Formatting.

[mandoc.git] / mdocterm.c

diff --git a/mdocterm.c b/mdocterm.c
index 6e449c697483e17edaf5253ed3a755e9d7a2565d..a8a06c1288866645f19a9ddd79063674dcf1ae08 100644 (file)
--- a/mdocterm.c
+++ b/mdocterm.c
@@ -1,4 +1,4 @@
-       /* $Id: mdocterm.c,v 1.1 2009/02/21 21:00:06 kristaps Exp $ */
+       /* $Id: mdocterm.c,v 1.10 2009/02/25 12:27:37 kristaps Exp $ */

 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *


@@ -16,274 +16,574 @@
  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  * PERFORMANCE OF THIS SOFTWARE.
  */
  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  * PERFORMANCE OF THIS SOFTWARE.
  */

-#include <sys/stat.h>
-#include <sys/param.h>
-

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

-#include <fcntl.h>
+#include <ctype.h>

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

-#include <unistd.h>

 
 

-#include "mdoc.h"
+#ifndef __OpenBSD__
+#include <time.h>
+#endif
+
+#include "mmain.h"

 #include "term.h"
 
 #include "term.h"
 

-#define        MD_LINE_SZ      (256)           /* Max input line size. */
-
-struct md_parse {
-       int               warn;         /* Warning flags. */
-#define        MD_WARN_SYNTAX   (1 << 0)       /* Show syntax warnings. */
-#define        MD_WARN_COMPAT   (1 << 1)       /* Show compat warnings. */
-#define        MD_WARN_ALL      (0x03)         /* Show all warnings. */
-#define        MD_WARN_ERR      (1 << 2)       /* Make warnings->errors. */
-       int               dbg;          /* Debug level. */
-       struct mdoc      *mdoc;         /* Active parser. */
-       char             *buf;          /* Input buffer. */
-       u_long            bufsz;        /* Input buffer size. */
-       char             *in;           /* Input file name. */
-       int               fdin;         /* Input file desc. */
-};
+#ifdef __NetBSD__
+#define xisspace(x) isspace((int)(x))
+#else
+#define        xisspace(x) isspace((x))
+#endif

 
 

-extern char             *__progname;
+enum   termstyle {
+       STYLE_CLEAR,
+       STYLE_BOLD,
+       STYLE_UNDERLINE
+};

 
 

-static void              usage(void);
-static int               getsopts(struct md_parse *, char *);
-static int               parse(struct md_parse *);
-static void              msg_msg(void *, int, int, const char *);
-static int               msg_err(void *, int, int, const char *);
-static int               msg_warn(void *, int, int, 
-                               enum mdoc_warn, const char *);
+static void              body(struct termp *,
+                               const struct mdoc_meta *,
+                               const struct mdoc_node *);
+static void              header(struct termp *,
+                               const struct mdoc_meta *);
+static void              footer(struct termp *,
+                               const struct mdoc_meta *);
+
+static void              pword(struct termp *, const char *, size_t);
+static void              pescape(struct termp *, 
+                               const char *, size_t *, size_t);
+static void              chara(struct termp *, char);
+static void              stringa(struct termp *, const char *);
+static void              style(struct termp *, enum termstyle);

 
 #ifdef __linux__
 
 #ifdef __linux__

-extern int               getsubopt(char **, char *const *, char **);
+extern size_t            strlcat(char *, const char *, size_t);
+extern size_t            strlcpy(char *, const char *, size_t);

 #endif
 
 #endif
 

+

 int
 main(int argc, char *argv[])
 {
 int
 main(int argc, char *argv[])
 {

-       struct md_parse  p;
-       struct mdoc_cb   cb;
-       struct stat      st;
-       int              c;
-       extern char     *optarg;
-       extern int       optind;
-
-       (void)memset(&p, 0, sizeof(struct md_parse));
-
-       while (-1 != (c = getopt(argc, argv, "vW:")))
-               switch (c) {
-               case ('v'):
-                       p.dbg++;
-                       break;
-               case ('W'):
-                       if ( ! getsopts(&p, optarg))
-                               return(0);
-                       break;
-               default:
-                       usage();
-                       return(0);
-               }
+       struct mmain    *p;
+       const struct mdoc *mdoc;
+       struct termp     termp;

 
 

-       argv += optind;
-       argc -= optind;
+       p = mmain_alloc();

 
 

-       /* Initialise the input file. */
+       if ( ! mmain_getopt(p, argc, argv, NULL, NULL, NULL, NULL))
+               mmain_exit(p, 1);

 
 

-       p.in = "-";
-       p.fdin = STDIN_FILENO;
+       if (NULL == (mdoc = mmain_mdoc(p)))
+               mmain_exit(p, 1);

 
 

-       if (argc > 0) {
-               p.in = *argv++;
-               p.fdin = open(p.in, O_RDONLY, 0);
-               if (-1 == p.fdin)
-                       err(1, "%s", p.in);
-       }
+       termp.maxrmargin = 80; /* XXX */
+       termp.rmargin = termp.maxrmargin;
+       termp.maxcols = 1024;
+       termp.offset = termp.col = 0;
+       termp.flags = TERMP_NOSPACE;

 
 

-       /* Allocate a buffer to be BUFSIZ/block size. */
+       if (NULL == (termp.buf = malloc(termp.maxcols)))
+               err(1, "malloc");

 
 

-       if (-1 == fstat(p.fdin, &st)) {
-               warn("%s", p.in);
-               p.bufsz = BUFSIZ;
-       } else 
-               p.bufsz = MAX(st.st_blksize, BUFSIZ);
+       header(&termp, mdoc_meta(mdoc));
+       body(&termp, mdoc_meta(mdoc), mdoc_node(mdoc));
+       footer(&termp, mdoc_meta(mdoc));

 
 

-       p.buf = malloc(p.bufsz);
-       if (NULL == p.buf)
-               err(1, "malloc");
+       free(termp.buf);

 
 

-       /* Allocate the parser. */
+       mmain_exit(p, 0);
+       /* NOTREACHED */
+}

 
 

-       cb.mdoc_err = msg_err;
-       cb.mdoc_warn = msg_warn;
-       cb.mdoc_msg = msg_msg;

 
 

-       p.mdoc = mdoc_alloc(&p, &cb);
+void
+flushln(struct termp *p)
+{
+       size_t           i, j, vsz, vis, maxvis;

 
 

-       /* Parse the input file. */
+       /*
+        * First, establish the maximum columns of "visible" content.
+        * This is usually the difference between the right-margin and
+        * an indentation, but can be, for tagged lists or columns, a
+        * small set of values.
+        */
+
+       assert(p->offset < p->rmargin);
+       maxvis = p->rmargin - p->offset;
+       vis = 0;
+
+       /*
+        * If in the standard case (left-justified), then begin with our
+        * indentation, otherwise (columns, etc.) just start spitting
+        * out text.
+        */

 
 

-       c = parse(&p);
-       free(p.buf);
+       if ( ! (p->flags & TERMP_NOLPAD))
+               /* LINTED */
+               for (j = 0; j < p->offset; j++)
+                       putchar(' ');

 
 

-       if (STDIN_FILENO != p.fdin && -1 == close(p.fdin))
-               warn("%s", p.in);
+       /*
+        * If we're literal, print out verbatim.
+        */
+       if (p->flags & TERMP_LITERAL) {
+               /* FIXME: count non-printing chars. */
+               for (i = 0; i < p->col; i++)
+                       putchar(p->buf[i]);
+               putchar('\n');
+               p->col = 0;
+               return;
+       }

 
 

-       if (0 == c) {
-               mdoc_free(p.mdoc);
-               return(EXIT_FAILURE);
+       for (i = 0; i < p->col; i++) {
+               /*
+                * Count up visible word characters.  Control sequences
+                * (starting with the CSI) aren't counted. 
+                */
+               assert( ! xisspace(p->buf[i]));
+
+               /* LINTED */
+               for (j = i, vsz = 0; j < p->col; j++) {
+                       if (xisspace(p->buf[j]))
+                               break;
+                       else if (27 == p->buf[j]) {
+                               assert(j + 4 <= p->col);
+                               j += 3;
+                       } else
+                               vsz++;
+               }
+               assert(vsz > 0);
+
+               /*
+                * If a word is too long and we're within a line, put it
+                * on the next line.  Puke if we're being asked to write
+                * something that will exceed the right margin (i.e.,
+                * from a fresh line or when we're not allowed to break
+                * the line with TERMP_NOBREAK).
+                */
+
+               if (vis && vis + vsz >= maxvis) {
+                       /* FIXME */
+                       if (p->flags & TERMP_NOBREAK)
+                               errx(1, "word breaks right margin");
+                       putchar('\n');
+                       for (j = 0; j < p->offset; j++)
+                               putchar(' ');
+                       vis = 0;
+               } else if (vis + vsz >= maxvis)
+                       /* FIXME */
+                       errx(1, "word breaks right margin");
+
+               /* 
+                * Write out the word and a trailing space.  Omit the
+                * space if we're the last word in the line.
+                */
+
+               for ( ; i < p->col; i++) {
+                       if (xisspace(p->buf[i]))
+                               break;
+                       putchar(p->buf[i]);
+               }
+               vis += vsz;
+               if (i < p->col) {
+                       putchar(' ');
+                       vis++;
+               }

        }
 
        }
 

-       /* If the parse succeeded, print it out. */
+       /*
+        * If we're not to right-marginalise it (newline), then instead
+        * pad to the right margin and stay off.
+        */

 
 

-       termprint(mdoc_node(p.mdoc), mdoc_meta(p.mdoc));
-       mdoc_free(p.mdoc);
+       if (p->flags & TERMP_NOBREAK) {
+               for ( ; vis <= maxvis; vis++)
+                       putchar(' ');
+       } else
+               putchar('\n');

 
 

-       return(EXIT_SUCCESS);
+       p->col = 0;

 }
 
 
 }
 
 

-static int
-getsopts(struct md_parse *p, char *arg)
+void
+newln(struct termp *p)

 {
 {

-       char            *v;
-       char            *toks[] = { "all", "compat", 
-                               "syntax", "error", NULL };
-
-       while (*arg) 
-               switch (getsubopt(&arg, toks, &v)) {
-               case (0):
-                       p->warn |= MD_WARN_ALL;
-                       break;
-               case (1):
-                       p->warn |= MD_WARN_COMPAT;
-                       break;
-               case (2):
-                       p->warn |= MD_WARN_SYNTAX;
-                       break;
-               case (3):
-                       p->warn |= MD_WARN_ERR;
-                       break;
-               default:
-                       usage();
-                       return(0);
-               }

 
 

-       return(1);
+       /* 
+        * A newline only breaks an existing line; it won't assert
+        * vertical space.
+        */
+       p->flags |= TERMP_NOSPACE;
+       if (0 == p->col) 
+               return;
+       flushln(p);

 }
 
 
 }
 
 

-static int
-parse(struct md_parse *p)
+void
+vspace(struct termp *p)

 {
 {

-       ssize_t          sz, i;
-       size_t           pos;
-       char             line[MD_LINE_SZ];
-       int              lnn;

 
        /*
 
        /*

-        * This is a little more complicated than fgets.  TODO: have
-        * some benchmarks that show it's faster (note that I want to
-        * check many, many manuals simultaneously, so speed is
-        * important).  Fill a buffer (sized to the block size) with a
-        * single read, then parse \n-terminated lines into a line
-        * buffer, which is passed to the parser.  Hard-code the line
-        * buffer to a particular size -- a reasonable assumption.
+        * Asserts a vertical space (a full, empty line-break between
+        * lines).

         */
         */

+       newln(p);
+       putchar('\n');
+}
+
+
+static void
+stringa(struct termp *p, const char *s)
+{
+
+       /* XXX - speed up if not passing to chara. */
+       for ( ; *s; s++)
+               chara(p, *s);
+}
+
+
+static void
+chara(struct termp *p, char c)
+{
+
+       /* TODO: dynamically expand the buffer. */
+       if (p->col + 1 >= p->maxcols)
+               errx(1, "line overrun");
+       p->buf[(p->col)++] = c;
+}
+
+
+static void
+style(struct termp *p, enum termstyle esc)
+{
+
+       if (p->col + 4 >= p->maxcols)
+               errx(1, "line overrun");
+
+       p->buf[(p->col)++] = 27;
+       p->buf[(p->col)++] = '[';
+       switch (esc) {
+       case (STYLE_CLEAR):
+               p->buf[(p->col)++] = '0';
+               break;
+       case (STYLE_BOLD):
+               p->buf[(p->col)++] = '1';
+               break;
+       case (STYLE_UNDERLINE):
+               p->buf[(p->col)++] = '4';
+               break;
+       default:
+               abort();
+               /* NOTREACHED */
+       }
+       p->buf[(p->col)++] = 'm';
+}
+
+
+static void
+pescape(struct termp *p, const char *word, size_t *i, size_t len)
+{
+
+       (*i)++;
+       assert(*i < len);
+
+       if ('(' == word[*i]) {
+               /* Two-character escapes. */
+               (*i)++;
+               assert(*i + 1 < len);
+
+               if ('r' == word[*i] && 'B' == word[*i + 1])
+                       chara(p, ']');
+               else if ('l' == word[*i] && 'B' == word[*i + 1])
+                       chara(p, '[');
+               else if ('<' == word[*i] && '-' == word[*i + 1])
+                       stringa(p, "<-");
+               else if ('-' == word[*i] && '>' == word[*i + 1])
+                       stringa(p, "->");
+               else if ('l' == word[*i] && 'q' == word[*i + 1])
+                       chara(p, '\"');
+               else if ('r' == word[*i] && 'q' == word[*i + 1])
+                       chara(p, '\"');
+
+               (*i)++;
+               return;

 
 

-       for (lnn = 1, pos = 0; ; ) {
-               if (-1 == (sz = read(p->fdin, p->buf, p->bufsz))) {
-                       warn("%s", p->in);
-                       return(0);
-               } else if (0 == sz) 
+       } else if ('[' != word[*i]) {
+               /* One-character escapes. */
+               switch (word[*i]) {
+               case ('\\'):
+                       /* FALLTHROUGH */
+               case ('\''):
+                       /* FALLTHROUGH */
+               case ('`'):
+                       /* FALLTHROUGH */
+               case ('-'):
+                       /* FALLTHROUGH */
+               case ('.'):
+                       chara(p, word[*i]);
+               default:

                        break;
                        break;

+               }
+               return;
+       }
+       /* n-character escapes. */
+}

 
 

-               for (i = 0; i < sz; i++) {
-                       if ('\n' != p->buf[i]) {
-                               if (pos < sizeof(line)) {
-                                       line[(int)pos++] = p->buf[(int)i];
-                                       continue;
-                               }
-                               warnx("%s: line %d too long", p->in, lnn);
-                               return(0);
-                       }
-       
-                       line[(int)pos] = 0;
-                       if ( ! mdoc_parseln(p->mdoc, lnn, line))
-                               return(0);

 
 

-                       lnn++;
-                       pos = 0;
+static void
+pword(struct termp *p, const char *word, size_t len)
+{
+       size_t           i;
+
+       /*assert(len > 0);*/ /* Can be, if literal. */
+
+       if ( ! (p->flags & TERMP_NOSPACE) && 
+                       ! (p->flags & TERMP_LITERAL))
+               chara(p, ' ');
+
+       p->flags &= ~TERMP_NOSPACE;
+
+       if (p->flags & TERMP_BOLD)
+               style(p, STYLE_BOLD);
+       if (p->flags & TERMP_UNDERLINE)
+               style(p, STYLE_UNDERLINE);
+
+       for (i = 0; i < len; i++) {
+               if ('\\' == word[i]) {
+                       pescape(p, word, &i, len);
+                       continue;

                }
                }

+               chara(p, word[i]);

        }
 
        }
 

-       return(mdoc_endparse(p->mdoc));
+       if (p->flags & TERMP_BOLD ||
+                       p->flags & TERMP_UNDERLINE)
+               style(p, STYLE_CLEAR);

 }
 
 
 }
 
 

-static int
-msg_err(void *arg, int line, int col, const char *msg)
+void
+word(struct termp *p, const char *word)

 {
 {

-       struct md_parse  *p;
+       size_t           i, j, len;
+
+       if (p->flags & TERMP_LITERAL) {
+               pword(p, word, strlen(word));
+               return;
+       }

 
 

-       p = (struct md_parse *)arg;
+       len = strlen(word);
+       assert(len > 0);

 
 

-       warnx("%s:%d: error: %s (column %d)", 
-                       p->in, line, msg, col);
-       return(0);
+       if (mdoc_isdelim(word)) {
+               if ( ! (p->flags & TERMP_IGNDELIM))
+                       p->flags |= TERMP_NOSPACE;
+               p->flags &= ~TERMP_IGNDELIM;
+       }
+
+       /* LINTED */
+       for (j = i = 0; i < len; i++) {
+               if ( ! xisspace(word[i])) {
+                       j++;
+                       continue;
+               }
+               if (0 == j)
+                       continue;
+               assert(i >= j);
+               pword(p, &word[i - j], j);
+               j = 0;
+       }
+       if (j > 0) {
+               assert(i >= j);
+               pword(p, &word[i - j], j);
+       }

 }
 
 
 static void
 }
 
 
 static void

-msg_msg(void *arg, int line, int col, const char *msg)
+body(struct termp *p, const struct mdoc_meta *meta,
+               const struct mdoc_node *node)

 {
 {

-       struct md_parse  *p;
+       int              dochild;
+       struct termpair  pair;

 
 

-       p = (struct md_parse *)arg;
+       /* Pre-processing. */

 
 

-       if (0 == p->dbg)
-               return;
+       dochild = 1;
+       pair.type = 0;
+       pair.offset = 0;
+       pair.flag = 0;
+
+       if (MDOC_TEXT != node->type) {
+               if (termacts[node->tok].pre)
+                       if ( ! (*termacts[node->tok].pre)(p, &pair, meta, node))
+                               dochild = 0;
+       } else /* MDOC_TEXT == node->type */
+               word(p, node->data.text.string);
+
+       /* Children. */
+
+       if (TERMPAIR_FLAG & pair.type)
+               p->flags |= pair.flag;
+
+       if (dochild && node->child)
+               body(p, meta, node->child);
+
+       if (TERMPAIR_FLAG & pair.type)
+               p->flags &= ~pair.flag;
+
+       /* Post-processing. */
+
+       if (MDOC_TEXT != node->type)
+               if (termacts[node->tok].post)
+                       (*termacts[node->tok].post)(p, &pair, meta, node);
+
+       /* Siblings. */
+
+       if (node->next)
+               body(p, meta, node->next);
+}

 
 

-       warnx("%s:%d: debug: %s (column %d)", 
-                       p->in, line, msg, col);
+
+static void
+footer(struct termp *p, const struct mdoc_meta *meta)
+{
+       struct tm       *tm;
+       char            *buf, *os;
+       size_t           sz, osz, ssz, i;
+
+       if (NULL == (buf = malloc(p->rmargin)))
+               err(1, "malloc");
+       if (NULL == (os = malloc(p->rmargin)))
+               err(1, "malloc");
+
+       tm = localtime(&meta->date);
+
+#ifdef __OpenBSD__
+       if (NULL == strftime(buf, p->rmargin, "%B %d, %Y", tm))
+#else
+       if (0 == strftime(buf, p->rmargin, "%B %d, %Y", tm))
+#endif
+               err(1, "strftime");
+
+       osz = strlcpy(os, meta->os, p->rmargin);
+
+       sz = strlen(buf);
+       ssz = sz + osz + 1;
+
+       if (ssz > p->rmargin) {
+               ssz -= p->rmargin;
+               assert(ssz <= osz);
+               os[osz - ssz] = 0;
+               ssz = 1;
+       } else
+               ssz = p->rmargin - ssz + 1;
+
+       printf("\n");
+       printf("%s", os);
+       for (i = 0; i < ssz; i++)
+               printf(" ");
+
+       printf("%s\n", buf);
+       fflush(stdout);
+
+       free(buf);
+       free(os);

 }
 
 
 }
 
 

-static int
-msg_warn(void *arg, int line, int col, 
-               enum mdoc_warn type, const char *msg)
+static void
+header(struct termp *p, const struct mdoc_meta *meta)

 {
 {

-       struct md_parse  *p;
+       char            *buf, *title;
+       const char      *pp, *msec;
+       size_t           ssz, tsz, ttsz, i;;

 
 

-       p = (struct md_parse *)arg;
+       if (NULL == (buf = malloc(p->rmargin)))
+               err(1, "malloc");
+       if (NULL == (title = malloc(p->rmargin)))
+               err(1, "malloc");

 
 

-       switch (type) {
-       case (WARN_COMPAT):
-               if (p->warn & MD_WARN_COMPAT)
+       if (NULL == (pp = mdoc_vol2a(meta->vol)))
+               switch (meta->msec) {
+               case (MSEC_1):
+                       /* FALLTHROUGH */
+               case (MSEC_6):
+                       /* FALLTHROUGH */
+               case (MSEC_7):
+                       pp = mdoc_vol2a(VOL_URM);
+                       break;
+               case (MSEC_8):
+                       pp = mdoc_vol2a(VOL_SMM);
+                       break;
+               case (MSEC_2):
+                       /* FALLTHROUGH */
+               case (MSEC_3):
+                       /* FALLTHROUGH */
+               case (MSEC_4):
+                       /* FALLTHROUGH */
+               case (MSEC_5):
+                       pp = mdoc_vol2a(VOL_PRM);

                        break;
                        break;

-               return(1);
-       case (WARN_SYNTAX):
-               if (p->warn & MD_WARN_SYNTAX)
+               case (MSEC_9):
+                       pp = mdoc_vol2a(VOL_KM);

                        break;
                        break;

-               return(1);
+               default:
+                       /* FIXME: capitalise. */
+                       if (NULL == (pp = mdoc_msec2a(meta->msec)))
+                               pp = mdoc_msec2a(MSEC_local);
+                       break;
+               }
+       assert(pp);
+
+       tsz = strlcpy(buf, pp, p->rmargin);
+       assert(tsz < p->rmargin);
+
+       if ((pp = mdoc_arch2a(meta->arch))) {
+               tsz = strlcat(buf, " (", p->rmargin);
+               assert(tsz < p->rmargin);
+               tsz = strlcat(buf, pp, p->rmargin);
+               assert(tsz < p->rmargin);
+               tsz = strlcat(buf, ")", p->rmargin);
+               assert(tsz < p->rmargin);

        }
 
        }
 

-       warnx("%s:%d: warning: %s (column %d)", 
-                       p->in, line, msg, col);
+       ttsz = strlcpy(title, meta->title, p->rmargin);

 
 

-       if ( ! (p->warn & MD_WARN_ERR))
-               return(1);
+       if (NULL == (msec = mdoc_msec2a(meta->msec)))
+               msec = "";

 
 

-       warnx("%s: considering warnings as errors", __progname);
-       return(0);
-}
+       ssz = (2 * (ttsz + 2 + strlen(msec))) + tsz + 2;

 
 

+       if (ssz > p->rmargin) {
+               if ((ssz -= p->rmargin) % 2)
+                       ssz++;
+               ssz /= 2;
+       
+               assert(ssz <= ttsz);
+               title[ttsz - ssz] = 0;
+               ssz = 1;
+       } else
+               ssz = ((p->rmargin - ssz) / 2) + 1;

 
 

-static void
-usage(void)
-{
+       printf("%s(%s)", title, msec);

 
 

-       warnx("usage: %s [-v] [-Wwarn...] [infile]", __progname);
-}
+       for (i = 0; i < ssz; i++)
+               printf(" ");

 
 

+       printf("%s", buf);
+
+       for (i = 0; i < ssz; i++)
+               printf(" ");
+
+       printf("%s(%s)\n", title, msec);
+       fflush(stdout);
+
+       free(title);
+       free(buf);
+}