=================================================================== RCS file: /cvs/pod2mdoc/pod2mdoc.c,v retrieving revision 1.42 retrieving revision 1.54 diff -u -p -r1.42 -r1.54 --- pod2mdoc/pod2mdoc.c 2015/02/14 10:35:02 1.42 +++ pod2mdoc/pod2mdoc.c 2015/02/19 15:26:45 1.54 @@ -1,4 +1,4 @@ -/* $Id: pod2mdoc.c,v 1.42 2015/02/14 10:35:02 schwarze Exp $ */ +/* $Id: pod2mdoc.c,v 1.54 2015/02/19 15:26:45 schwarze Exp $ */ /* * Copyright (c) 2014 Kristaps Dzonsons * Copyright (c) 2014, 2015 Ingo Schwarze @@ -185,7 +185,11 @@ outbuf_flush(struct state *st) if (OUST_TXT == st->oust && st->wantws) putchar(' '); - fputs(st->outbuf, stdout); + if (OUST_MAC == st->oust && '"' == *st->outbuf) + printf("\\(dq%s", st->outbuf + 1); + else + fputs(st->outbuf, stdout); + *st->outbuf = '\0'; st->outbuflen = 0; @@ -480,7 +484,6 @@ formatcode(struct state *st, const char *buf, size_t * { size_t i, j, dsz; enum fmt fmt; - int wantws; unsigned char uc; assert(*start + 1 < end); @@ -559,45 +562,43 @@ formatcode(struct state *st, const char *buf, size_t * if (FMT__MAX != fmt && !nomacro) { /* - * We may already have wantws if there was whitespace - * before the code ("text Boust ? st->wantws : ! st->outbuflen); - - /* * If we are on a text line and there is no * whitespace before our content, we have to make * the previous word a prefix to the macro line. - * In the following, mdoc_newln() must not be used - * lest we clobber out output state. */ - if (OUST_MAC != st->oust && ! wantws) { + if (OUST_MAC != st->oust && ' ' != buf[*start] && + st->outbuflen) { if (OUST_NL != st->oust) - putchar('\n'); + mdoc_newln(st); printf(".Pf "); - st->wantws = 0; + st->oust = OUST_MAC; + st->wantws = 1; } outbuf_flush(st); - /* Whitespace is easier to suppress on macro lines. */ + /* + * Whitespace is easier to suppress on macro lines. + * We may already have wantws if there was whitespace + * before the code ("text Boust && ! wantws) - printf(" Ns "); + if (OUST_MAC == st->oust && ' ' != buf[*start] && + ! st->wantws) + printf(" Ns"); /* Unless we are on a macro line, start one. */ - if (OUST_MAC != st->oust && wantws) { + if (OUST_MAC != st->oust) { if (OUST_NL != st->oust) - putchar('\n'); + mdoc_newln(st); putchar('.'); + st->oust = OUST_MAC; } else putchar(' '); + st->wantws = 1; /* * Print the macro corresponding to this format code, @@ -657,8 +658,6 @@ formatcode(struct state *st, const char *buf, size_t * default: abort(); } - st->oust = OUST_MAC; - st->wantws = 1; } else outbuf_flush(st); @@ -1059,11 +1058,26 @@ verbatim(struct state *st, char *buf, size_t start, si { size_t i, ift, ifo, ifa, ifc, inl; char *cp, *cp2; - int nopen; + int indisplay, nopen, wantsp; - if ( ! st->parsing || st->paused || start == end) + if (st->paused || ! st->parsing) return; + + indisplay = wantsp = 0; + again: + if (start == end) { + if (indisplay) + puts(".Ed"); + return; + } + + if ('\n' == buf[start]) { + wantsp = 1; + start++; + goto again; + } + /* * If we're in the SYNOPSIS, see if we're an #include block. * If we are, then print the "In" macro and re-loop. @@ -1075,7 +1089,7 @@ again: while (i < end && buf[i] == ' ') i++; if (i == end) - return; + goto again; /* We're an include block! */ if (end - i > 10 && @@ -1083,6 +1097,9 @@ again: start = i + 10; while (start < end && ' ' == buf[start]) start++; + if (indisplay) + puts(".Ed"); + indisplay = wantsp = 0; fputs(".In ", stdout); /* Stop til the '>' marker or we hit eoln. */ while (start < end && @@ -1098,6 +1115,9 @@ again: /* Other preprocessor directives. */ if ('#' == buf[i]) { + if (indisplay) + puts(".Ed"); + indisplay = wantsp = 0; fputs(".Fd ", stdout); start = i; while(start < end && '\n' != buf[start]) @@ -1105,6 +1125,24 @@ again: putchar('\n'); if (start < end && '\n' == buf[start]) start++; + + /* Remember #define for Dv or Fn. */ + + if (strncmp(buf + i + 1, "define", 6) || + ! isspace((unsigned char)buf[i + 7])) + goto again; + + ifo = i + 7; + while (ifo < start && + isspace((unsigned char)buf[ifo])) + ifo++; + ifa = ifo; + while ('_' == buf[ifa] || + isalnum((unsigned char)buf[ifa])) + ifa++; + dict_put(buf + ifo, ifa - ifo, + '(' == buf[ifa] ? MDOC_Fo : MDOC_Dv); + goto again; } @@ -1120,6 +1158,8 @@ again: break; } switch (buf[i]) { + case '\t': + /* FALLTHROUGH */ case ' ': if ( ! ifa) ifo = i; @@ -1157,6 +1197,9 @@ again: buf[i] = ' '; buf[ifo++] = '\0'; register_type(buf + ift); + if (indisplay) + puts(".Ed"); + indisplay = wantsp = 0; printf(".Ft %s", buf + ift); if (buf[ifo] == '*') { fputs(" *", stdout); @@ -1180,10 +1223,13 @@ again: if ('\0' != *cp2) dict_put(cp2, 0, MDOC_Fa); register_type(buf + ifa); - printf(".Fa \"%s\"\n", buf + ifa); + if (strchr(buf + ifa, ' ') == NULL) + printf(".Fa %s\n", buf + ifa); + else + printf(".Fa \"%s\"\n", buf + ifa); if (cp == NULL) break; - while (*cp == ' ') + while (*cp == ' ' || *cp == '\t') cp++; ifa = cp - buf; } @@ -1194,28 +1240,37 @@ again: buf[inl] = '\0'; puts(buf + ifc); } - start = inl + 1; - if (start < end) - goto again; - return; + start = inl < end ? inl + 1 : end; + goto again; } } - - puts(".Bd -literal"); - for (last = ' '; start < end; start++) { + + if ( ! indisplay) + puts(".Bd -literal"); + else if (wantsp) + putchar('\n'); + indisplay = 1; + wantsp = 0; + + for (last = '\n'; start < end; start++) { /* * Handle accidental macros (newline starting with * control character) and escapes. */ - if ('\n' == last) + if ('\n' == last) { + if ('\n' == buf[start]) + goto again; if ('.' == buf[start] || '\'' == buf[start]) printf("\\&"); + } putchar(last = buf[start]); if ('\\' == buf[start]) printf("e"); } - putchar(last = '\n'); - puts(".Ed"); + if ('\n' != last) + putchar('\n'); + if (indisplay) + puts(".Ed"); } /* @@ -1285,7 +1340,7 @@ donamenm(struct state *st, const char *buf, size_t *st assert(OUST_NL == st->oust); assert(st->wantws); - while (*start < end && ' ' == buf[*start]) + while (*start < end && isspace((unsigned char)buf[*start])) (*start)++; if (end == *start) { @@ -1306,7 +1361,7 @@ donamenm(struct state *st, const char *buf, size_t *st printf(" ,"); mdoc_newln(st); (*start)++; - while (*start < end && ' ' == buf[*start]) + while (*start < end && isspace((unsigned char)buf[*start])) (*start)++; } } @@ -1328,8 +1383,10 @@ donamenm(struct state *st, const char *buf, size_t *st static void ordinary(struct state *st, const char *buf, size_t start, size_t end) { - size_t i, j, opstack; - int seq; + size_t i, j, opstack, wend; + enum mdoc_type mtype; + int eos, noeos, seq; + char savechar; if ( ! st->parsing || st->paused) return; @@ -1342,7 +1399,8 @@ ordinary(struct state *st, const char *buf, size_t sta */ if (SECT_NAME == st->sect) { for (i = end - 2; i > start; i--) - if ('-' == buf[i] && ' ' == buf[i + 1]) + if ('-' == buf[i] && + isspace((unsigned char)buf[i + 1])) break; if ('-' == buf[i]) { j = i; @@ -1352,7 +1410,8 @@ ordinary(struct state *st, const char *buf, size_t sta break; donamenm(st, buf, &start, i + 1); start = j + 1; - while (start < end && ' ' == buf[start]) + while (start < end && + isspace((unsigned char)buf[start])) start++; formatcodeln(st, "Nd", buf, &start, end, 1); mdoc_newln(st); @@ -1398,23 +1457,67 @@ ordinary(struct state *st, const char *buf, size_t sta /* Merely buffer non-whitespace. */ last = buf[start++]; - if ( ! isspace(last)) { + if ( ! isspace(last)) outbuf_addchar(st); + if (start < end && + ! isspace((unsigned char)buf[start - 1]) && + ! isspace((unsigned char)buf[start])) continue; - } - /* Detect function names. */ + /* + * Found the end of a word. + * Rewind trailing delimiters. + */ - if (st->outbuflen > 2 && - ')' == st->outbuf[st->outbuflen - 1] && - '(' == st->outbuf[st->outbuflen - 2] && - dict_get(st->outbuf, st->outbuflen - 2) == - MDOC_Fo) { - st->outbuflen -= 2; - st->outbuf[st->outbuflen] = '\0'; - mdoc_newln(st); - fputs(".Fn ", stdout); - st->oust = OUST_MAC; + eos = noeos = 0; + for (wend = st->outbuflen; wend; wend--) + if ('.' == st->outbuf[wend - 1] || + '!' == st->outbuf[wend - 1] || + '?' == st->outbuf[wend - 1]) + eos = 1; + else if ('|' == st->outbuf[wend - 1] || + ',' == st->outbuf[wend - 1] || + ';' == st->outbuf[wend - 1] || + ':' == st->outbuf[wend - 1]) + noeos = 1; + else if ('\'' != st->outbuf[wend - 1] && + '"' != st->outbuf[wend - 1] && + ')' != st->outbuf[wend - 1] && + ']' != st->outbuf[wend - 1]) + break; + eos &= ! noeos; + + /* + * Detect function names. + */ + + mtype = MDOC_Fa; + savechar = '\0'; + if (wend && ')' == st->outbuf[wend] && + '(' == st->outbuf[wend - 1]) { + mtype = dict_get(st->outbuf, --wend); + if (MDOC_Dv == mtype) + mtype = MDOC_Fo; + if (MDOC_Fo == mtype || MDOC_MAX == mtype) { + st->outbuflen = wend; + st->outbuf[wend] = '\0'; + mdoc_newln(st); + if (MDOC_Fo == mtype) + fputs(".Fn ", stdout); + else + fputs(".Xr ", stdout); + st->oust = OUST_MAC; + } + } else { + mtype = dict_get(st->outbuf, wend); + if (MDOC_Dv == mtype) { + savechar = st->outbuf[wend]; + st->outbuf[wend] = '\0'; + mdoc_newln(st); + fputs(".Dv ", stdout); + st->oust = OUST_MAC; + } else + mtype = MDOC_Fa; } /* @@ -1429,17 +1532,26 @@ ordinary(struct state *st, const char *buf, size_t sta * end text lines at the end of sentences. */ - if (OUST_MAC == st->oust || (start > 3 && - ('.' == buf[start - 2] || - '!' == buf[start - 2] || - '?' == buf[start - 2]) && - islower((unsigned char)buf[start - 3]) && - islower((unsigned char)buf[start - 4]))) + if (OUST_MAC == st->oust || (eos && wend > 1 && + islower((unsigned char)st->outbuf[wend - 1]))) { + if (MDOC_MAX == mtype) + fputs(" 3", stdout); + if (MDOC_Fa != mtype) { + if (MDOC_Dv == mtype) + st->outbuf[wend] = savechar; + else + wend += 2; + while ('\0' != st->outbuf[wend]) + printf(" %c", + st->outbuf[wend++]); + } mdoc_newln(st); + } /* Advance to the next word. */ - while (isspace((unsigned char)buf[start])) + while ('\n' != buf[start] && + isspace((unsigned char)buf[start])) start++; st->wantws = 1; } @@ -1453,12 +1565,11 @@ ordinary(struct state *st, const char *buf, size_t sta * XXX Some punctuation characters * are not handled yet. */ - if ((start == end - 1 || - (start < end - 1 && - (' ' == buf[start + 1] || - '\n' == buf[start + 1]))) && - ('.' == buf[start] || - ',' == buf[start])) { + if ((start == end - 1 || + (start < end - 1 && + (' ' == buf[start + 1] || + '\n' == buf[start + 1]))) && + NULL != strchr("|.,;:?!)]", buf[start])) { putchar(' '); putchar(buf[start++]); } @@ -1546,9 +1657,10 @@ dofile(const struct args *args, const char *fname, { char datebuf[64]; struct state st; - const char *fbase, *fext, *section, *date; + const char *fbase, *fext, *section, *date, *format; char *title, *cp; - size_t sup, end, i, cur = 0; + size_t cur, end; + int verb; if (0 == sz) return; @@ -1591,8 +1703,12 @@ dofile(const struct args *args, const char *fname, /* Date. Or the given "tm" if not supplied. */ - if (NULL == (date = args->date)) { - strftime(datebuf, sizeof(datebuf), "%B %d, %Y", tm); + date = args->date; + format = (NULL == date) ? "%B %d, %Y" : + strcmp(date, "Mdocdate") ? NULL : "$" "Mdocdate: %B %d %Y $"; + + if (NULL != format) { + strftime(datebuf, sizeof(datebuf), format, tm); date = datebuf; } @@ -1616,24 +1732,32 @@ dofile(const struct args *args, const char *fname, /* Main loop over file contents. */ - while (cur < sz) { + cur = 0; + for (;;) { + while (cur < sz && '\n' == buf[cur]) + cur++; + if (cur >= sz) + break; + + verb = isspace((unsigned char)buf[cur]); + /* Read until next paragraph. */ - for (i = cur + 1; i < sz; i++) - if ('\n' == buf[i] && '\n' == buf[i - 1]) { - /* Consume blank paragraphs. */ - while (i + 1 < sz && '\n' == buf[i + 1]) - i++; + + for (end = cur + 1; end + 1 < sz; end++) + if ('\n' == buf[end] && '\n' == buf[end + 1] && + !(verb && end + 2 < sz && + isspace((unsigned char)buf[end + 2]))) break; - } /* Adjust end marker for EOF. */ - end = i < sz ? i - 1 : - ('\n' == buf[sz - 1] ? sz - 1 : sz); - sup = i < sz ? end + 2 : sz; + if (end < sz && '\n' != buf[end]) + end++; + /* Process paragraph and adjust start. */ + dopar(&st, buf, cur, end); - cur = sup; + cur = end + 2; } dict_destroy(); }