=================================================================== RCS file: /cvs/texi2mdoc/util.c,v retrieving revision 1.25 retrieving revision 1.30 diff -u -p -r1.25 -r1.30 --- texi2mdoc/util.c 2015/03/03 22:37:24 1.25 +++ texi2mdoc/util.c 2015/03/11 12:51:41 1.30 @@ -1,4 +1,4 @@ -/* $Id: util.c,v 1.25 2015/03/03 22:37:24 kristaps Exp $ */ +/* $Id: util.c,v 1.30 2015/03/11 12:51:41 kristaps Exp $ */ /* * Copyright (c) 2015 Kristaps Dzonsons * @@ -31,6 +31,44 @@ #include "extern.h" /* + * Table of macros. + * These ABSOLUTELY MUST BE 2 or three characters long. + */ +static const char *const mdocs[] = { + "Ap", "Dd", "Dt", "Os", + "Sh", "Ss", "Pp", "D1", + "Dl", "Bd", "Ed", "Bl", + "El", "It", "Ad", "An", + "Ar", "Cd", "Cm", "Dv", + "Er", "Ev", "Ex", "Fa", + "Fd", "Fl", "Fn", "Ft", + "Ic", "In", "Li", "Nd", + "Nm", "Op", "Ot", "Pa", + "Rv", "St", "Va", "Vt", + "Xr", "%A", "%B", "%D", + "%I", "%J", "%N", "%O", + "%P", "%R", "%T", "%V", + "Ac", "Ao", "Aq", "At", + "Bc", "Bf", "Bo", "Bq", + "Bsx", "Bx", "Db", "Dc", + "Do", "Dq", "Ec", "Ef", + "Em", "Eo", "Fx", "Ms", + "No", "Ns", "Nx", "Ox", + "Pc", "Pf", "Po", "Pq", + "Qc", "Ql", "Qo", "Qq", + "Re", "Rs", "Sc", "So", + "Sq", "Sm", "Sx", "Sy", + "Tn", "Ux", "Xc", "Xo", + "Fo", "Fc", "Oo", "Oc", + "Bk", "Ek", "Bt", "Hf", + "Fr", "Ud", "Lb", "Lp", + "Lk", "Mt", "Brq", "Bro", + "Brc", "%C", "Es", "En", + "Dx", "%Q", "br", "sp", + "%U", "Ta", "ll", NULL, + }; + +/* * Unmap the top-most file in the stack of files currently opened (that * is, nested calls to parsefile()). */ @@ -65,7 +103,121 @@ texivaluefree(struct texivalue *p) free(p->value); } +static void +texidex_free(struct texidex *p) +{ + size_t i; + + for (i = 0; i < p->indexsz; i++) + free(p->index[i]); + + free(p->index); + free(p->name); + p->index = NULL; + p->indexsz = 0; +} + /* + * Add the text beginning at "index" and of "sz" bytes to the index + * named "tok" with name size "toksz". + * This will also output the necessary mdoc(7) to construct the index. + */ +void +texindex(struct texi *p, const char *tok, + size_t toksz, const char *index, size_t sz) +{ + size_t i; +#ifdef HAVE_INDEX + char *cp; +#endif + + if (0 == sz) { + texiwarn(p, "zero-length index entry"); + return; + } + + /* Look for the index. (Must be found.) */ + for (i = 0; i < p->indexsz; i++) { + if (strlen(p->indexs[i].name) != toksz) + continue; + if (strncmp(p->indexs[i].name, tok, toksz)) + continue; + break; + } + + assert(i < p->indexsz); + /* Reallocate index's terms. */ + p->indexs[i].index = realloc + (p->indexs[i].index, + (p->indexs[i].indexsz + 1) * + sizeof(char *)); + if (NULL == p->indexs[i].index) + texiabort(p, NULL); + + /* Add term to term array. */ + p->indexs[i].index[p->indexs[i].indexsz] = + malloc(sz + 1); + if (NULL == p->indexs[i].index[p->indexs[i].indexsz]) + texiabort(p, NULL); + memcpy(p->indexs[i].index[p->indexs[i].indexsz], + index, sz); + p->indexs[i].index[p->indexs[i].indexsz][sz] = '\0'; + + /* Output mdoc(7) for index. */ +#ifdef HAVE_INDEX + p->seenvs = -1; + teximacroopen(p, "Ix"); + texiputchars(p, "idx"); + texiputchars(p, p->indexs[i].name); + cp = p->indexs[i].index[p->indexs[i].indexsz]; + while ('\n' != *cp) { + assert('\0' != *cp); + texiputchar(p, *cp++); + } + teximacroclose(p); +#endif + p->indexs[i].indexsz++; +} + +/* + * Add an index entry named "tok" of length "sz". + * This usually consists of two letters, e.g., "cp" or "vr". + * This does nothing if the index exists or is zero-sized. + */ +void +texindex_add(struct texi *p, const char *tok, size_t sz) +{ + size_t i; + char *cp; + + if (0 == sz) + return; + + /* Make sure we don't have a duplicate. */ + for (i = 0; i < p->indexsz; i++) { + if (strlen(p->indexs[i].name) != sz) + continue; + if (strncmp(p->indexs[i].name, tok, sz)) + continue; + return; + } + + /* Reallocate indices. */ + p->indexs = realloc(p->indexs, + sizeof(struct texidex) * (p->indexsz + 1)); + if (NULL == p->indexs) + texiabort(p, NULL); + if (NULL == (cp = malloc(sz + 1))) + texiabort(p, NULL); + memcpy(cp, tok, sz); + cp[sz] = '\0'; + p->indexs[p->indexsz].name = cp; + p->indexs[p->indexsz].index = NULL; + p->indexs[p->indexsz].indexsz = 0; + p->indexsz++; +} + +/* * Unmap all files that we're currently using and free all resources * that we've allocated during the parse. * The utility should exit(...) after this is called. @@ -90,7 +242,7 @@ texiexit(struct texi *p) for (i = 0; i < p->dirsz; i++) free(p->dirs[i]); for (i = 0; i < p->indexsz; i++) - free(p->indexs[i]); + texidex_free(&p->indexs[i]); for (i = 0; i < p->valsz; i++) texivaluefree(&p->vals[i]); @@ -100,6 +252,7 @@ texiexit(struct texi *p) free(p->dirs); free(p->subtitle); free(p->title); + free(p->copying); } /* @@ -190,7 +343,6 @@ texiputchar(struct texi *p, char c) fputc(c, p->outfile); if ('\\' == c) fputc('e', p->outfile); - p->seenvs = 0; if ('\n' == c) { p->outcol = 0; p->seenws = 0; @@ -219,7 +371,6 @@ texiputchars(struct texi *p, const char *s) ((unsigned int)*s), p->outfile); else p->outcol += fputs(s, p->outfile); - p->seenvs = 0; } /* @@ -243,7 +394,7 @@ void teximacroclose(struct texi *p) { - if (p->ign) + if (p->ign || p->literal|| TEXILIST_TABLE == p->list) return; if (0 == --p->outmacro) { @@ -263,7 +414,7 @@ teximacroopen(struct texi *p, const char *s) { int rc; - if (p->ign) + if (p->ign || p->literal|| TEXILIST_TABLE == p->list) return; if (p->outcol && 0 == p->outmacro) { @@ -271,6 +422,9 @@ teximacroopen(struct texi *p, const char *s) p->outcol = 0; } + if (p->seenvs > 0 && 0 == p->outmacro) + fputs(".Pp\n", p->outfile); + if (0 == p->outmacro) fputc('.', p->outfile); else @@ -282,7 +436,7 @@ teximacroopen(struct texi *p, const char *s) fputc(' ', p->outfile); p->outcol++; p->outmacro++; - p->seenws = 0; + p->seenws = p->seenvs = 0; } /* @@ -299,9 +453,10 @@ teximacro(struct texi *p, const char *s) texierr(p, "\"%s\" in open line scope!?", s); if (p->literal) texierr(p, "\"%s\" in a literal scope!?", s); - if (p->outcol) fputc('\n', p->outfile); + if (p->seenvs > 0) + fputs(".Pp\n", p->outfile); fputc('.', p->outfile); fputs(s, p->outfile); @@ -316,10 +471,8 @@ void texivspace(struct texi *p) { - if (p->seenvs || TEXILIST_TABLE == p->list) - return; - teximacro(p, "Pp"); - p->seenvs = 1; + if (TEXILIST_TABLE != p->list && p->seenvs >= 0) + p->seenvs = 1; } /* @@ -382,7 +535,7 @@ texipunctuate(struct texi *p, size_t *pos) if (end == *pos) return; if (end + 1 == BUFSZ(p) || ' ' == BUF(p)[end] || - '\n' == BUF(p)[end]) { + '@' == BUF(p)[end] || '\n' == BUF(p)[end]) { for ( ; start < end; start++) { texiputchar(p, ' '); texiputchar(p, BUF(p)[start]); @@ -410,13 +563,9 @@ advancenext(struct texi *p, size_t *pos) while (*pos < BUFSZ(p) && ismspace(BUF(p)[*pos])) { p->seenws = 1; - /* - * If it looks like we've printed a double-line, then - * output a paragraph. - * FIXME: this is stupid. - */ - if (*pos && '\n' == BUF(p)[*pos] && '\n' == BUF(p)[*pos - 1]) - texivspace(p); + if (0 == p->seenvs && '\n' == BUF(p)[*pos]) + if (*pos + 1 < BUFSZ(p) && '\n' == BUF(p)[*pos + 1]) + p->seenvs = 1; advance(p, pos); } return(*pos); @@ -568,8 +717,23 @@ texiexecmacro(struct texi *p, struct teximacro *m, siz static void parseword(struct texi *p, size_t *pos, char extra) { + size_t i, end, len; + int c; /* + * If a prior word had a terminating double-newline, then begin + * this text block with a `Pp'. + * We don't do this if we're in a literal context (we'll print + * out the newlines themselves) nor in a `TS' table. + */ + if (p->seenvs > 0 && 0 == p->literal && TEXILIST_TABLE != p->list) { + if (p->outcol > 0) + fputc('\n', p->outfile); + fputs(".Pp\n", p->outfile); + p->outcol = 0; + } + + /* * Some line control: if we (non-macro, non-literal) already * have more than 72 characters written to the screen, then * output a newline before getting started. @@ -584,6 +748,46 @@ parseword(struct texi *p, size_t *pos, char extra) p->seenws = 0; + /* + * If we're in a macro line, we might want to print text that + * happens to be the same as an mdoc(7) macro. + * Obviously, we need to escape these words. + */ + if (p->outmacro) { + end = *pos; + /* Read ahead to get the word length. */ + while (end < BUFSZ(p) && ! ismspace(BUF(p)[end])) { + switch ((c = BUF(p)[end])) { + case ('@'): + case ('}'): + case ('{'): + break; + default: + if ('\0' != extra && extra == c) + break; + end++; + continue; + } + break; + } + len = end - *pos; + /* See if we have a match. */ + for (i = 0; NULL != mdocs[i]; i++) { + /* All macros are 2 or three letters. */ + if (len < 2 || len > 3) + continue; + /* Check the macro word length. */ + if ('\0' == mdocs[i][2] && 2 != len) + continue; + else if ('\0' == mdocs[i][3] && 3 != len) + continue; + if (strncmp(mdocs[i], &BUF(p)[*pos], len)) + continue; + texiputchars(p, "\\&"); + break; + } + } + while (*pos < BUFSZ(p) && ! ismspace(BUF(p)[*pos])) { switch (BUF(p)[*pos]) { case ('@'): @@ -593,7 +797,28 @@ parseword(struct texi *p, size_t *pos, char extra) } if ('\0' != extra && BUF(p)[*pos] == extra) return; - if (*pos < BUFSZ(p) - 1 && + + if (p->literal) { + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + continue; + } + + if ('"' == BUF(p)[*pos]) { + texiputchars(p, "\\(dq"); + } else if (*pos < BUFSZ(p) - 2 && + '-' == BUF(p)[*pos] && + '-' == BUF(p)[*pos + 1] && + '-' == BUF(p)[*pos + 2]) { + texiputchars(p, "\\(em"); + advance(p, pos); + advance(p, pos); + } else if (*pos < BUFSZ(p) - 1 && + '-' == BUF(p)[*pos] && + '-' == BUF(p)[*pos + 1]) { + texiputchars(p, "\\(en"); + advance(p, pos); + } else if (*pos < BUFSZ(p) - 1 && '`' == BUF(p)[*pos] && '`' == BUF(p)[*pos + 1]) { texiputchars(p, "\\(lq"); @@ -605,6 +830,7 @@ parseword(struct texi *p, size_t *pos, char extra) advance(p, pos); } else texiputchar(p, BUF(p)[*pos]); + advance(p, pos); } @@ -613,9 +839,18 @@ parseword(struct texi *p, size_t *pos, char extra) * period at the end of the last printed word, then open a * newline. */ - if (0 == p->literal && 0 == p->outmacro && - *pos < BUFSZ(p) && '.' == BUF(p)[*pos - 1]) - texiputchar(p, '\n'); + if (0 == p->literal && 0 == p->outmacro && *pos < BUFSZ(p)) + switch (BUF(p)[*pos - 1]) { + case ('.'): + case ('!'): + case ('?'): + texiputchar(p, '\n'); + break; + default: + break; + } + + p->seenvs = 0; } /* @@ -669,10 +904,10 @@ texicmd(const struct texi *p, size_t pos, size_t *end, /* Look for it in our indices. */ for (i = 0; i < p->indexsz; i++) { - toksz = strlen(p->indexs[i]); + toksz = strlen(p->indexs[i].name); if (len != 5 + toksz) continue; - if (strncmp(&BUF(p)[pos], p->indexs[i], toksz)) + if (strncmp(&BUF(p)[pos], p->indexs[i].name, toksz)) continue; if (0 == strncmp(&BUF(p)[pos + toksz], "index", 5)) return(TEXICMD_USER_INDEX); @@ -837,6 +1072,8 @@ parseeoln(struct texi *p, size_t *pos) texiwarn(p, "unexpected \"{\""); advance(p, pos); continue; + case ('\n'): + continue; case ('@'): break; default: @@ -859,6 +1096,18 @@ parseeoln(struct texi *p, size_t *pos) advance(p, pos); } +enum texicmd +peeklinecmd(const struct texi *p, size_t pos) +{ + size_t end; + + while (pos < BUFSZ(p) && isws(BUF(p)[pos])) + pos++; + if (pos == BUFSZ(p) || '@' != BUF(p)[pos]) + return(TEXICMD__MAX); + return(texicmd(p, pos, &end, NULL)); +} + /* * Peek to see if there's a command after subsequent whitespace. * If so, return the macro identifier. @@ -1043,6 +1292,9 @@ parseto(struct texi *p, size_t *pos, const char *endto if (NULL != texitoks[cmd].fp) (*texitoks[cmd].fp)(p, cmd, pos); } + + if (*pos == BUFSZ(p)) + texiwarn(p, "EOF expecting \"%s\" end\n", endtoken); } /* @@ -1468,6 +1720,7 @@ teximdocopen(struct texi *p, size_t *pos) t = time(NULL); strftime(date, sizeof(date), "%F", localtime(&t)); + p->seenvs = -1; teximacroopen(p, "Dd"); texiputchars(p, date); teximacroclose(p); @@ -1493,6 +1746,5 @@ teximdocopen(struct texi *p, size_t *pos) } else texiputchars(p, "Unknown description"); teximacroclose(p); - p->seenvs = 1; }