+ if (0 == rc)
+ break;
+ }
+ else if (ROFF_EQN == rr)
+ rc = curp->mdoc ?
+ mdoc_addeqn(curp->mdoc,
+ roff_eqn(curp->roff)) :
+ man_addeqn(curp->man,
+ roff_eqn(curp->roff));
+ else if (curp->man || curp->mdoc)
+ rc = curp->man ?
+ man_parseln(curp->man,
+ curp->line, ln.buf, of) :
+ mdoc_parseln(curp->mdoc,
+ curp->line, ln.buf, of);
+
+ if (0 == rc) {
+ assert(MANDOCLEVEL_FATAL <= curp->file_status);
break;
}
/* Temporary buffers typically are not full. */
+
if (0 == start && '\0' == blk.buf[i])
break;
/* Start the next input line. */
+
pos = 0;
}
free(ln.buf);
}
-
static void
-pset(const char *buf, int pos, struct curparse *curp,
- struct man **man, struct mdoc **mdoc)
+pset(const char *buf, int pos, struct curparse *curp)
{
int i;
* explicitly. If passed as -mandoc, then try to guess from the
* line: either skip dot-lines, use -mdoc when finding `.Dt', or
* default to -man, which is more lenient.
+ *
+ * Separate out pmdoc/pman from mdoc/man: the first persists
+ * through all parsers, while the latter is used per-parse.
*/
if ('.' == buf[0] || '\'' == buf[0]) {
switch (curp->inttype) {
case (INTT_MDOC):
- if (NULL == curp->mdoc)
- curp->mdoc = mdoc_alloc(&curp->regs, curp, mmsg);
- assert(curp->mdoc);
- *mdoc = curp->mdoc;
+ if (NULL == curp->pmdoc)
+ curp->pmdoc = mdoc_alloc
+ (&curp->regs, curp, mmsg);
+ assert(curp->pmdoc);
+ curp->mdoc = curp->pmdoc;
return;
case (INTT_MAN):
- if (NULL == curp->man)
- curp->man = man_alloc(&curp->regs, curp, mmsg);
- assert(curp->man);
- *man = curp->man;
+ if (NULL == curp->pman)
+ curp->pman = man_alloc
+ (&curp->regs, curp, mmsg);
+ assert(curp->pman);
+ curp->man = curp->pman;
return;
default:
break;
}
if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3)) {
- if (NULL == curp->mdoc)
- curp->mdoc = mdoc_alloc(&curp->regs, curp, mmsg);
- assert(curp->mdoc);
- *mdoc = curp->mdoc;
+ if (NULL == curp->pmdoc)
+ curp->pmdoc = mdoc_alloc
+ (&curp->regs, curp, mmsg);
+ assert(curp->pmdoc);
+ curp->mdoc = curp->pmdoc;
return;
}
- if (NULL == curp->man)
- curp->man = man_alloc(&curp->regs, curp, mmsg);
- assert(curp->man);
- *man = curp->man;
+ if (NULL == curp->pman)
+ curp->pman = man_alloc(&curp->regs, curp, mmsg);
+ assert(curp->pman);
+ curp->man = curp->pman;
}
-
static int
moptions(enum intt *tflags, char *arg)
{
return(1);
}
-
static int
toptions(struct curparse *curp, char *arg)
{
return(1);
}
-
static int
woptions(struct curparse *curp, char *arg)
{
return(1);
}
-
static int
mmsg(enum mandocerr t, void *arg, int ln, int col, const char *msg)
{
fprintf(stderr, ": %s", msg);
fputc('\n', stderr);
- if (exit_status < level)
- exit_status = level;
+ if (cp->file_status < level)
+ cp->file_status = level;
return(level < MANDOCLEVEL_FATAL);
}
git.ckatri.com / mandoc.git / blobdiff