=================================================================== RCS file: /cvs/pod2mdoc/pod2mdoc.c,v retrieving revision 1.19 retrieving revision 1.23 diff -u -p -r1.19 -r1.23 --- pod2mdoc/pod2mdoc.c 2014/04/03 10:17:14 1.19 +++ pod2mdoc/pod2mdoc.c 2014/04/07 11:46:03 1.23 @@ -1,4 +1,4 @@ -/* $Id: pod2mdoc.c,v 1.19 2014/04/03 10:17:14 kristaps Exp $ */ +/* $Id: pod2mdoc.c,v 1.23 2014/04/07 11:46:03 kristaps Exp $ */ /* * Copyright (c) 2014 Kristaps Dzonsons * @@ -179,18 +179,26 @@ formatescape(const char *buf, size_t *start, size_t en static int trylink(const char *buf, size_t *start, size_t end, size_t dsz) { - size_t linkstart, realend, linkend, i, j, textsz; + size_t linkstart, realend, linkend, + i, j, textsz, stack; const char *text; /* * Scan to the start of the terminus. * This function is more or less replicated in the formatcode() * for null or index formatting codes. + * However, we're slightly different because we might have + * nested escapes we need to ignore. */ + stack = 0; for (linkstart = realend = *start; realend < end; realend++) { + if ('<' == buf[realend]) + stack++; if ('>' != buf[realend]) continue; - else if (dsz == 1) + else if (stack-- > 0) + continue; + if (dsz == 1) break; assert(realend > 0); if (' ' != buf[realend - 1]) @@ -215,8 +223,8 @@ trylink(const char *buf, size_t *start, size_t end, si if ('|' == buf[textsz] || '/' == buf[textsz]) break; - /* If we have a title, find the section. */ if (textsz < linkend && '|' == buf[textsz]) { + /* With title: set start, then end at section. */ linkstart = textsz + 1; textsz = textsz - *start; for (i = linkstart; i < linkend; i++) @@ -224,32 +232,43 @@ trylink(const char *buf, size_t *start, size_t end, si break; if (i < linkend) linkend = i; - } else { + } else if (textsz < linkend && '/' == buf[textsz]) { + /* With section: set end at section. */ + linkend = textsz; textsz = 0; - if (textsz < linkend && '/' == buf[textsz]) - linkend = textsz; - } + } else + /* No title, no section. */ + textsz = 0; *start = realend; - j = linkend - linkstart; - if (0 == j) + /* Do we have only subsection material? */ + if (0 == j && '/' == buf[linkend]) { + linkstart = linkend + 1; + linkend = dsz > 1 ? realend - 1 : realend; + if (0 == (j = linkend - linkstart)) + return(0); + printf("Sx %.*s", (int)j, &buf[linkstart]); + return(1); + } else if (0 == j) return(0); /* See if we qualify as being a link or not. */ - if ((j > 5 && 0 == memcmp("http:", &buf[linkstart], j)) || - (j > 6 && 0 == memcmp("https:", &buf[linkstart], j)) || - (j > 4 && 0 == memcmp("ftp:", &buf[linkstart], j)) || - (j > 5 && 0 == memcmp("sftp:", &buf[linkstart], j)) || - (j > 4 && 0 == memcmp("smb:", &buf[linkstart], j)) || - (j > 4 && 0 == memcmp("afs:", &buf[linkstart], j))) { - printf("Lk %.*s", (int)j, &buf[linkstart]); + if ((j > 4 && 0 == memcmp("http:", &buf[linkstart], j)) || + (j > 5 && 0 == memcmp("https:", &buf[linkstart], j)) || + (j > 3 && 0 == memcmp("ftp:", &buf[linkstart], j)) || + (j > 4 && 0 == memcmp("sftp:", &buf[linkstart], j)) || + (j > 3 && 0 == memcmp("smb:", &buf[linkstart], j)) || + (j > 3 && 0 == memcmp("afs:", &buf[linkstart], j))) { + /* Gross. */ + printf("Lk %.*s", (int)((dsz > 1 ? realend - 1 : + realend) - linkstart), &buf[linkstart]); return(1); } /* See if we qualify as a mailto. */ - if (j > 7 && 0 == memcmp("mailto:", &buf[linkstart], j)) { + if (j > 6 && 0 == memcmp("mailto:", &buf[linkstart], j)) { printf("Mt %.*s", (int)j, &buf[linkstart]); return(1); } @@ -837,10 +856,47 @@ static void verbatim(struct state *st, const char *buf, size_t start, size_t end) { int last; + size_t i; if ( ! st->parsing || st->paused) return; - +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. + * This handles any number of inclusions, but only when they + * come before the remaining parts... + */ + if (SECT_SYNOPSIS == st->sect) { + i = start; + for (i = start; i < end && ' ' == buf[i]; i++) + /* Spin. */ ; + if (i == end) + return; + /* We're an include block! */ + if (end - i > 10 && + 0 == memcmp(&buf[i], "#include <", 10)) { + start = i + 10; + while (start < end && ' ' == buf[start]) + start++; + fputs(".In ", stdout); + /* Stop til the '>' marker or we hit eoln. */ + while (start < end && + '>' != buf[start] && '\n' != buf[start]) + putchar(buf[start++]); + putchar('\n'); + if (start < end && '>' == buf[start]) + start++; + if (start < end && '\n' == buf[start]) + start++; + if (start < end) + goto again; + return; + } + } + + if (start == end) + return; puts(".Bd -literal"); for (last = ' '; start < end; start++) { /*