The CommonMark specification allows list markers fo the form "number) "

[mandoc.git] / read.c

diff --git a/read.c b/read.c
index 2a7eeb6e113fb4afa5daca8fffabf069fd78257a..883f5412750893b9e4525ffb25ec57068d3be809 100644 (file)
--- a/read.c
+++ b/read.c
@@ -1,7 +1,7 @@
-/*     $Id: read.c,v 1.140 2015/07/19 06:05:16 schwarze Exp $ */
+/*     $Id: read.c,v 1.163 2017/03/07 20:00:08 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


@@ -19,17 +19,17 @@
 #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>

-#include <stdint.h>

 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>


@@ -47,9 +47,8 @@
 #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 */


@@ -67,7 +66,7 @@ struct        mparse {
 
 static void      choose_parser(struct mparse *);
 static void      resize_buf(struct buf *, size_t);
 
 static void      choose_parser(struct mparse *);
 static void      resize_buf(struct buf *, size_t);

-static void      mparse_buf_r(struct mparse *, struct buf, size_t, int);
+static int       mparse_buf_r(struct mparse *, struct buf, size_t, int);

 static int       read_whole_file(struct mparse *, const char *, int,
                                struct buf *, int *);
 static void      mparse_end(struct mparse *);
 static int       read_whole_file(struct mparse *, const char *, int,
                                struct buf *, int *);
 static void      mparse_end(struct mparse *);


@@ -108,11 +107,13 @@ 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 \"\"",
        "missing description line, using \"\"",

+       "description line outside NAME section",

        "sections out of conventional order",
        "duplicate section title",
        "unexpected section",
        "sections out of conventional order",
        "duplicate section title",
        "unexpected section",


@@ -129,7 +130,6 @@ static      const char * const      mandocerrs[MANDOCERR_MAX] = {
        "blocks badly nested",
        "nested displays are not portable",
        "moving content out of list",
        "blocks badly nested",
        "nested displays are not portable",
        "moving content out of list",

-       ".Vt block has child macro",

        "fill mode already enabled, skipping",
        "fill mode already disabled, skipping",
        "line scope broken",
        "fill mode already enabled, skipping",
        "fill mode already disabled, skipping",
        "line scope broken",


@@ -142,7 +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",


@@ -151,6 +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 \"\"",


@@ -176,6 +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 \"\"",


@@ -216,6 +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",


@@ -289,13 +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;


@@ -314,7 +310,7 @@ choose_parser(struct mparse *curp)
  * macros, inline equations, and input line traps)
  * and indirectly (for .so file inclusion).
  */
  * macros, inline equations, and input line traps)
  * and indirectly (for .so file inclusion).
  */

-static void
+static int

 mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
 {
        const struct tbl_span   *span;
 mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start)
 {
        const struct tbl_span   *span;


@@ -322,6 +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 */


@@ -427,14 +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 */


@@ -507,13 +511,16 @@ rerun:
 
                switch (rr) {
                case ROFF_REPARSE:
 
                switch (rr) {
                case ROFF_REPARSE:

-                       if (REPARSE_LIMIT >= ++curp->reparse_count)
-                               mparse_buf_r(curp, ln, of, 0);
-                       else
+                       if (++curp->reparse_count > REPARSE_LIMIT)

                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
                                    curp->line, pos, NULL);
                                mandoc_msg(MANDOCERR_ROFFLOOP, curp,
                                    curp->line, pos, NULL);

-                       pos = 0;
-                       continue;
+                       else if (mparse_buf_r(curp, ln, of, 0) == 1 ||
+                           start == 1) {
+                               pos = 0;
+                               continue;
+                       }
+                       free(ln.buf);
+                       return 0;

                case ROFF_APPEND:
                        pos = strlen(ln.buf);
                        continue;
                case ROFF_APPEND:
                        pos = strlen(ln.buf);
                        continue;


@@ -527,7 +534,7 @@ rerun:
                            (i >= blk.sz || blk.buf[i] == '\0')) {
                                curp->sodest = mandoc_strdup(ln.buf + of);
                                free(ln.buf);
                            (i >= blk.sz || blk.buf[i] == '\0')) {
                                curp->sodest = mandoc_strdup(ln.buf + of);
                                free(ln.buf);

-                               return;
+                               return 1;

                        }
                        /*
                         * We remove `so' clauses from our lookaside
                        }
                        /*
                         * We remove `so' clauses from our lookaside


@@ -537,9 +544,9 @@ 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;


@@ -560,15 +567,7 @@ 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);
 
                /*


@@ -601,22 +600,20 @@ rerun:
        }
 
        free(ln.buf);
        }
 
        free(ln.buf);

+       return 1;

 }
 
 static int
 read_whole_file(struct mparse *curp, const char *file, int fd,
                struct buf *fb, int *with_mmap)
 {
 }
 
 static int
 read_whole_file(struct mparse *curp, const char *file, int fd,
                struct buf *fb, int *with_mmap)
 {

+       struct stat      st;

        gzFile           gz;
        size_t           off;
        ssize_t          ssz;
 
        gzFile           gz;
        size_t           off;
        ssize_t          ssz;
 

-#if HAVE_MMAP
-       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


@@ -628,21 +625,18 @@ 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;
 


@@ -669,27 +663,21 @@ 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)


@@ -749,7 +737,7 @@ 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;

 }
 
 /*
 }
 
 /*


@@ -769,24 +757,19 @@ 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, '.');


@@ -794,8 +777,8 @@ 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


@@ -804,23 +787,23 @@ 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;
 


@@ -831,8 +814,7 @@ 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) {


@@ -843,24 +825,23 @@ 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


@@ -868,8 +849,7 @@ 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);
 


@@ -891,6 +871,13 @@ 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, ...)


@@ -929,13 +916,13 @@ 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


@@ -951,5 +938,5 @@ 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;

 }
 }