Switch to autogenerated dependency rules; they are less error-prone.

[mandoc.git] / read.c

diff --git a/read.c b/read.c
index a99ca0584ccfb1fffaf289b9b82d5d39c46da93c..95c0fe40890e4eb71823d567e6cc4eb8d3869420 100644 (file)
--- a/read.c
+++ b/read.c
@@ -1,4 +1,4 @@
-/*     $Id: read.c,v 1.45 2014/03/19 22:20:43 schwarze Exp $ */
+/*     $Id: read.c,v 1.78 2014/08/01 17:27:44 schwarze Exp $ */

 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>


@@ -37,6 +37,7 @@
 #include <unistd.h>
 
 #include "mandoc.h"
 #include <unistd.h>
 
 #include "mandoc.h"

+#include "mandoc_aux.h"

 #include "libmandoc.h"
 #include "mdoc.h"
 #include "man.h"
 #include "libmandoc.h"
 #include "mdoc.h"
 #include "man.h"


@@ -45,7 +46,7 @@
 #define        REPARSE_LIMIT   1000
 
 struct buf {
 #define        REPARSE_LIMIT   1000
 
 struct buf {

-       char             *buf; /* binary input buffer */
+       char             *buf; /* binary input buffer */

        size_t            sz; /* size of binary buffer */
 };
 
        size_t            sz; /* size of binary buffer */
 };
 


@@ -62,9 +63,9 @@ struct        mparse {
        char             *sodest; /* filename pointed to by .so */
        int               reparse_count; /* finite interp. stack */
        mandocmsg         mmsg; /* warning/error message handler */
        char             *sodest; /* filename pointed to by .so */
        int               reparse_count; /* finite interp. stack */
        mandocmsg         mmsg; /* warning/error message handler */

-       const char       *file; 
+       const char       *file;

        struct buf       *secondary;
        struct buf       *secondary;

-       char             *defos; /* default operating system */
+       const char       *defos; /* default operating system */

 };
 
 static void      resize_buf(struct buf *, size_t);
 };
 
 static void      resize_buf(struct buf *, size_t);


@@ -92,67 +93,76 @@ static      const char * const      mandocerrs[MANDOCERR_MAX] = {
        "generic warning",
 
        /* related to the prologue */
        "generic warning",
 
        /* related to the prologue */

-       "no title in document",
-       "document title should be all caps",
+       "missing .TH macro, using \"unknown 1\"",
+       "lower case character in document title",

        "unknown manual section",
        "unknown manual volume or arch",
        "unknown manual section",
        "unknown manual volume or arch",

-       "date missing, using today's date",
+       "missing date, using today's date",

        "cannot parse date, using it verbatim",
        "prologue macros out of order",
        "duplicate prologue macro",
        "cannot parse date, using it verbatim",
        "prologue macros out of order",
        "duplicate prologue macro",

-       "macro not allowed in prologue",
-       "macro not allowed in body",
+       "incomplete prologue, terminated by",
+       "skipping prologue macro in body",

 
        /* related to document structure */
        ".so is fragile, better use ln(1)",
 
        /* related to document structure */
        ".so is fragile, better use ln(1)",

-       "NAME section must come first",
+       "no document body",
+       "content before first section header",
+       "first section is not \"NAME\"",

        "bad NAME section contents",
        "sections out of conventional order",
        "bad NAME section contents",
        "sections out of conventional order",

-       "duplicate section name",
-       "section header suited to sections 2, 3, and 9 only",
+       "duplicate section title",
+       "unexpected section",

 
        /* related to macros and nesting */
 
        /* related to macros and nesting */

-       "skipping obsolete macro",
+       "obsolete macro",

        "skipping paragraph macro",
        "moving paragraph macro out of list",
        "skipping no-space macro",
        "blocks badly nested",
        "skipping paragraph macro",
        "moving paragraph macro out of list",
        "skipping no-space macro",
        "blocks badly nested",

-       "child violates parent syntax",

        "nested displays are not portable",
        "nested displays are not portable",

-       "already in literal mode",
+       "moving content out of list",
+       ".Vt block has child macro",
+       "fill mode already enabled, skipping",
+       "fill mode already disabled, skipping",

        "line scope broken",
 
        /* related to missing macro arguments */
        "line scope broken",
 
        /* related to missing macro arguments */

+       "skipping empty request",
+       "conditional request controls empty scope",

        "skipping empty macro",
        "skipping empty macro",

+       "empty argument, using 0n",

        "argument count wrong",
        "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",
+       "missing display type, using -ragged",
+       "list type is not the first argument",
+       "missing -width in -tag list, using 8n",
+       "missing utility name, using \"\"",
+       "empty head in list item",
+       "empty list item",
+       "missing font type, using \\fR",
+       "unknown font type, using \\fR",
+       "missing -std argument, adding it",

 
        /* related to bad macro arguments */
 
        /* related to bad macro arguments */

-       "skipping argument",
+       "unterminated quoted argument",

        "duplicate argument",
        "duplicate argument",

-       "duplicate display type",
-       "duplicate list type",
+       "skipping duplicate argument",
+       "skipping duplicate display type",
+       "skipping duplicate list type",
+       "skipping -width argument",

        "unknown AT&T UNIX version",
        "unknown AT&T UNIX version",

-       "bad Boolean value",
-       "unknown font",
-       "unknown standard specifier",
-       "bad width argument",
+       "invalid content in Rs block",
+       "invalid Boolean argument",
+       "unknown font, skipping request",

 
        /* related to plain text */
 
        /* related to plain text */

-       "blank line in non-literal context",
-       "tab in non-literal context",
-       "end of line whitespace",
+       "blank line in fill mode, using .sp",
+       "tab in filled text",
+       "whitespace at end of input line",

        "bad comment style",
        "bad comment style",

-       "bad escape sequence",
-       "unterminated quoted string",
+       "invalid escape sequence",
+       "undefined string, using \"\"",

 
 

-       /* related to equations */
-       "unexpected literal in equation",
-       

        "generic error",
 
        /* related to equations */
        "generic error",
 
        /* related to equations */


@@ -172,43 +182,36 @@ static    const char * const      mandocerrs[MANDOCERR_MAX] = {
        "data block still open",
        "ignoring extra data cells",
 
        "data block still open",
        "ignoring extra data cells",
 

+       /* related to document structure and macros */

        "input stack limit exceeded, infinite loop?",
        "skipping bad character",
        "input stack limit exceeded, infinite loop?",
        "skipping bad character",

-       "escaped character not allowed in a name",
-       "manual name not yet set",
-       "skipping text before the first section header",

        "skipping unknown macro",
        "skipping unknown macro",

-       "NOT IMPLEMENTED, please use groff: skipping request",
-       "argument count wrong",
+       "skipping item outside list",

        "skipping column outside column list",
        "skipping end of block that is not open",
        "skipping column outside column list",
        "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)",
-       "request requires a numeric argument",
-       "missing list type",
-       "line argument(s) will be lost",
-       "body argument(s) will be lost",
+       "inserting missing end of block",
+       "appending missing end of block",
+
+       /* related to request and macro arguments */
+       "escaped character not allowed in a name",
+       "argument count wrong",
+       "missing list type, using -item",
+       "missing manual name, using \"\"",
+       "uname(3) system call failed, using UNKNOWN",
+       "unknown standard specifier",
+       "skipping request without numeric argument",
+       "skipping all arguments",
+       "skipping excess arguments",

 
        "generic fatal error",
 
        "input too large",
 
        "generic fatal error",
 
        "input too large",

-       "not a manual",
-       "column syntax is inconsistent",
-       "NOT IMPLEMENTED: .Bd -file",
-       "argument count wrong, violates syntax",
-       "child violates parent syntax",
-       "argument count wrong, violates syntax",
+       "NOT IMPLEMENTED: Bd -file",

        "NOT IMPLEMENTED: .so with absolute path or \"..\"",
        "NOT IMPLEMENTED: .so with absolute path or \"..\"",

-       "no document body",
-       "no document prologue",
-       "static buffer exhausted",
+       ".so request failed",

 
        /* system errors */
 
        /* system errors */

-       "cannot open file",
+       NULL,

        "cannot stat file",
        "cannot read file",
 };
        "cannot stat file",
        "cannot read file",
 };


@@ -223,6 +226,7 @@ static      const char * const      mandoclevels[MANDOCLEVEL_MAX] = {
        "SYSERR"
 };
 
        "SYSERR"
 };
 

+

 static void
 resize_buf(struct buf *buf, size_t initial)
 {
 static void
 resize_buf(struct buf *buf, size_t initial)
 {


@@ -256,33 +260,24 @@ pset(const char *buf, int pos, struct mparse *curp)
        }
 
        if (MPARSE_MDOC & curp->options) {
        }
 
        if (MPARSE_MDOC & curp->options) {

-               if (NULL == curp->pmdoc) 
-                       curp->pmdoc = mdoc_alloc(
-                           curp->roff, curp, curp->defos,
-                           MPARSE_QUICK & curp->options ? 1 : 0);
-               assert(curp->pmdoc);

                curp->mdoc = curp->pmdoc;
                return;
        } else if (MPARSE_MAN & curp->options) {
                curp->mdoc = curp->pmdoc;
                return;
        } else if (MPARSE_MAN & curp->options) {

-               if (NULL == curp->pman) 
-                       curp->pman = man_alloc(curp->roff, curp,
-                           MPARSE_QUICK & curp->options ? 1 : 0);
-               assert(curp->pman);

                curp->man = curp->pman;
                return;
        }
 
        if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3))  {
                curp->man = curp->pman;
                return;
        }
 
        if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3))  {

-               if (NULL == curp->pmdoc) 
+               if (NULL == curp->pmdoc)

                        curp->pmdoc = mdoc_alloc(
                            curp->roff, curp, curp->defos,
                            MPARSE_QUICK & curp->options ? 1 : 0);
                assert(curp->pmdoc);
                curp->mdoc = curp->pmdoc;
                return;
                        curp->pmdoc = mdoc_alloc(
                            curp->roff, curp, curp->defos,
                            MPARSE_QUICK & curp->options ? 1 : 0);
                assert(curp->pmdoc);
                curp->mdoc = curp->pmdoc;
                return;

-       } 
+       }

 
 

-       if (NULL == curp->pman) 
+       if (NULL == curp->pman)

                curp->pman = man_alloc(curp->roff, curp,
                    MPARSE_QUICK & curp->options ? 1 : 0);
        assert(curp->pman);
                curp->pman = man_alloc(curp->roff, curp,
                    MPARSE_QUICK & curp->options ? 1 : 0);
        assert(curp->pman);


@@ -307,8 +302,8 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
 
        memset(&ln, 0, sizeof(struct buf));
 
 
        memset(&ln, 0, sizeof(struct buf));
 

-       lnn = curp->line; 
-       pos = 0; 
+       lnn = curp->line;
+       pos = 0;

 
        for (i = 0; i < (int)blk.sz; ) {
                if (0 == pos && '\0' == blk.buf[i])
 
        for (i = 0; i < (int)blk.sz; ) {
                if (0 == pos && '\0' == blk.buf[i])


@@ -345,7 +340,7 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
                        if (pos + 2 >= (int)ln.sz)
                                resize_buf(&ln, 256);
 
                        if (pos + 2 >= (int)ln.sz)
                                resize_buf(&ln, 256);
 

-                       /* 
+                       /*

                         * Warn about bogus characters.  If you're using
                         * non-ASCII encoding, you're screwing your
                         * readers.  Since I'd rather this not happen,
                         * Warn about bogus characters.  If you're using
                         * non-ASCII encoding, you're screwing your
                         * readers.  Since I'd rather this not happen,


@@ -356,10 +351,10 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
 
                        c = (unsigned char) blk.buf[i];
 
 
                        c = (unsigned char) blk.buf[i];
 

-                       if ( ! (isascii(c) && 
-                                       (isgraph(c) || isblank(c)))) {
-                               mandoc_msg(MANDOCERR_BADCHAR, curp,
-                                               curp->line, pos, NULL);
+                       if ( ! (isascii(c) &&
+                           (isgraph(c) || isblank(c)))) {
+                               mandoc_vmsg(MANDOCERR_BADCHAR, curp,
+                                   curp->line, pos, "0x%x", c);

                                i++;
                                ln.buf[pos++] = '?';
                                continue;
                                i++;
                                ln.buf[pos++] = '?';
                                continue;


@@ -413,10 +408,10 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
 
                        c = (unsigned char) blk.buf[i+1];
 
 
                        c = (unsigned char) blk.buf[i+1];
 

-                       if ( ! (isascii(c) && 
-                                       (isgraph(c) || isblank(c)))) {
-                               mandoc_msg(MANDOCERR_BADCHAR, curp,
-                                               curp->line, pos, NULL);
+                       if ( ! (isascii(c) &&
+                           (isgraph(c) || isblank(c)))) {
+                               mandoc_vmsg(MANDOCERR_BADCHAR, curp,
+                                   curp->line, pos, "0x%x", c);

                                i += 2;
                                ln.buf[pos++] = '?';
                                continue;
                                i += 2;
                                ln.buf[pos++] = '?';
                                continue;


@@ -428,7 +423,7 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
                        ln.buf[pos++] = blk.buf[i++];
                }
 
                        ln.buf[pos++] = blk.buf[i++];
                }
 

-               if (pos >= (int)ln.sz)
+               if (pos >= (int)ln.sz)

                        resize_buf(&ln, 256);
 
                ln.buf[pos] = '\0';
                        resize_buf(&ln, 256);
 
                ln.buf[pos] = '\0';


@@ -451,13 +446,12 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
                 */
 
                if (curp->secondary) {
                 */
 
                if (curp->secondary) {

-                       curp->secondary->buf = 
-                               mandoc_realloc
-                               (curp->secondary->buf, 
-                                curp->secondary->sz + pos + 2);
-                       memcpy(curp->secondary->buf + 
-                                       curp->secondary->sz, 
-                                       ln.buf, pos);
+                       curp->secondary->buf = mandoc_realloc(
+                           curp->secondary->buf,
+                           curp->secondary->sz + pos + 2);
+                       memcpy(curp->secondary->buf +
+                           curp->secondary->sz,
+                           ln.buf, pos);

                        curp->secondary->sz += pos;
                        curp->secondary->buf
                                [curp->secondary->sz] = '\n';
                        curp->secondary->sz += pos;
                        curp->secondary->buf
                                [curp->secondary->sz] = '\n';


@@ -466,31 +460,30 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int start)
                                [curp->secondary->sz] = '\0';
                }
 rerun:
                                [curp->secondary->sz] = '\0';
                }
 rerun:

-               rr = roff_parseln
-                       (curp->roff, curp->line, 
-                        &ln.buf, &ln.sz, of, &of);
+               rr = roff_parseln(curp->roff, curp->line,
+                   &ln.buf, &ln.sz, of, &of);

 
                switch (rr) {
 
                switch (rr) {

-               case (ROFF_REPARSE):
+               case ROFF_REPARSE:

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

-                                       curp->line, pos, NULL);
+                                   curp->line, pos, NULL);

                        pos = 0;
                        continue;
                        pos = 0;
                        continue;

-               case (ROFF_APPEND):
+               case ROFF_APPEND:

                        pos = (int)strlen(ln.buf);
                        continue;
                        pos = (int)strlen(ln.buf);
                        continue;

-               case (ROFF_RERUN):
+               case ROFF_RERUN:

                        goto rerun;
                        goto rerun;

-               case (ROFF_IGN):
+               case ROFF_IGN:

                        pos = 0;
                        continue;
                        pos = 0;
                        continue;

-               case (ROFF_ERR):
+               case ROFF_ERR:

                        assert(MANDOCLEVEL_FATAL <= curp->file_status);
                        break;
                        assert(MANDOCLEVEL_FATAL <= curp->file_status);
                        break;

-               case (ROFF_SO):
+               case ROFF_SO:

                        if (0 == (MPARSE_SO & curp->options) &&
                            (i >= (int)blk.sz || '\0' == blk.buf[i])) {
                                curp->sodest = mandoc_strdup(ln.buf + of);
                        if (0 == (MPARSE_SO & curp->options) &&
                            (i >= (int)blk.sz || '\0' == blk.buf[i])) {
                                curp->sodest = mandoc_strdup(ln.buf + of);


@@ -502,11 +495,15 @@ rerun:
                         * buffer because we're going to descend into
                         * the file recursively.
                         */
                         * buffer because we're going to descend into
                         * the file recursively.
                         */

-                       if (curp->secondary) 
+                       if (curp->secondary)

                                curp->secondary->sz -= pos + 1;
                        mparse_readfd(curp, -1, ln.buf + of);
                                curp->secondary->sz -= pos + 1;
                        mparse_readfd(curp, -1, ln.buf + of);

-                       if (MANDOCLEVEL_FATAL <= curp->file_status)
+                       if (MANDOCLEVEL_FATAL <= curp->file_status) {
+                               mandoc_vmsg(MANDOCERR_SO_FAIL,
+                                   curp, curp->line, pos,
+                                   ".so %s", ln.buf + of);

                                break;
                                break;

+                       }

                        pos = 0;
                        continue;
                default:
                        pos = 0;
                        continue;
                default:


@@ -531,7 +528,7 @@ rerun:
                if ( ! (curp->man || curp->mdoc))
                        pset(ln.buf + of, pos - of, curp);
 
                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.
                 * Lastly, push down into the parsers themselves.  One
                 * of these will have already been set in the pset()
                 * routine.


@@ -547,23 +544,23 @@ rerun:
                if (ROFF_TBL == rr)
                        while (NULL != (span = roff_span(curp->roff))) {
                                rc = curp->man ?
                if (ROFF_TBL == rr)
                        while (NULL != (span = roff_span(curp->roff))) {
                                rc = curp->man ?

-                                       man_addspan(curp->man, span) :
-                                       mdoc_addspan(curp->mdoc, span);
+                                   man_addspan(curp->man, span) :
+                                   mdoc_addspan(curp->mdoc, span);

                                if (0 == rc)
                                        break;
                        }
                else if (ROFF_EQN == rr)
                                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));
+                       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 ?
                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);
+                           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);
 
                if (0 == rc) {
                        assert(MANDOCLEVEL_FATAL <= curp->file_status);


@@ -673,6 +670,19 @@ mparse_end(struct mparse *curp)
        if (MANDOCLEVEL_FATAL <= curp->file_status)
                return;
 
        if (MANDOCLEVEL_FATAL <= curp->file_status)
                return;
 

+       if (curp->mdoc == NULL &&
+           curp->man == NULL &&
+           curp->sodest == NULL) {
+               if (curp->options & MPARSE_MDOC)
+                       curp->mdoc = curp->pmdoc;
+               else {
+                       if (curp->pman == NULL)
+                               curp->pman = man_alloc(curp->roff, curp,
+                                   curp->options & MPARSE_QUICK ? 1 : 0);
+                       curp->man = curp->pman;
+               }
+       }
+

        if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) {
                assert(MANDOCLEVEL_FATAL <= curp->file_status);
                return;
        if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) {
                assert(MANDOCLEVEL_FATAL <= curp->file_status);
                return;


@@ -683,12 +693,6 @@ mparse_end(struct mparse *curp)
                return;
        }
 
                return;
        }
 

-       if ( ! (curp->mdoc || curp->man || curp->sodest)) {
-               mandoc_msg(MANDOCERR_NOTMANUAL, curp, 1, 0, NULL);
-               curp->file_status = MANDOCLEVEL_FATAL;
-               return;
-       }
-

        roff_endparse(curp->roff);
 }
 
        roff_endparse(curp->roff);
 }
 


@@ -772,7 +776,7 @@ out:
 
 struct mparse *
 mparse_alloc(int options, enum mandoclevel wlevel,
 
 struct mparse *
 mparse_alloc(int options, enum mandoclevel wlevel,

-               mandocmsg mmsg, char *defos)
+               mandocmsg mmsg, const char *defos)

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


@@ -786,6 +790,14 @@ mparse_alloc(int options, enum mandoclevel wlevel,
        curp->defos = defos;
 
        curp->roff = roff_alloc(curp, options);
        curp->defos = defos;
 
        curp->roff = roff_alloc(curp, options);

+       if (curp->options & MPARSE_MDOC)
+               curp->pmdoc = mdoc_alloc(
+                   curp->roff, curp, curp->defos,
+                   curp->options & MPARSE_QUICK ? 1 : 0);
+       if (curp->options & MPARSE_MAN)
+               curp->pman = man_alloc(curp->roff, curp,
+                   curp->options & MPARSE_QUICK ? 1 : 0);
+

        return(curp);
 }
 
        return(curp);
 }
 


@@ -852,14 +864,14 @@ mandoc_vmsg(enum mandocerr t, struct mparse *m,
        va_list          ap;
 
        va_start(ap, fmt);
        va_list          ap;
 
        va_start(ap, fmt);

-       vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
+       (void)vsnprintf(buf, sizeof(buf), fmt, ap);

        va_end(ap);
 
        mandoc_msg(t, m, ln, pos, buf);
 }
 
 void
        va_end(ap);
 
        mandoc_msg(t, m, ln, pos, buf);
 }
 
 void

-mandoc_msg(enum mandocerr er, struct mparse *m, 
+mandoc_msg(enum mandocerr er, struct mparse *m,

                int ln, int col, const char *msg)
 {
        enum mandoclevel level;
                int ln, int col, const char *msg)
 {
        enum mandoclevel level;