=================================================================== RCS file: /cvs/texi2mdoc/main.c,v retrieving revision 1.49 retrieving revision 1.66 diff -u -p -r1.49 -r1.66 --- texi2mdoc/main.c 2015/02/25 15:50:56 1.49 +++ texi2mdoc/main.c 2015/03/11 12:51:41 1.66 @@ -1,4 +1,4 @@ -/* $Id: main.c,v 1.49 2015/02/25 15:50:56 kristaps Exp $ */ +/* $Id: main.c,v 1.66 2015/03/11 12:51:41 kristaps Exp $ */ /* * Copyright (c) 2015 Kristaps Dzonsons * @@ -26,23 +26,31 @@ #include #include #include -#include #include #include "extern.h" +#define HAVE_INDEX 1 + +/* + * Texinfo can change the "meaning" of its section headings: chapter, + * section, subsection, etc., can be promoted and/or demoted to other + * levels of heading. + * Thus, we use an offset and just jump into this array. + */ #define SECTSZ 4 static const char *const sects[SECTSZ] = { - "Sh", - "Ss", - "Em", - "No", + "Sh", /* Chapters (sections) */ + "Ss", /* Sections (subsections) */ + "Em", /* Subsections (subsubsection) */ + "Sy", /* Subsubsections (...). */ }; static void doaccent(struct texi *, enum texicmd, size_t *); static void doblock(struct texi *, enum texicmd, size_t *); static void dobracket(struct texi *, enum texicmd, size_t *); static void dobye(struct texi *, enum texicmd, size_t *); +static void docopying(struct texi *, enum texicmd, size_t *); static void dodefindex(struct texi *, enum texicmd, size_t *); static void dodefn(struct texi *, enum texicmd, size_t *); static void dodisplay(struct texi *, enum texicmd, size_t *); @@ -53,22 +61,25 @@ static void doignargn(struct texi *, enum texicmd, siz static void doignblock(struct texi *, enum texicmd, size_t *); static void doignbracket(struct texi *, enum texicmd, size_t *); static void doignline(struct texi *, enum texicmd, size_t *); +static void doindex(struct texi *, enum texicmd, size_t *); static void doinline(struct texi *, enum texicmd, size_t *); static void doinclude(struct texi *, enum texicmd, size_t *); +static void doinsertcopying(struct texi *, enum texicmd, size_t *); static void doitem(struct texi *, enum texicmd, size_t *); static void doitemize(struct texi *, enum texicmd, size_t *); static void dolink(struct texi *, enum texicmd, size_t *); static void domacro(struct texi *, enum texicmd, size_t *); static void domath(struct texi *, enum texicmd, size_t *); +static void domenu(struct texi *, enum texicmd, size_t *); static void domultitable(struct texi *, enum texicmd, size_t *); +static void donode(struct texi *, enum texicmd, size_t *); +static void doprintindex(struct texi *, enum texicmd, size_t *); static void doquotation(struct texi *, enum texicmd, size_t *); static void dotable(struct texi *, enum texicmd, size_t *); static void dotop(struct texi *, enum texicmd, size_t *); static void dosecoffs(struct texi *, enum texicmd, size_t *); static void dosection(struct texi *, enum texicmd, size_t *); static void dosp(struct texi *, enum texicmd, size_t *); -static void dosubsection(struct texi *, enum texicmd, size_t *); -static void dosubsubsection(struct texi *, enum texicmd, size_t *); static void dosymbol(struct texi *, enum texicmd, size_t *); static void dotab(struct texi *, enum texicmd, size_t *); static void dotitle(struct texi *, enum texicmd, size_t *); @@ -89,13 +100,14 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { doignbracket, "anchor", 6 }, /* TEXICMD_ANCHOR */ { dosection, "appendix", 8 }, /* TEXICMD_APPENDIX */ { dosection, "appendixsec", 11 }, /* TEXICMD_APPENDIXSEC */ - { dosubsection, "appendixsubsec", 14 }, /* TEXICMD_APPENDIXSUBSEC */ - { dosubsubsection, "appendixsubsubsec", 17 }, /* TEXICMD_APPENDIXSUBSUBSEC */ + { dosection, "appendixsubsec", 14 }, /* TEXICMD_APPENDIXSUBSEC */ + { dosection, "appendixsubsubsec", 17 }, /* TEXICMD_APPENDIXSUBSUBSEC */ { doinline, "asis", 4 }, /* TEXICMD_ASIS */ { dosymbol, "*", 1 }, /* TEXICMD_ASTERISK */ { dosymbol, "@", 1 }, /* TEXICMD_AT */ { doignline, "author", 6 }, /* TEXICMD_AUTHOR */ - { doinline, "b", 1 }, /* TEXICMD_BOLD */ + { doinline, "b", 1 }, /* TEXICMD_B */ + { dosymbol, "\\", 1 }, /* TEXICMD_BACKSLASH */ { dosymbol, "!", 1 }, /* TEXICMD_BANG */ { dosymbol, "bullet", 6 }, /* TEXICMD_BULLET */ { dobye, "bye", 3 }, /* TEXICMD_BYE */ @@ -103,7 +115,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { doaccent, ",", 1 }, /* TEXICMD_CEDILLA */ { doignline, "center", 6 }, /* TEXICMD_CENTER */ { dosection, "chapter", 7 }, /* TEXICMD_CHAPTER */ - { doignline, "cindex", 6 }, /* TEXICMD_CINDEX */ + { doindex, "cindex", 6 }, /* TEXICMD_CINDEX */ { doaccent, "^", 1 }, /* TEXICMD_CIRCUMFLEX */ { doinline, "cite", 4 }, /* TEXICMD_CITE */ { dovalue, "clear", 5 }, /* TEXICMD_CLEAR */ @@ -115,7 +127,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { doignline, "c", 1 }, /* TEXICMD_COMMENT */ { doignline, "comment", 7 }, /* TEXICMD_COMMENT_LONG */ { doignline, "contents", 8 }, /* TEXICMD_CONTENTS */ - { doignblock, "copying", 7 }, /* TEXICMD_COPYING */ + { docopying, "copying", 7 }, /* TEXICMD_COPYING */ { dosymbol, "copyright", 9 }, /* TEXICMD_COPYRIGHT */ { dodefindex, "defcodeindex", 12 }, /* TEXICMD_DEFCODEINDEX */ { dodefn, "deffn", 5 }, /* TEXICMD_DEFFN */ @@ -141,7 +153,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dodefn, "defvarx", 7 }, /* TEXICMD_DEFVARX */ { dodefn, "defvr", 5 }, /* TEXICMD_DEFVR */ { dodefn, "defvrx", 6 }, /* TEXICMD_DEFVRX */ - { doignblock, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */ + { domenu, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */ { doinline, "dfn", 3 }, /* TEXICMD_DFN */ { dosymbol, "DH", 2 }, /* TEXICMD_DH */ { dosymbol, "dh", 2 }, /* TEXICMD_DHSMALL */ @@ -170,7 +182,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dosymbol, "expansion", 9 }, /* TEXICMD_EXPANSION */ { doinline, "file", 4 }, /* TEXICMD_FILE */ { doignline, "finalout", 8 }, /* TEXICMD_FINALOUT */ - { doignline, "findex", 6 }, /* TEXICMD_FINDEX */ + { doindex, "findex", 6 }, /* TEXICMD_FINDEX */ { doblock, "flushleft", 9 }, /* TEXICMD_FLUSHLEFT */ { doblock, "flushright", 10 }, /* TEXICMD_FLUSHRIGHT */ { doignline, "firstparagraphindent", 20 }, /* TEXICMD_FIRSTPARAGRAPHINDENT */ @@ -214,7 +226,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dodisplay, "indentblock", 11 }, /* TEXICMD_INDENTBLOCK */ { dolink, "indicateurl", 11 }, /* TEXICMD_INDICATEURL */ { dolink, "inforef", 7 }, /* TEXICMD_INFOREF */ - { doignline, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */ + { doinsertcopying, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */ { doitem, "item", 4 }, /* TEXICMD_ITEM */ { doitemize, "itemize", 7 }, /* TEXICMD_ITEMIZE */ { doitem, "itemx", 5 }, /* TEXICMD_ITEMX */ @@ -229,12 +241,12 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { domacro, "macro", 5 }, /* TEXICMD_MACRO */ { doaccent, "=", 1 }, /* TEXICMD_MACRON */ { domath, "math", 4 }, /* TEXICMD_MATH */ - { doignblock, "menu", 4 }, /* TEXICMD_MENU */ + { domenu, "menu", 4 }, /* TEXICMD_MENU */ { dosymbol, "minus", 5 }, /* TEXICMD_MINUS */ { domultitable, "multitable", 10 }, /* TEXICMD_MULTITABLE */ { doignline, "need", 4 }, /* TEXICMD_NEED */ { dosymbol, "\n", 1 }, /* TEXICMD_NEWLINE */ - { doignline, "node", 4 }, /* TEXICMD_NODE */ + { donode, "node", 4 }, /* TEXICMD_NODE */ { doignline, "noindent", 8 }, /* TEXICMD_NOINDENT */ { dosymbol, "O", 1 }, /* TEXICMD_O */ { dosymbol, "OE", 2 }, /* TEXICMD_OE */ @@ -249,7 +261,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dosymbol, ".", 1 }, /* TEXICMD_PERIOD */ { doignline, "pindex", 6 }, /* TEXICMD_PINDEX */ { dosymbol, "pounds", 6 }, /* TEXICMD_POUNDS */ - { doignline, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ + { doprintindex, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ { dolink, "pxref", 5 }, /* TEXICMD_PXREF */ { dosymbol, "questiondown", 12 }, /* TEXICMD_QUESTIONDOWN */ { dosymbol, "?", 1 }, /* TEXICMD_QUESTIONMARK */ @@ -269,7 +281,7 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { doaccent, "ringaccent", 10 }, /* TEXICMD_RINGACCENT */ { doinline, "samp", 4 }, /* TEXICMD_SAMP */ { doinline, "sansserif", 9 }, /* TEXICMD_SANSSERIF */ - { dobracket, "sc", 2 }, /* TEXICMD_SC */ + { doinline, "sc", 2 }, /* TEXICMD_SC */ { dosection, "section", 7 }, /* TEXICMD_SECTION */ { dovalue, "set", 3 }, /* TEXICMD_SET */ { doignline, "setchapternewpage", 17 }, /* TEXICMD_SETCHAPNEWPAGE */ @@ -290,14 +302,14 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dosymbol, "}", 1 }, /* TEXICMD_SQUIGGLE_RIGHT */ { dosymbol, "ss", 2 }, /* TEXICMD_SS */ { doinline, "strong", 6 }, /* TEXICMD_STRONG */ - { dosubsection, "subheading", 10 }, /* TEXICMD_SUBHEADING */ - { dosubsection, "subsection", 10 }, /* TEXICMD_SUBSECTION */ - { dosubsubsection, "subsubheading", 13 }, /* TEXICMD_SUBSUBHEADING */ - { dosubsubsection, "subsubsection", 13 }, /* TEXICMD_SUBSUBSECTION */ + { dosection, "subheading", 10 }, /* TEXICMD_SUBHEADING */ + { dosection, "subsection", 10 }, /* TEXICMD_SUBSECTION */ + { dosection, "subsubheading", 13 }, /* TEXICMD_SUBSUBHEADING */ + { dosection, "subsubsection", 13 }, /* TEXICMD_SUBSUBSECTION */ { doignline, "subtitle", 8 }, /* TEXICMD_SUBTITLE */ { doignline, "summarycontents", 15 }, /* TEXICMD_SUMMARYCONTENTS */ - { doignline, "synindex", 8 }, /* TEXICMD_SYNINDEX */ - { doignline, "syncodeindex", 12 }, /* TEXICMD_SYNCODEINDEX */ + { dodefindex, "synindex", 8 }, /* TEXICMD_SYNINDEX */ + { dodefindex, "syncodeindex", 12 }, /* TEXICMD_SYNCODEINDEX */ { doinline, "t", 1 }, /* TEXICMD_T */ { dotab, "tab", 3 }, /* TEXICMD_TAB */ { dosymbol, "\t", 1 }, /* TEXICMD_TABSYM */ @@ -310,19 +322,20 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dosymbol, "tie", 3 }, /* TEXICMD_TIE */ { doaccent, "tieaccent", 9 }, /* TEXICMD_TIEACCENT */ { doaccent, "~", 1 }, /* TEXICMD_TILDE */ - { doignline, "tindex", 6 }, /* TEXICMD_TINDEX */ + { doindex, "tindex", 6 }, /* TEXICMD_TINDEX */ { doignline, "title", 5 }, /* TEXICMD_TITLE */ { dobracket, "titlefont", 9 }, /* TEXICMD_TITLEFONT */ { doignblock, "titlepage", 9 }, /* TEXICMD_TITLEPAGE */ { dotop, "top", 3 }, /* TEXICMD_TOP */ + { doindex, "tpindex", 7 }, /* TEXICMD_TPINDEX */ { doaccent, "u", 1 }, /* TEXICMD_U */ { doaccent, "ubaraccent", 10 }, /* TEXICMD_UBARACCENT */ { doaccent, "udotaccent", 10 }, /* TEXICMD_UDOTACCENT */ { doaccent, "\"", 1 }, /* TEXICMD_UMLAUT */ { dosection, "unnumbered", 10 }, /* TEXICMD_UNNUMBERED */ { dosection, "unnumberedsec", 13 }, /* TEXICMD_UNNUMBEREDSEC */ - { dosubsection, "unnumberedsubsec", 16 }, /* TEXICMD_UNNUMBEREDSUBSEC */ - { dosubsubsection, "unnumberedsubsubsec", 19 }, /* TEXICMD_UNNUMBEREDSUBSUBSEC */ + { dosection, "unnumberedsubsec", 16 }, /* TEXICMD_UNNUMBEREDSUBSEC */ + { dosection, "unnumberedsubsubsec", 19 }, /* TEXICMD_UNNUMBEREDSUBSUBSEC */ { dolink, "uref", 4 }, /* TEXICMD_UREF */ { dolink, "url", 3 }, /* TEXICMD_URL */ { doignline, "", 0 }, /* TEXICMD_USER_INDEX */ @@ -332,7 +345,8 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { doverb, "verb", 4 }, /* TEXICMD_VERB */ { doverbatim, "verbatim", 8 }, /* TEXICMD_VERBATIM */ { doverbinclude, "verbatiminclude", 15 }, /* TEXICMD_VERBATIMINCLUDE */ - { doignline, "vindex", 6 }, /* TEXICMD_VINDEX */ + { doindex, "vindex", 6 }, /* TEXICMD_VINDEX */ + { doindex, "vrindex", 7 }, /* TEXICMD_VRINDEX */ { dosp, "vskip", 5 }, /* TEXICMD_VSKIP */ { dotable, "vtable", 6 }, /* TEXICMD_VTABLE */ { dobracket, "w", 1 }, /* TEXICMD_W */ @@ -342,36 +356,35 @@ static const struct texitok __texitoks[TEXICMD__MAX] = const struct texitok *const texitoks = __texitoks; +/* + * Define new indexes either by assignment or aliasing (both of these + * accept the first argument as the new index). + */ static void dodefindex(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t start, end; - char *cp; + size_t start; while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) advance(p, pos); - - start = end = *pos; - while (end < BUFSZ(p) && ! ismspace(BUF(p)[end])) - end++; - - if (start == end) { - advanceeoln(p, pos, 1); + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + if (*pos == BUFSZ(p)) { + texiwarn(p, "unexpected EOF"); return; - } else if (NULL == (cp = malloc(end - start + 1))) - texiabort(p, NULL); - - memcpy(cp, &BUF(p)[start], end - start); - cp[end - start] = '\0'; - - p->indexs = realloc(p->indexs, - sizeof(char *) * (p->indexsz + 1)); - - if (NULL == p->indexs) - texiabort(p, NULL); - p->indexs[p->indexsz++] = cp; + } + if (0 == *pos - start) + texiwarn(p, "zero-length index definition"); + else + texindex_add(p, &BUF(p)[start], *pos - start); + advance(p, pos); } +/* + * Handle both possible "define function" (type, etc.) classes: where + * we'll have a body and without one (suffixed with "x"). + */ static void dodefn(struct texi *p, enum texicmd cmd, size_t *pos) { @@ -403,8 +416,10 @@ dodefn(struct texi *p, enum texicmd cmd, size_t *pos) return; } - if (NULL != blk) - texivspace(p); + if (p->seenvs >= 0) { + teximacro(p, "Pp"); + p->seenvs = -1; + } switch (cmd) { case (TEXICMD_DEFTYPEMETHOD): @@ -432,6 +447,7 @@ dodefn(struct texi *p, enum texicmd cmd, size_t *pos) break; } + p->seenvs = 0; texiputchar(p, ':'); texiputchar(p, '\n'); @@ -498,9 +514,54 @@ dodefn(struct texi *p, enum texicmd cmd, size_t *pos) abort(); } - texivspace(p); - if (NULL != blk) + if (NULL == blk) + return; + + /* + * All "block" definitions have their block bodies indented + * unless they have the "x" form of the command following. + * E.g., + * @deffn some function + * @deffnx another + * An explanation. + * @end deffn + * With this loop, we delay opening the indented block until we + * skipped past conformant macros. + */ + for (;;) { + switch (peekcmd(p, *pos)) { + case (TEXICMD_DEFFNX): + case (TEXICMD_DEFMACX): + case (TEXICMD_DEFTPX): + case (TEXICMD_DEFTYPEFNX): + case (TEXICMD_DEFTYPEFUNX): + case (TEXICMD_DEFTYPEMETHODX): + case (TEXICMD_DEFTYPEVARX): + case (TEXICMD_DEFTYPEVRX): + case (TEXICMD_DEFUNX): + case (TEXICMD_DEFVARX): + case (TEXICMD_DEFVRX): + texivspace(p); + parseeoln(p, pos); + continue; + default: + break; + } + break; + } + + if (TEXICMD_END == peekcmd(p, *pos)) { parseto(p, pos, blk); + return; + } + + texivspace(p); + teximacro(p, "Bd -filled -offset indent -compact"); + p->seenvs = -1; + parseto(p, pos, blk); + p->seenvs = 0; + teximacro(p, "Ed"); + texivspace(p); } static void @@ -614,8 +675,8 @@ doignblock(struct texi *p, enum texicmd cmd, size_t *p * Thus, we keep track of scopes for matching "end" blocks. */ while (stack > 0 && *pos < BUFSZ(p)) { - if (stack > 10) - abort(); + if (stack > 64) + texierr(p, "run-away nested stack?"); endt = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, esz); startt = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, start, ssz); if (NULL == endt) { @@ -655,6 +716,7 @@ doinline(struct texi *p, enum texicmd cmd, size_t *pos switch (cmd) { case (TEXICMD_CODE): case (TEXICMD_KBD): + /* FIXME: quote around @samp{} */ case (TEXICMD_SAMP): case (TEXICMD_T): macro = "Li"; @@ -690,13 +752,34 @@ doinline(struct texi *p, enum texicmd cmd, size_t *pos } if (NULL == macro || p->literal || TEXILIST_TABLE == p->list) { - parsebracket(p, pos); + if (TEXICMD_SC == cmd) + p->uppercase++; + parsebracket(p, pos, 0); + if (TEXICMD_SC == cmd) + p->uppercase--; return; } + /* + * If we haven't seen any whitespace, then we don't want the + * subsequent macro to insert any whitespace. + */ + if (p->outmacro && 0 == p->seenws) { + teximacroopen(p, "Ns"); + teximacroclose(p); + } + teximacroopen(p, macro); p->seenws = 0; - parsebracket(p, pos); + if (TEXICMD_CODE == cmd) + p->literal++; + if (TEXICMD_SC == cmd) + p->uppercase++; + parsebracket(p, pos, 0); + if (TEXICMD_SC == cmd) + p->uppercase--; + if (TEXICMD_CODE == cmd) + p->literal--; texipunctuate(p, pos); teximacroclose(p); } @@ -728,7 +811,7 @@ doverb(struct texi *p, enum texicmd cmd, size_t *pos) break; advance(p, pos); } - if (*pos == BUFSZ(p) - 1) + if (*pos >= BUFSZ(p) - 1) return; texiputbuf(p, start, *pos); @@ -740,6 +823,52 @@ doverb(struct texi *p, enum texicmd cmd, size_t *pos) } static void +doinsertcopying(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + advanceeoln(p, pos, 0); + if (NULL != p->copying) + texisplice(p, p->copying, p->copyingsz, *pos); +} + +static void +docopying(struct texi *p, enum texicmd cmd, size_t *pos) +{ + const char *end, *term; + size_t endsz, endpos; + + /* We retain our starting (but not ending) newlines. */ + end = "\n@end copying\n"; + endsz = strlen(end); + advanceeoln(p, pos, 0); + if (*pos == BUFSZ(p)) { + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); + return; + } + + term = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, endsz); + if (NULL == term) { + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); + endpos = BUFSZ(p); + } else + endpos = *pos + (size_t)(term - &BUF(p)[*pos]); + + assert(endpos <= BUFSZ(p)); + assert('\n' == BUF(p)[*pos]); + advance(p, pos); + + p->copying = malloc(endpos - *pos + 1); + p->copyingsz = endpos - *pos; + memcpy(p->copying, &BUF(p)[*pos], p->copyingsz); + p->copying[endpos - *pos] = '\0'; + + while (*pos < endpos) + advance(p, pos); + if (*pos < BUFSZ(p)) + advanceto(p, pos, endpos + endsz); +} + +static void doverbatim(struct texi *p, enum texicmd cmd, size_t *pos) { const char *end, *term; @@ -750,13 +879,13 @@ doverbatim(struct texi *p, enum texicmd cmd, size_t *p endsz = strlen(end); advanceeoln(p, pos, 0); if (*pos == BUFSZ(p)) { - texiwarn(p, "unexpected end of file"); + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); return; } term = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, endsz); if (NULL == term) { - texiwarn(p, "unterminated verbatim block"); + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); endpos = BUFSZ(p); } else endpos = *pos + (size_t)(term - &BUF(p)[*pos]); @@ -764,12 +893,16 @@ doverbatim(struct texi *p, enum texicmd cmd, size_t *p assert(endpos <= BUFSZ(p)); assert('\n' == BUF(p)[*pos]); advance(p, pos); - teximacro(p, "Bd -literal -offset indent"); + texivspace(p); + teximacro(p, "Bd -literal -offset indent -compact"); + p->seenvs = -1; while (*pos < endpos) { texiputchar(p, BUF(p)[*pos]); advance(p, pos); } + p->seenvs = 0; teximacro(p, "Ed"); + texivspace(p); if (*pos < BUFSZ(p)) advanceto(p, pos, endpos + endsz); } @@ -900,41 +1033,47 @@ static void dobracket(struct texi *p, enum texicmd cmd, size_t *pos) { - parsebracket(p, pos); + parsebracket(p, pos, 0); } static void dodisplay(struct texi *p, enum texicmd cmd, size_t *pos) { + advanceeoln(p, pos, 1); + texivspace(p); + switch (cmd) { case (TEXICMD_FORMAT): case (TEXICMD_SMALLFORMAT): - teximacro(p, "Bd -filled"); + teximacro(p, "Bd -filled -compact"); break; default: - teximacro(p, "Bd -filled -offset indent"); + teximacro(p, "Bd -filled -offset indent -compact"); break; } - p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, pos, 1); + p->seenvs = -1; parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "Ed"); + texivspace(p); } static void doexample(struct texi *p, enum texicmd cmd, size_t *pos) { - teximacro(p, "Bd -literal -offset indent"); - /* FIXME: ignore and parseeoln. */ advanceeoln(p, pos, 1); + + texivspace(p); + teximacro(p, "Bd -literal -offset indent -compact"); p->literal++; parseto(p, pos, texitoks[cmd].tok); p->literal--; + p->seenvs = 0; teximacro(p, "Ed"); + texivspace(p); } static void @@ -948,20 +1087,27 @@ dobye(struct texi *p, enum texicmd cmd, size_t *pos) static void dotitle(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t start, end; + size_t start; while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) advance(p, pos); - start = end = *pos; - while (end < BUFSZ(p) && '\n' != BUF(p)[end]) - end++; - advanceeoln(p, pos, 1); + + /* We want to suck down the entire line, inclusive \n. */ + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) { + if ('@' == BUF(p)[*pos]) + advance(p, pos); + advance(p, pos); + } + if (*pos < BUFSZ(p)) + advance(p, pos); + + /* Copy this into a buffer. */ free(p->subtitle); - p->subtitle = malloc(end - start + 1); - if (NULL == p->subtitle) + if (NULL == (p->subtitle = malloc(*pos - start + 1))) texiabort(p, NULL); - memcpy(p->subtitle, &BUF(p)[start], end - start); - p->subtitle[end - start] = '\0'; + memcpy(p->subtitle, &BUF(p)[start], *pos - start); + p->subtitle[*pos - start] = '\0'; } static void @@ -1154,6 +1300,9 @@ dosymbol(struct texi *p, enum texicmd cmd, size_t *pos case (TEXICMD_AT): texiputchar(p, '@'); break; + case (TEXICMD_BACKSLASH): + texiputchar(p, '\\'); + break; case (TEXICMD_BANG): texiputchar(p, '!'); break; @@ -1325,41 +1474,259 @@ doquotation(struct texi *p, enum texicmd cmd, size_t * teximacro(p, "Qc"); } +static int +indexcmp(const void *p1, const void *p2) +{ + + return(strcasecmp(*(const char **)p1, *(const char **)p2)); +} + static void -domath(struct texi *p, enum texicmd cmd, size_t *pos) +doprintindex(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t nest, start; + size_t i, j, start, end, len; +#if HAVE_INDEX + char *cp; +#endif - /* - * Math handling is different from everything else. - * We don't allow any subcomponents, and we ignore the rules in - * terms of @-commands. - * This departs from GNU's rules, but whatever. - */ while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) advance(p, pos); - if (*pos == BUFSZ(p) || '{' != BUF(p)[*pos]) + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + if ((end = *pos) == BUFSZ(p)) { + texiwarn(p, "unexpected EOF"); return; + } + advance(p, pos); - if (p->seenws && p->outcol && 0 == p->literal) - texiputchar(p, ' '); - p->seenws = 0; - for (nest = 1, start = *pos; *pos < BUFSZ(p) && nest > 0; ) { - if ('{' == BUF(p)[*pos]) - nest++; - else if ('}' == BUF(p)[*pos]) - if (0 == --nest) - continue; + if (0 == (len = end - start)) { + texiwarn(p, "zero-length index"); + return; + } + + /* Look for the index in our table. */ + for (i = 0; i < p->indexsz; i++) { + if (strlen(p->indexs[i].name) != len) + continue; + if (strncmp(p->indexs[i].name, &BUF(p)[start], len)) + continue; + break; + } + + if (i == p->indexsz) { + texiwarn(p, "cannot find index"); + return; + } else if (0 == p->indexs[i].indexsz) + return; + + /* Alphabetically sort our indices. */ + qsort(p->indexs[i].index, + p->indexs[i].indexsz, sizeof(char *), indexcmp); + + texivspace(p); + teximacro(p, "Bl -tag -width Ds -compact"); + for (j = 0; j < p->indexs[i].indexsz; j++) { + teximacroopen(p, "It"); +#if HAVE_INDEX + teximacroopen(p, "Lkx"); + texiputchars(p, "\"idx"); + texiputchars(p, p->indexs[i].name); + cp = p->indexs[i].index[j]; + while ('\n' != *cp) { + assert('\0' != *cp); + texiputchar(p, *cp++); + } + texiputchars(p, "\" \""); + p->literal++; +#endif + texisplice(p, p->indexs[i].index[j], + strlen(p->indexs[i].index[j]), *pos); + parseeoln(p, pos); +#if HAVE_INDEX + p->literal--; + texiputchars(p, "\""); + teximacroclose(p); +#endif + teximacroclose(p); + } + p->seenvs = 0; + teximacro(p, "El"); + texivspace(p); +} + +static void +donode(struct texi *p, enum texicmd cmd, size_t *pos) +{ + int sv = p->seenvs; + + if (NULL != p->chapters) { + advanceeoln(p, pos, 1); + return; + } + +#if HAVE_INDEX + p->seenvs = -1; + teximacroopen(p, "Ix"); + texiputchars(p, "node"); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) advance(p, pos); + while (*pos < BUFSZ(p)) { + if (BUF(p)[*pos] == ',') + break; + else if (BUF(p)[*pos] == '\n') + break; + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); } - if (*pos == BUFSZ(p)) + teximacroclose(p); +#endif + advanceeoln(p, pos, 1); + p->seenvs = sv; +} + +/* + * This handles both menu and detailedmenu. + * The syntax of these is fairly... unspecific, but what we do here + * seems to work with most manuals. + */ +static void +domenu(struct texi *p, enum texicmd cmd, size_t *pos) +{ + size_t start, end, sv; + + if (NULL != p->chapters) { + doignblock(p, cmd, pos); return; - assert('}' == BUF(p)[*pos]); - texiputbuf(p, start, *pos); - advance(p, pos); + } + + advanceeoln(p, pos, 1); + + /* + * Parse past initial stuff. + * TODO: the manual says we're supposed to make this in bold or + * something. + */ + while (*pos < BUFSZ(p)) { + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + if ('*' != BUF(p)[*pos]) { + if (TEXICMD_END == peeklinecmd(p, *pos)) + break; + parseeoln(p, pos); + } else + break; + } + + texivspace(p); + teximacro(p, "Bl -tag -width Ds -compact"); + while (*pos < BUFSZ(p)) { + /* + * Read to next menu item. + * We simply parse every line until we get a magic '*'. + * These lines might occur interspersed OR as the + * description of an entry. + * Either way it's in the `It' block. + */ + if (0 == p->seenws) + p->seenws = *pos < BUFSZ(p) && isws(BUF(p)[*pos]); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + if ('*' != BUF(p)[*pos]) { + if (TEXICMD_END == peeklinecmd(p, *pos)) + break; + parseeoln(p, pos); + continue; + } + + /* Now we're parsing a menu item. */ + advance(p, pos); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + start = sv = *pos; + while (*pos < BUFSZ(p) && ':' != BUF(p)[*pos]) + advance(p, pos); + end = *pos; + advance(p, pos); + + if (*pos == BUFSZ(p)) { + texiwarn(p, "bad menu syntax"); + break; + } else if (':' != BUF(p)[*pos]) { + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + teximacroopen(p, "It"); +#ifdef HAVE_INDEX + teximacroopen(p, "Lkx"); + texiputchars(p, "\"node"); +#endif + while (*pos < BUFSZ(p)) { + switch (BUF(p)[*pos]) { + case ('\t'): + case ('\n'): + case (','): + break; + case ('.'): + if (*pos + 1 == BUFSZ(p)) { + advance(p, pos); + continue; + } + if (' ' == BUF(p)[*pos + 1]) { + advance(p, pos); + break; + } + /* FALLTHROUGH */ + default: + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + continue; + } + advance(p, pos); + break; + } +#ifdef HAVE_INDEX + texiputchars(p, "\" \""); + for (start = sv; start < end; start++) + texiputchar(p, BUF(p)[start]); + texiputchar(p, '"'); +#endif + } else { + advance(p, pos); + p->seenvs = 0; + teximacroopen(p, "It"); +#ifdef HAVE_INDEX + teximacroopen(p, "Lkx"); + texiputchars(p, "\"node"); + for (start = sv; start < end; start++) + texiputchar(p, BUF(p)[start]); + texiputchars(p, "\" \""); +#endif + for (start = sv; start < end; start++) + texiputchar(p, BUF(p)[start]); +#ifdef HAVE_INDEX + texiputchar(p, '"'); +#endif + } +#ifdef HAVE_INDEX + teximacroclose(p); +#endif + teximacroclose(p); + } + + advanceeoln(p, pos, 0); + p->seenvs = 0; + teximacro(p, "El"); + texivspace(p); } static void +domath(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + parsebracket(p, pos, 1); +} + +static void dovalue(struct texi *p, enum texicmd cmd, size_t *pos) { size_t start, end; @@ -1402,7 +1769,7 @@ dovalue(struct texi *p, enum texicmd cmd, size_t *pos) texiputchar(p, ' '); p->seenws = 0; if (NULL != (cp = valueblookup(p, pos))) - texisplice(p, cp, strlen(cp), pos); + texisplice(p, cp, strlen(cp), *pos); else texiputchars(p, "{No value}"); } else if (TEXICMD_IFCLEAR == cmd) { @@ -1492,43 +1859,6 @@ sectioner(struct texi *p, int sec) } static void -dosubsubsection(struct texi *p, enum texicmd cmd, size_t *pos) -{ - int sec; - - sec = sectioner(p, 3); - - /* We don't have a subsubsubsection, so make one up. */ - texivspace(p); - teximacroopen(p, sects[sec]); - parseeoln(p, pos); - teximacroclose(p); - texivspace(p); -} - -static void -dosubsection(struct texi *p, enum texicmd cmd, size_t *pos) -{ - int sec; - - sec = sectioner(p, 2); - - if (p->outmacro) - texierr(p, "\"%s\" in open line scope!?", sects[sec]); - else if (p->literal) - texierr(p, "\"%s\" in a literal scope!?", sects[sec]); - - /* We don't have a subsubsection, so make one up. */ - if (sec > 1) - texivspace(p); - teximacroopen(p, sects[sec]); - parseeoln(p, pos); - teximacroclose(p); - if (sec > 1) - texivspace(p); -} - -static void dosecoffs(struct texi *p, enum texicmd cmd, size_t *pos) { @@ -1544,9 +1874,11 @@ dosection(struct texi *p, enum texicmd cmd, size_t *po int sec; switch (cmd) { + case (TEXICMD_TOP): + sec = 0; + break; case (TEXICMD_APPENDIX): case (TEXICMD_CHAPTER): - case (TEXICMD_TOP): case (TEXICMD_UNNUMBERED): sec = sectioner(p, 0); break; @@ -1556,6 +1888,18 @@ dosection(struct texi *p, enum texicmd cmd, size_t *po case (TEXICMD_UNNUMBEREDSEC): sec = sectioner(p, 1); break; + case (TEXICMD_APPENDIXSUBSEC): + case (TEXICMD_SUBHEADING): + case (TEXICMD_SUBSECTION): + case (TEXICMD_UNNUMBEREDSUBSEC): + sec = sectioner(p, 2); + break; + case (TEXICMD_APPENDIXSUBSUBSEC): + case (TEXICMD_SUBSUBHEADING): + case (TEXICMD_SUBSUBSECTION): + case (TEXICMD_UNNUMBEREDSUBSUBSEC): + sec = sectioner(p, 3); + break; default: abort(); } @@ -1565,67 +1909,47 @@ dosection(struct texi *p, enum texicmd cmd, size_t *po else if (p->literal) texierr(p, "\"%s\" in a literal scope!?", sects[sec]); + if (sec < 2) + p->seenvs = -1; + else + texivspace(p); + + if (0 == sec && NULL != p->chapters) { + teximdocclose(p, 0); + teximdocopen(p, pos); + } + teximacroopen(p, sects[sec]); parseeoln(p, pos); teximacroclose(p); - p->seenvs = 1; -} -static void -dosp(struct texi *p, enum texicmd cmd, size_t *pos) -{ - - if (p->literal) - texiputchar(p, '\n'); + if (sec < 2) + p->seenvs = -1; else texivspace(p); - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, pos, 1); } static void dotop(struct texi *p, enum texicmd cmd, size_t *pos) { - const char *cp; - time_t t; - char date[32]; if (--p->ign) texierr(p, "@top command while ignoring"); - /* - * Here we print our standard mdoc(7) prologue. - * We use the title set with @settitle for the `Nd' description - * and the source document filename (the first one as invoked on - * the command line) for the title. - * The date is set to the current date. - */ - t = time(NULL); - strftime(date, sizeof(date), "%F", localtime(&t)); + if (NULL == p->chapters) + teximdocopen(p, pos); + dosection(p, cmd, pos); +} - teximacroopen(p, "Dd"); - texiputchars(p, date); - teximacroclose(p); - teximacroopen(p, "Dt"); - for (cp = p->title; '\0' != *cp; cp++) - texiputchar(p, toupper((unsigned int)*cp)); - texiputchars(p, " 7"); - teximacroclose(p); - teximacro(p, "Os"); - teximacro(p, "Sh NAME"); - teximacroopen(p, "Nm"); - for (cp = p->title; '\0' != *cp; cp++) - texiputchar(p, *cp); - teximacroclose(p); - teximacroopen(p, "Nd"); - if (NULL != p->subtitle) - for (cp = p->subtitle; '\0' != *cp; cp++) - texiputchar(p, *cp); +static void +dosp(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + advanceeoln(p, pos, 1); + if (p->literal) + texiputchar(p, '\n'); else - texiputchars(p, "Unknown description"); - teximacroclose(p); - p->seenvs = 1; - dosection(p, cmd, pos); + texivspace(p); } static void @@ -1634,7 +1958,8 @@ doitem(struct texi *p, enum texicmd cmd, size_t *pos) /* Multitable is using raw tbl(7). */ if (TEXILIST_TABLE == p->list) { - texiputchar(p, '\n'); + if (p->outcol > 0) + texiputchar(p, '\n'); return; } @@ -1645,9 +1970,11 @@ doitem(struct texi *p, enum texicmd cmd, size_t *pos) switch (p->list) { case (TEXILIST_ITEM): + p->seenvs = -1; teximacroopen(p, "It"); break; case (TEXILIST_NOITEM): + p->seenvs = -1; teximacro(p, "It"); break; default: @@ -1656,7 +1983,6 @@ doitem(struct texi *p, enum texicmd cmd, size_t *pos) } /* Trick so we don't start with Pp. */ - p->seenvs = 1; parseeoln(p, pos); if (TEXILIST_ITEM == p->list) @@ -1670,7 +1996,7 @@ dotab(struct texi *p, enum texicmd cmd, size_t *pos) { /* This command is only useful in @multitable. */ - if (TEXILIST_TABLE == p->list) + if (TEXILIST_TABLE == p->list && p->outcol) texiputchar(p, '\t'); } @@ -1682,6 +2008,7 @@ domultitable(struct texi *p, enum texicmd cmd, size_t enum texicmd type; size_t i, end, columns; + texivspace(p); p->list = TEXILIST_TABLE; /* * TS/TE blocks aren't "in mdoc(7)", so we can disregard the @@ -1697,7 +2024,7 @@ domultitable(struct texi *p, enum texicmd cmd, size_t /* Make sure we don't print anything when scanning. */ p->ign++; - if ('@' == BUF(p)[*pos]) { + if (*pos < BUFSZ(p) && '@' == BUF(p)[*pos]) { /* * Look for @columnfractions. * We ignore these, but we do use the number of @@ -1735,13 +2062,16 @@ domultitable(struct texi *p, enum texicmd cmd, size_t texiputchar(p, ' '); texiputchar(p, 'l'); } - texiputchars(p, ".\n"); + + texiputchar(p, '.'); + texiputchar(p, '\n'); p->outmacro++; parseto(p, pos, texitoks[cmd].tok); p->outmacro--; teximacro(p, "TE"); p->literal = svliteral; p->list = sv; + texivspace(p); } static void @@ -1749,13 +2079,15 @@ dotable(struct texi *p, enum texicmd cmd, size_t *pos) { enum texilist sv = p->list; - p->list = TEXILIST_ITEM; - teximacro(p, "Bl -tag -width Ds"); - /* FIXME: ignore and parseeoln. */ advanceeoln(p, pos, 1); - p->seenvs = 1; + + p->list = TEXILIST_ITEM; + texivspace(p); + teximacro(p, "Bl -tag -width Ds -compact"); parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "El"); + texivspace(p); p->list = sv; } @@ -1770,7 +2102,8 @@ doend(struct texi *p, enum texicmd cmd, size_t *pos) while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) advance(p, pos); - texiwarn(p, "unexpected \"end\": %.*s", (int)(*pos - start), &BUF(p)[start]); + texiwarn(p, "unexpected \"end\": %.*s", + (int)(*pos - start), &BUF(p)[start]); advanceeoln(p, pos, 1); } @@ -1779,13 +2112,15 @@ doenumerate(struct texi *p, enum texicmd cmd, size_t * { enum texilist sv = p->list; - p->list = TEXILIST_NOITEM; - teximacro(p, "Bl -enum"); - p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ advanceeoln(p, pos, 1); - parseto(p, pos, "enumerate"); + + p->list = TEXILIST_NOITEM; + texivspace(p); + teximacro(p, "Bl -enum -compact"); + parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "El"); + texivspace(p); p->list = sv; } @@ -1794,13 +2129,15 @@ doitemize(struct texi *p, enum texicmd cmd, size_t *po { enum texilist sv = p->list; - p->list = TEXILIST_NOITEM; - teximacro(p, "Bl -bullet"); - p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ advanceeoln(p, pos, 1); - parseto(p, pos, "itemize"); + + p->list = TEXILIST_NOITEM; + texivspace(p); + teximacro(p, "Bl -bullet -compact"); + parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "El"); + texivspace(p); p->list = sv; } @@ -1809,16 +2146,65 @@ doignbracket(struct texi *p, enum texicmd cmd, size_t { p->ign++; - parsebracket(p, pos); + parsebracket(p, pos, 0); p->ign--; } static void +doindex(struct texi *p, enum texicmd cmd, size_t *pos) +{ + size_t start, end, len; + + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + + if (*pos == BUFSZ(p)) { + texiwarn(p, "unexpected EOF"); + return; + } + + advance(p, pos); + end = *pos; + if (0 == (len = end - start)) { + texiwarn(p, "zero-length index"); + return; + } + + /* Two-letter combos we can look up verbatim. */ + if (7 == texitoks[cmd].len) { + texindex(p, texitoks[cmd].tok, 2, &BUF(p)[start], len); + return; + } + + assert(6 == texitoks[cmd].len); + /* Newer one-letter combos need to be mapped. */ + switch (texitoks[cmd].tok[0]) { + case ('c'): + texindex(p, "cp", 2, &BUF(p)[start], len); + break; + case ('v'): + texindex(p, "vr", 2, &BUF(p)[start], len); + break; + case ('f'): + texindex(p, "fn", 2, &BUF(p)[start], len); + break; + case ('t'): + texindex(p, "tp", 2, &BUF(p)[start], len); + break; + default: + abort(); + } +} + +static void doignline(struct texi *p, enum texicmd cmd, size_t *pos) { - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, pos, 1); + advanceeoln(p, pos, 1); } /* @@ -1876,10 +2262,15 @@ main(int argc, char *argv[]) memset(&texi, 0, sizeof(struct texi)); texi.ign = 1; + texi.outfile = stdout; + texi.seenvs = -1; Idir = NULL; - while (-1 != (c = getopt(argc, argv, "I:"))) + while (-1 != (c = getopt(argc, argv, "C:I:"))) switch (c) { + case ('C'): + texi.chapters = optarg; + break; case ('I'): Idir = optarg; break; @@ -1890,6 +2281,12 @@ main(int argc, char *argv[]) argv += optind; argc -= optind; + /* Add the default Texinfo indices. */ + texindex_add(&texi, "cp", 2); + texindex_add(&texi, "vr", 2); + texindex_add(&texi, "tp", 2); + texindex_add(&texi, "fn", 2); + if (argc > 0) { if (NULL == (dirpath = strdup(argv[0]))) texiabort(&texi, NULL); @@ -1913,8 +2310,8 @@ main(int argc, char *argv[]) } texiexit(&texi); - return(EXIT_FAILURE); + exit(EXIT_SUCCESS); usage: - fprintf(stderr, "usage: %s [-Idirs] [file]\n", progname); + fprintf(stderr, "usage: %s [-Cdir] [-Idirs] [file]\n", progname); return(EXIT_FAILURE); }