In the SYNOPSIS, .Nm blocks can get broken if one of their children

[mandoc.git] / read.c

diff --git a/read.c b/read.c
index fd7b5ce1b8bb0a44078828ff04ea6488805a5485..6d831c48b3921f538db45cfc58f6eb19bbae6198 100644 (file)
--- a/read.c
+++ b/read.c
@@ -1,7 +1,7 @@
-/*     $Id: read.c,v 1.141 2015/09/14 15:36:14 schwarze Exp $ */
+/*     $Id: read.c,v 1.160 2017/02/03 18:18:23 schwarze Exp $ */

 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>

- * Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org>
+ * Copyright (c) 2010-2017 Ingo Schwarze <schwarze@openbsd.org>

  * Copyright (c) 2010, 2012 Joerg Sonnenberger <joerg@netbsd.org>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * Copyright (c) 2010, 2012 Joerg Sonnenberger <joerg@netbsd.org>
  *
  * Permission to use, copy, modify, and distribute this software for any


@@ -19,13 +19,14 @@
 #include "config.h"
 
 #include <sys/types.h>
 #include "config.h"
 
 #include <sys/types.h>

-#if HAVE_MMAP

 #include <sys/mman.h>
 #include <sys/stat.h>
 #include <sys/mman.h>
 #include <sys/stat.h>

-#endif

 
 #include <assert.h>
 #include <ctype.h>
 
 #include <assert.h>
 #include <ctype.h>

+#if HAVE_ERR
+#include <err.h>
+#endif

 #include <errno.h>
 #include <fcntl.h>
 #include <stdarg.h>
 #include <errno.h>
 #include <fcntl.h>
 #include <stdarg.h>


@@ -47,9 +48,8 @@
 #define        REPARSE_LIMIT   1000
 
 struct mparse {
 #define        REPARSE_LIMIT   1000
 
 struct mparse {

-       struct roff_man  *man; /* man parser */

        struct roff      *roff; /* roff parser (!NULL) */
        struct roff      *roff; /* roff parser (!NULL) */

-       const struct mchars *mchars; /* character table */
+       struct roff_man  *man; /* man parser */

        char             *sodest; /* filename pointed to by .so */
        const char       *file; /* filename of current input file */
        struct buf       *primary; /* buffer currently being parsed */
        char             *sodest; /* filename pointed to by .so */
        const char       *file; /* filename of current input file */
        struct buf       *primary; /* buffer currently being parsed */


@@ -108,10 +108,11 @@ static    const char * const      mandocerrs[MANDOCERR_MAX] = {
        "no document body",
        "content before first section header",
        "first section is not \"NAME\"",
        "no document body",
        "content before first section header",
        "first section is not \"NAME\"",

-       "NAME section without name",
+       "NAME section without Nm before Nd",

        "NAME section without description",
        "description not at the end of NAME",
        "bad NAME section content",
        "NAME section without description",
        "description not at the end of NAME",
        "bad NAME section content",

+       "missing comma before name",

        "missing description line, using \"\"",
        "sections out of conventional order",
        "duplicate section title",
        "missing description line, using \"\"",
        "sections out of conventional order",
        "duplicate section title",


@@ -141,7 +142,7 @@ static      const char * const      mandocerrs[MANDOCERR_MAX] = {
        "empty argument, using 0n",
        "missing display type, using -ragged",
        "list type is not the first argument",
        "empty argument, using 0n",
        "missing display type, using -ragged",
        "list type is not the first argument",

-       "missing -width in -tag list, using 8n",
+       "missing -width in -tag list, using 6n",

        "missing utility name, using \"\"",
        "missing function name, using \"\"",
        "empty head in list item",
        "missing utility name, using \"\"",
        "missing function name, using \"\"",
        "empty head in list item",


@@ -150,6 +151,7 @@ static      const char * const      mandocerrs[MANDOCERR_MAX] = {
        "unknown font type, using \\fR",
        "nothing follows prefix",
        "empty reference block",
        "unknown font type, using \\fR",
        "nothing follows prefix",
        "empty reference block",

+       "missing section argument",

        "missing -std argument, adding it",
        "missing option string, using \"\"",
        "missing resource identifier, using \"\"",
        "missing -std argument, adding it",
        "missing option string, using \"\"",
        "missing resource identifier, using \"\"",


@@ -175,6 +177,7 @@ static      const char * const      mandocerrs[MANDOCERR_MAX] = {
        "blank line in fill mode, using .sp",
        "tab in filled text",
        "whitespace at end of input line",
        "blank line in fill mode, using .sp",
        "tab in filled text",
        "whitespace at end of input line",

+       "new sentence, new line",

        "bad comment style",
        "invalid escape sequence",
        "undefined string, using \"\"",
        "bad comment style",
        "invalid escape sequence",
        "undefined string, using \"\"",


@@ -215,6 +218,7 @@ static      const char * const      mandocerrs[MANDOCERR_MAX] = {
        /* related to request and macro arguments */
        "escaped character not allowed in a name",
        "NOT IMPLEMENTED: Bd -file",
        /* related to request and macro arguments */
        "escaped character not allowed in a name",
        "NOT IMPLEMENTED: Bd -file",

+       "skipping display without arguments",

        "missing list type, using -item",
        "missing manual name, using \"\"",
        "uname(3) system call failed, using UNKNOWN",
        "missing list type, using -item",
        "missing manual name, using \"\"",
        "uname(3) system call failed, using UNKNOWN",


@@ -288,13 +292,6 @@ choose_parser(struct mparse *curp)
                }
        }
 
                }
        }
 

-       if (curp->man == NULL) {
-               curp->man = roff_man_alloc(curp->roff, curp, curp->defos,
-                   curp->options & MPARSE_QUICK ? 1 : 0);
-               curp->man->macroset = MACROSET_MAN;
-               curp->man->first->tok = TOKEN_NONE;
-       }
-

        if (format == MPARSE_MDOC) {
                mdoc_hash_init();
                curp->man->macroset = MACROSET_MDOC;
        if (format == MPARSE_MDOC) {
                mdoc_hash_init();
                curp->man->macroset = MACROSET_MDOC;


@@ -321,6 +318,7 @@ mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
        const char      *save_file;
        char            *cp;
        size_t           pos; /* byte number in the ln buffer */
        const char      *save_file;
        char            *cp;
        size_t           pos; /* byte number in the ln buffer */

+       size_t           j;  /* auxiliary byte number in the blk buffer */

        enum rofferr     rr;
        int              of;
        int              lnn; /* line number in the real file */
        enum rofferr     rr;
        int              of;
        int              lnn; /* line number in the real file */


@@ -426,14 +424,21 @@ mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
                        }
 
                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {
                        }
 
                        if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) {

+                               j = i;

                                i += 2;
                                /* Comment, skip to end of line */
                                for (; i < blk.sz; ++i) {
                                i += 2;
                                /* Comment, skip to end of line */
                                for (; i < blk.sz; ++i) {

-                                       if ('\n' == blk.buf[i]) {
-                                               ++i;
-                                               ++lnn;
-                                               break;
-                                       }
+                                       if (blk.buf[i] != '\n')
+                                               continue;
+                                       if (blk.buf[i - 1] == ' ' ||
+                                           blk.buf[i - 1] == '\t')
+                                               mandoc_msg(
+                                                   MANDOCERR_SPACE_EOL,
+                                                   curp, curp->line,
+                                                   pos + i-1 - j, NULL);
+                                       ++i;
+                                       ++lnn;
+                                       break;

                                }
 
                                /* Backout trailing whitespaces */
                                }
 
                                /* Backout trailing whitespaces */


@@ -536,9 +541,9 @@ rerun:
                        if (curp->secondary)
                                curp->secondary->sz -= pos + 1;
                        save_file = curp->file;
                        if (curp->secondary)
                                curp->secondary->sz -= pos + 1;
                        save_file = curp->file;

-                       if (mparse_open(curp, &fd, ln.buf + of) ==
-                           MANDOCLEVEL_OK) {
+                       if ((fd = mparse_open(curp, ln.buf + of)) != -1) {

                                mparse_readfd(curp, fd, ln.buf + of);
                                mparse_readfd(curp, fd, ln.buf + of);

+                               close(fd);

                                curp->file = save_file;
                        } else {
                                curp->file = save_file;
                                curp->file = save_file;
                        } else {
                                curp->file = save_file;


@@ -559,15 +564,7 @@ rerun:
                        break;
                }
 
                        break;
                }
 

-               /*
-                * If input parsers have not been allocated, do so now.
-                * We keep these instanced between parsers, but set them
-                * locally per parse routine since we can use different
-                * parsers with each one.
-                */
-
-               if (curp->man == NULL ||
-                   curp->man->macroset == MACROSET_NONE)
+               if (curp->man->macroset == MACROSET_NONE)

                        choose_parser(curp);
 
                /*
                        choose_parser(curp);
 
                /*


@@ -610,12 +607,10 @@ read_whole_file(struct mparse *curp, const char *file, int fd,
        size_t           off;
        ssize_t          ssz;
 
        size_t           off;
        ssize_t          ssz;
 

-#if HAVE_MMAP

        struct stat      st;
        struct stat      st;

-       if (-1 == fstat(fd, &st)) {
-               perror(file);
-               exit((int)MANDOCLEVEL_SYSERR);
-       }
+
+       if (fstat(fd, &st) == -1)
+               err((int)MANDOCLEVEL_SYSERR, "%s", file);

 
        /*
         * If we're a regular file, try just reading in the whole entry
 
        /*
         * If we're a regular file, try just reading in the whole entry


@@ -627,21 +622,18 @@ read_whole_file(struct mparse *curp, const char *file, int fd,
        if (curp->gzip == 0 && S_ISREG(st.st_mode)) {
                if (st.st_size > 0x7fffffff) {
                        mandoc_msg(MANDOCERR_TOOLARGE, curp, 0, 0, NULL);
        if (curp->gzip == 0 && S_ISREG(st.st_mode)) {
                if (st.st_size > 0x7fffffff) {
                        mandoc_msg(MANDOCERR_TOOLARGE, curp, 0, 0, NULL);

-                       return(0);
+                       return 0;

                }
                *with_mmap = 1;
                fb->sz = (size_t)st.st_size;
                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
                if (fb->buf != MAP_FAILED)
                }
                *with_mmap = 1;
                fb->sz = (size_t)st.st_size;
                fb->buf = mmap(NULL, fb->sz, PROT_READ, MAP_SHARED, fd, 0);
                if (fb->buf != MAP_FAILED)

-                       return(1);
+                       return 1;

        }
        }

-#endif

 
        if (curp->gzip) {
 
        if (curp->gzip) {

-               if ((gz = gzdopen(fd, "rb")) == NULL) {
-                       perror(file);
-                       exit((int)MANDOCLEVEL_SYSERR);
-               }
+               if ((gz = gzdopen(fd, "rb")) == NULL)
+                       err((int)MANDOCLEVEL_SYSERR, "%s", file);

        } else
                gz = NULL;
 
        } else
                gz = NULL;
 


@@ -668,27 +660,21 @@ read_whole_file(struct mparse *curp, const char *file, int fd,
                    read(fd, fb->buf + (int)off, fb->sz - off);
                if (ssz == 0) {
                        fb->sz = off;
                    read(fd, fb->buf + (int)off, fb->sz - off);
                if (ssz == 0) {
                        fb->sz = off;

-                       return(1);
-               }
-               if (ssz == -1) {
-                       perror(file);
-                       exit((int)MANDOCLEVEL_SYSERR);
+                       return 1;

                }
                }

+               if (ssz == -1)
+                       err((int)MANDOCLEVEL_SYSERR, "%s", file);

                off += (size_t)ssz;
        }
 
        free(fb->buf);
        fb->buf = NULL;
                off += (size_t)ssz;
        }
 
        free(fb->buf);
        fb->buf = NULL;

-       return(0);
+       return 0;

 }
 
 static void
 mparse_end(struct mparse *curp)
 {
 }
 
 static void
 mparse_end(struct mparse *curp)
 {

-
-       if (curp->man == NULL && curp->sodest == NULL)
-               curp->man = roff_man_alloc(curp->roff, curp, curp->defos,
-                   curp->options & MPARSE_QUICK ? 1 : 0);

        if (curp->man->macroset == MACROSET_NONE)
                curp->man->macroset = MACROSET_MAN;
        if (curp->man->macroset == MACROSET_MDOC)
        if (curp->man->macroset == MACROSET_NONE)
                curp->man->macroset = MACROSET_MAN;
        if (curp->man->macroset == MACROSET_MDOC)


@@ -748,7 +734,7 @@ mparse_readmem(struct mparse *curp, void *buf, size_t len,
        blk.sz = len;
 
        mparse_parse_buffer(curp, blk, file);
        blk.sz = len;
 
        mparse_parse_buffer(curp, blk, file);

-       return(curp->file_status);
+       return curp->file_status;

 }
 
 /*
 }
 
 /*


@@ -768,24 +754,19 @@ mparse_readfd(struct mparse *curp, int fd, const char *file)
                    (MPARSE_UTF8 | MPARSE_LATIN1);
                mparse_parse_buffer(curp, blk, file);
                curp->filenc = save_filenc;
                    (MPARSE_UTF8 | MPARSE_LATIN1);
                mparse_parse_buffer(curp, blk, file);
                curp->filenc = save_filenc;

-#if HAVE_MMAP

                if (with_mmap)
                        munmap(blk.buf, blk.sz);
                else
                if (with_mmap)
                        munmap(blk.buf, blk.sz);
                else

-#endif

                        free(blk.buf);
        }
                        free(blk.buf);
        }

-
-       if (fd != STDIN_FILENO && close(fd) == -1)
-               perror(file);
-
-       return(curp->file_status);
+       return curp->file_status;

 }
 
 }
 

-enum mandoclevel
-mparse_open(struct mparse *curp, int *fd, const char *file)
+int
+mparse_open(struct mparse *curp, const char *file)

 {
        char             *cp;
 {
        char             *cp;

+       int               fd;

 
        curp->file = file;
        cp = strrchr(file, '.');
 
        curp->file = file;
        cp = strrchr(file, '.');


@@ -793,8 +774,8 @@ mparse_open(struct mparse *curp, int *fd, const char *file)
 
        /* First try to use the filename as it is. */
 
 
        /* First try to use the filename as it is. */
 

-       if ((*fd = open(file, O_RDONLY)) != -1)
-               return(MANDOCLEVEL_OK);
+       if ((fd = open(file, O_RDONLY)) != -1)
+               return fd;

 
        /*
         * If that doesn't work and the filename doesn't
 
        /*
         * If that doesn't work and the filename doesn't


@@ -803,23 +784,23 @@ mparse_open(struct mparse *curp, int *fd, const char *file)
 
        if ( ! curp->gzip) {
                mandoc_asprintf(&cp, "%s.gz", file);
 
        if ( ! curp->gzip) {
                mandoc_asprintf(&cp, "%s.gz", file);

-               *fd = open(file, O_RDONLY);
+               fd = open(cp, O_RDONLY);

                free(cp);
                free(cp);

-               if (*fd != -1) {
+               if (fd != -1) {

                        curp->gzip = 1;
                        curp->gzip = 1;

-                       return(MANDOCLEVEL_OK);
+                       return fd;

                }
        }
 
        /* Neither worked, give up. */
 
        mandoc_msg(MANDOCERR_FILE, curp, 0, 0, strerror(errno));
                }
        }
 
        /* Neither worked, give up. */
 
        mandoc_msg(MANDOCERR_FILE, curp, 0, 0, strerror(errno));

-       return(MANDOCLEVEL_ERROR);
+       return -1;

 }
 
 struct mparse *
 mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
 }
 
 struct mparse *
 mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,

-    const struct mchars *mchars, const char *defos)
+    const char *defos)

 {
        struct mparse   *curp;
 
 {
        struct mparse   *curp;
 


@@ -830,8 +811,7 @@ mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
        curp->mmsg = mmsg;
        curp->defos = defos;
 
        curp->mmsg = mmsg;
        curp->defos = defos;
 

-       curp->mchars = mchars;
-       curp->roff = roff_alloc(curp, curp->mchars, options);
+       curp->roff = roff_alloc(curp, options);

        curp->man = roff_man_alloc( curp->roff, curp, curp->defos,
                curp->options & MPARSE_QUICK ? 1 : 0);
        if (curp->options & MPARSE_MDOC) {
        curp->man = roff_man_alloc( curp->roff, curp, curp->defos,
                curp->options & MPARSE_QUICK ? 1 : 0);
        if (curp->options & MPARSE_MDOC) {


@@ -842,24 +822,23 @@ mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg,
                curp->man->macroset = MACROSET_MAN;
        }
        curp->man->first->tok = TOKEN_NONE;
                curp->man->macroset = MACROSET_MAN;
        }
        curp->man->first->tok = TOKEN_NONE;

-       return(curp);
+       return curp;

 }
 
 void
 mparse_reset(struct mparse *curp)
 {
 }
 
 void
 mparse_reset(struct mparse *curp)
 {

-

        roff_reset(curp->roff);
        roff_reset(curp->roff);

+       roff_man_reset(curp->man);
+
+       free(curp->sodest);
+       curp->sodest = NULL;

 
 

-       if (curp->man != NULL)
-               roff_man_reset(curp->man);

        if (curp->secondary)
                curp->secondary->sz = 0;
 
        curp->file_status = MANDOCLEVEL_OK;
        if (curp->secondary)
                curp->secondary->sz = 0;
 
        curp->file_status = MANDOCLEVEL_OK;

-
-       free(curp->sodest);
-       curp->sodest = NULL;
+       curp->gzip = 0;

 }
 
 void
 }
 
 void


@@ -867,8 +846,7 @@ mparse_free(struct mparse *curp)
 {
 
        roff_man_free(curp->man);
 {
 
        roff_man_free(curp->man);

-       if (curp->roff)
-               roff_free(curp->roff);
+       roff_free(curp->roff);

        if (curp->secondary)
                free(curp->secondary->buf);
 
        if (curp->secondary)
                free(curp->secondary->buf);
 


@@ -890,6 +868,13 @@ mparse_result(struct mparse *curp, struct roff_man **man,
                *man = curp->man;
 }
 
                *man = curp->man;
 }
 

+void
+mparse_updaterc(struct mparse *curp, enum mandoclevel *rc)
+{
+       if (curp->file_status > *rc)
+               *rc = curp->file_status;
+}
+

 void
 mandoc_vmsg(enum mandocerr t, struct mparse *m,
                int ln, int pos, const char *fmt, ...)
 void
 mandoc_vmsg(enum mandocerr t, struct mparse *m,
                int ln, int pos, const char *fmt, ...)


@@ -928,13 +913,13 @@ const char *
 mparse_strerror(enum mandocerr er)
 {
 
 mparse_strerror(enum mandocerr er)
 {
 

-       return(mandocerrs[er]);
+       return mandocerrs[er];

 }
 
 const char *
 mparse_strlevel(enum mandoclevel lvl)
 {
 }
 
 const char *
 mparse_strlevel(enum mandoclevel lvl)
 {

-       return(mandoclevels[lvl]);
+       return mandoclevels[lvl];

 }
 
 void
 }
 
 void


@@ -950,5 +935,5 @@ mparse_getkeep(const struct mparse *p)
 {
 
        assert(p->secondary);
 {
 
        assert(p->secondary);

-       return(p->secondary->sz ? p->secondary->buf : NULL);
+       return p->secondary->sz ? p->secondary->buf : NULL;

 }
 }