static void text_heading(textfile *tf, word *tprefix, word *nprefix, word *text, alignstruct align, int indent, int width, textconfig *cfg) { rdstring t = { 0, 0, NULL }; int margin, length; int firstlinewidth, wrapwidth; wrappedline *wrapping, *p; if (align.just_numbers && nprefix) { text_rdaddw(&t, nprefix, NULL, cfg); rdadds(&t, align.number_suffix); } else if (!align.just_numbers && tprefix) { text_rdaddw(&t, tprefix, NULL, cfg); rdadds(&t, align.number_suffix); } margin = length = ustrwid(t.text ? t.text : L"", cfg->charset); if (align.align == LEFTPLUS) { margin = indent - margin; if (margin < 0) margin = 0; firstlinewidth = indent + width - margin - length; wrapwidth = width; } else if (align.align == LEFT || align.align == CENTRE) { margin = 0; firstlinewidth = indent + width - length; wrapwidth = indent + width; } wrapping = wrap_para(text, firstlinewidth, wrapwidth, text_width, cfg, 0); for (p = wrapping; p; p = p->next) { text_rdaddw(&t, p->begin, p->end, cfg); length = ustrwid(t.text ? t.text : L"", cfg->charset); if (align.align == CENTRE) { margin = (indent + width - length)/2; if (margin < 0) margin = 0; } text_output_many(tf, margin, L' '); text_output(tf, t.text); text_output(tf, L"\n"); if (*align.underline) { text_output_many(tf, margin, L' '); while (length > 0) { text_output(tf, align.underline); length -= ustrwid(align.underline, cfg->charset); } text_output(tf, L"\n"); } if (align.align == LEFTPLUS) margin = indent; else margin = 0; sfree(t.text); t = empty_rdstring; } wrap_free(wrapping); text_output(tf, L"\n"); sfree(t.text); }
static void text_rdaddw(rdstring *rs, word *text, word *end, textconfig *cfg) { for (; text && text != end; text = text->next) switch (text->type) { case word_HyperLink: case word_HyperEnd: case word_UpperXref: case word_LowerXref: case word_XrefEnd: case word_IndexRef: break; case word_Normal: case word_Emph: case word_Code: case word_WeakCode: case word_WhiteSpace: case word_EmphSpace: case word_CodeSpace: case word_WkCodeSpace: case word_Quote: case word_EmphQuote: case word_CodeQuote: case word_WkCodeQuote: assert(text->type != word_CodeQuote && text->type != word_WkCodeQuote); if (towordstyle(text->type) == word_Emph && (attraux(text->aux) == attr_First || attraux(text->aux) == attr_Only)) rdadds(rs, cfg->startemph); else if (towordstyle(text->type) == word_Code && (attraux(text->aux) == attr_First || attraux(text->aux) == attr_Only)) rdadds(rs, cfg->lquote); if (removeattr(text->type) == word_Normal) { if (cvt_ok(cfg->charset, text->text) || !text->alt) rdadds(rs, text->text); else text_rdaddw(rs, text->alt, NULL, cfg); } else if (removeattr(text->type) == word_WhiteSpace) { rdadd(rs, L' '); } else if (removeattr(text->type) == word_Quote) { rdadds(rs, quoteaux(text->aux) == quote_Open ? cfg->lquote : cfg->rquote); } if (towordstyle(text->type) == word_Emph && (attraux(text->aux) == attr_Last || attraux(text->aux) == attr_Only)) rdadds(rs, cfg->endemph); else if (towordstyle(text->type) == word_Code && (attraux(text->aux) == attr_Last || attraux(text->aux) == attr_Only)) rdadds(rs, cfg->rquote); break; } }
static void text_para(textfile *tf, word *prefix, wchar_t *prefixextra, word *text, int indent, int extraindent, int width, textconfig *cfg) { wrappedline *wrapping, *p; rdstring pfx = { 0, 0, NULL }; int e; int firstlinewidth = width; if (prefix) { text_rdaddw(&pfx, prefix, NULL, cfg); if (prefixextra) rdadds(&pfx, prefixextra); text_output_many(tf, indent, L' '); text_output(tf, pfx.text); /* If the prefix is too long, shorten the first line to fit. */ e = extraindent - ustrwid(pfx.text ? pfx.text : L"", cfg->charset); if (e < 0) { firstlinewidth += e; /* this decreases it, since e < 0 */ if (firstlinewidth < 0) { e = indent + extraindent; firstlinewidth = width; text_output(tf, L"\n"); } else e = 0; } sfree(pfx.text); } else e = indent + extraindent; wrapping = wrap_para(text, firstlinewidth, width, text_width, cfg, 0); for (p = wrapping; p; p = p->next) { rdstring t = { 0, 0, NULL }; text_rdaddw(&t, p->begin, p->end, cfg); text_output_many(tf, e, L' '); text_output(tf, t.text); text_output(tf, L"\n"); e = indent + extraindent; sfree(t.text); } wrap_free(wrapping); text_output(tf, L"\n"); }
/* * Reads a single file (ie until get() returns EOF) */ static void read_file(paragraph *** ret, input * in, indexdata * idx) { token t; paragraph par; word wd, **whptr, **idximplicit; tree234 *macros; wchar_t utext[2], *wdtext; int style, spcstyle; int already; int iswhite, seenwhite; int type; struct stack_item { enum { stack_nop = 0, /* do nothing (for error recovery) */ stack_ualt = 1, /* \u alternative */ stack_style = 2, /* \e, \c, \cw */ stack_idx = 4, /* \I, \i, \ii */ stack_hyper = 8, /* \W */ stack_quote = 16, /* \q */ } type; word **whptr; /* to restore from \u alternatives */ word **idximplicit; /* to restore from \u alternatives */ } *sitem; stack parsestk; word *indexword=NULL, *uword=NULL, *iword=NULL; word *idxwordlist; rdstring indexstr; int index_downcase=0, index_visible=0, indexing=0; const rdstring nullrs = { 0, 0, NULL }; wchar_t uchr; t.text = NULL; macros = newtree234(macrocmp); already = FALSE; /* * Loop on each paragraph. */ while (1) { int start_cmd = c__invalid; par.words = NULL; par.keyword = NULL; whptr = &par.words; /* * Get a token. */ if (!already) { dtor(t), t = get_token(in); } already = FALSE; if (t.type == tok_eof) break; /* * Parse code paragraphs separately. */ if (t.type == tok_cmd && t.cmd == c_c && !isbrace(in)) { par.type = para_Code; par.fpos = t.pos; while (1) { dtor(t), t = get_codepar_token(in); wd.type = word_WeakCode; wd.breaks = FALSE; /* shouldn't need this... */ wd.text = ustrdup(t.text); wd.alt = NULL; wd.fpos = t.pos; addword(wd, &whptr); dtor(t), t = get_token(in); if (t.type == tok_white) { /* * The newline after a code-paragraph line */ dtor(t), t = get_token(in); } if (t.type == tok_eop || t.type == tok_eof) break; else if (t.type != tok_cmd || t.cmd != c_c) { error(err_brokencodepara, &t.pos); addpara(par, ret); while (t.type != tok_eop) /* error recovery: */ dtor(t), t = get_token(in); /* eat rest of paragraph */ goto codeparabroken; /* ick, but such is life */ } } addpara(par, ret); codeparabroken: continue; } while (t.type == tok_cmd && macrolookup(macros, in, t.text, &t.pos)) { dtor(t), t = get_token(in); } /* * This token begins a paragraph. See if it's one of the * special commands that define a paragraph type. * * (note that \# is special in a way, and \nocite takes no * text) */ par.type = para_Normal; if (t.type == tok_cmd) { int needkw=0; int is_macro = FALSE; par.fpos = t.pos; switch (t.cmd) { default: needkw = -1; break; case c__invalid: error(err_badparatype, t.text, &t.pos); needkw = 4; break; case c__comment: if (isbrace(in)) break; /* `\#{': isn't a comment para */ do { dtor(t), t = get_token(in); } while (t.type != tok_eop && t.type != tok_eof); continue; /* next paragraph */ /* * `needkw' values: * * 1 -- exactly one keyword * 2 -- at least one keyword * 4 -- any number of keywords including zero * 8 -- at least one keyword and then nothing else * 16 -- nothing at all! no keywords, no body * 32 -- no keywords at all */ case c_A: needkw = 2; par.type = para_Appendix; break; case c_B: needkw = 2; par.type = para_Biblio; break; case c_BR: needkw = 1; par.type = para_BR; start_cmd = c_BR; break; case c_C: needkw = 2; par.type = para_Chapter; break; case c_H: needkw = 2; par.type = para_Heading; par.aux = 0; break; case c_IM: needkw = 2; par.type = para_IM; start_cmd = c_IM; break; case c_S: needkw = 2; par.type = para_Subsect; par.aux = t.aux; break; case c_U: needkw = 32; par.type = para_UnnumberedChapter; break; /* For \b and \n the keyword is optional */ case c_b: needkw = 4; par.type = para_Bullet; break; case c_n: needkw = 4; par.type = para_NumberedList; break; case c_cfg: needkw = 8; par.type = para_Config; start_cmd = c_cfg; break; case c_copyright: needkw = 32; par.type = para_Copyright; break; case c_define: is_macro = TRUE; needkw = 1; break; /* For \nocite the keyword is _everything_ */ case c_nocite: needkw = 8; par.type = para_NoCite; break; case c_preamble: needkw = 32; par.type = para_Preamble; break; case c_rule: needkw = 16; par.type = para_Rule; break; case c_title: needkw = 32; par.type = para_Title; break; case c_versionid: needkw = 32; par.type = para_VersionID; break; } if (needkw > 0) { rdstring rs = { 0, 0, NULL }; int nkeys = 0; filepos fp; /* Get keywords. */ dtor(t), t = get_token(in); fp = t.pos; while (t.type == tok_lbrace) { /* This is a keyword. */ nkeys++; /* FIXME: there will be bugs if anyone specifies an * empty keyword (\foo{}), so trap this case. */ while (dtor(t), t = get_token(in), t.type == tok_word || t.type == tok_white || (t.type == tok_cmd && t.cmd == c__nbsp) || (t.type == tok_cmd && t.cmd == c__escaped)) { if (t.type == tok_white || (t.type == tok_cmd && t.cmd == c__nbsp)) rdadd(&rs, ' '); else rdadds(&rs, t.text); } if (t.type != tok_rbrace) { error(err_kwunclosed, &t.pos); continue; } rdadd(&rs, 0); /* add string terminator */ dtor(t), t = get_token(in); /* eat right brace */ } rdadd(&rs, 0); /* add string terminator */ /* See whether we have the right number of keywords. */ if ((needkw & 48) && nkeys > 0) error(err_kwillegal, &fp); if ((needkw & 11) && nkeys == 0) error(err_kwexpected, &fp); if ((needkw & 5) && nkeys > 1) error(err_kwtoomany, &fp); if (is_macro) { /* * Macro definition. Get the rest of the line * as a code-paragraph token, repeatedly until * there's nothing more left of it. Separate * with newlines. */ rdstring macrotext = { 0, 0, NULL }; while (1) { dtor(t), t = get_codepar_token(in); if (macrotext.pos > 0) rdadd(¯otext, L'\n'); rdadds(¯otext, t.text); dtor(t), t = get_token(in); if (t.type == tok_eop) break; } macrodef(macros, rs.text, macrotext.text, fp); continue; /* next paragraph */ } par.keyword = rdtrim(&rs); /* Move to EOP in case of needkw==8 or 16 (no body) */ if (needkw & 24) { /* We allow whitespace even when we expect no para body */ while (t.type == tok_white) dtor(t), t = get_token(in); if (t.type != tok_eop && t.type != tok_eof && (start_cmd == c__invalid || t.type != tok_cmd || t.cmd != start_cmd)) { error(err_bodyillegal, &t.pos); /* Error recovery: eat the rest of the paragraph */ while (t.type != tok_eop && t.type != tok_eof && (start_cmd == c__invalid || t.type != tok_cmd || t.cmd != start_cmd)) dtor(t), t = get_token(in); } if (t.type == tok_cmd) already = TRUE; /* inhibit get_token at top of loop */ addpara(par, ret); continue; /* next paragraph */ } } } /* * Now read the actual paragraph, word by word, adding to * the paragraph list. * * Mid-paragraph commands: * * \K \k * \c \cw * \e * \i \ii * \I * \u * \W * \date * \\ \{ \} */ parsestk = stk_new(); style = word_Normal; spcstyle = word_WhiteSpace; indexing = FALSE; seenwhite = TRUE; while (t.type != tok_eop && t.type != tok_eof) { iswhite = FALSE; already = FALSE; /* Handle implicit paragraph breaks after \IM, \BR etc */ if (start_cmd != c__invalid && t.type == tok_cmd && t.cmd == start_cmd) { already = TRUE; /* inhibit get_token at top of loop */ break; } if (t.type == tok_cmd && t.cmd == c__escaped) { t.type = tok_word; /* nice and simple */ t.aux = 0; /* even if `\-' - nonbreaking! */ } if (t.type == tok_cmd && t.cmd == c__nbsp) { t.type = tok_word; /* nice and simple */ sfree(t.text); t.text = ustrdup(L" "); /* text is ` ' not `_' */ t.aux = 0; /* (nonbreaking) */ } switch (t.type) { case tok_white: if (whptr == &par.words) break; /* strip whitespace at start of para */ wd.text = NULL; wd.type = spcstyle; wd.alt = NULL; wd.aux = 0; wd.fpos = t.pos; wd.breaks = FALSE; /* * Inhibit use of whitespace if it's (probably the * newline) before a repeat \IM / \BR type * directive. */ if (start_cmd != c__invalid) { dtor(t), t = get_token(in); already = TRUE; if (t.type == tok_cmd && t.cmd == start_cmd) break; } if (indexing) rdadd(&indexstr, ' '); if (!indexing || index_visible) addword(wd, &whptr); if (indexing) addword(wd, &idximplicit); iswhite = TRUE; break; case tok_word: if (indexing) rdadds(&indexstr, t.text); wd.type = style; wd.alt = NULL; wd.aux = 0; wd.fpos = t.pos; wd.breaks = t.aux; if (!indexing || index_visible) { wd.text = ustrdup(t.text); addword(wd, &whptr); } if (indexing) { wd.text = ustrdup(t.text); addword(wd, &idximplicit); } break; case tok_lbrace: error(err_unexbrace, &t.pos); /* Error recovery: push nop */ sitem = mknew(struct stack_item); sitem->type = stack_nop; stk_push(parsestk, sitem); break; case tok_rbrace: sitem = stk_pop(parsestk); if (!sitem) error(err_unexbrace, &t.pos); else { if (sitem->type & stack_ualt) { whptr = sitem->whptr; idximplicit = sitem->idximplicit; } if (sitem->type & stack_style) { style = word_Normal; spcstyle = word_WhiteSpace; } if (sitem->type & stack_idx ) { indexword->text = ustrdup(indexstr.text); if (index_downcase) ustrlow(indexword->text); indexing = FALSE; rdadd(&indexstr, L'\0'); index_merge(idx, FALSE, indexstr.text, idxwordlist); sfree(indexstr.text); } if (sitem->type & stack_hyper) { wd.text = NULL; wd.type = word_HyperEnd; wd.alt = NULL; wd.aux = 0; wd.fpos = t.pos; wd.breaks = FALSE; if (!indexing || index_visible) addword(wd, &whptr); if (indexing) addword(wd, &idximplicit); } if (sitem->type & stack_quote) { wd.text = NULL; wd.type = toquotestyle(style); wd.alt = NULL; wd.aux = quote_Close; wd.fpos = t.pos; wd.breaks = FALSE; if (!indexing || index_visible) addword(wd, &whptr); if (indexing) { rdadd(&indexstr, L'"'); addword(wd, &idximplicit); } } } sfree(sitem); break; case tok_cmd: switch (t.cmd) { case c__comment: /* * In-paragraph comment: \#{ balanced braces } * * Anything goes here; even tok_eop. We should * eat whitespace after the close brace _if_ * there was whitespace before the \#. */ dtor(t), t = get_token(in); if (t.type != tok_lbrace) { error(err_explbr, &t.pos); } else { int braces = 1; while (braces > 0) { dtor(t), t = get_token(in); if (t.type == tok_lbrace) braces++; else if (t.type == tok_rbrace) braces--; else if (t.type == tok_eof) { error(err_commenteof, &t.pos); break; } } } if (seenwhite) { already = TRUE; dtor(t), t = get_token(in); if (t.type == tok_white) { iswhite = TRUE; already = FALSE; } } break; case c_q: dtor(t), t = get_token(in); if (t.type != tok_lbrace) { error(err_explbr, &t.pos); } else { wd.text = NULL; wd.type = toquotestyle(style); wd.alt = NULL; wd.aux = quote_Open; wd.fpos = t.pos; wd.breaks = FALSE; if (!indexing || index_visible) addword(wd, &whptr); if (indexing) { rdadd(&indexstr, L'"'); addword(wd, &idximplicit); } sitem = mknew(struct stack_item); sitem->type = stack_quote; stk_push(parsestk, sitem); } break; case c_K: case c_k: case c_R: case c_W: case c_L: case c_date: /* * Keyword, hyperlink, or \date. We expect a * left brace, some text, and then a right * brace. No nesting; no arguments. */ wd.fpos = t.pos; wd.breaks = FALSE; if (t.cmd == c_K) wd.type = word_UpperXref; else if (t.cmd == c_k) wd.type = word_LowerXref; else if (t.cmd == c_R) wd.type = word_FreeTextXref; else if (t.cmd == c_W) wd.type = word_HyperLink; else if (t.cmd == c_L) wd.type = word_LocalHyperLink; else wd.type = word_Normal; dtor(t), t = get_token(in); if (t.type != tok_lbrace) { if (wd.type == word_Normal) { time_t thetime = time(NULL); struct tm *broken = localtime(&thetime); already = TRUE; wdtext = ustrftime(NULL, broken); wd.type = style; } else { error(err_explbr, &t.pos); wdtext = NULL; } } else { rdstring rs = { 0, 0, NULL }; while (dtor(t), t = get_token(in), t.type == tok_word || t.type == tok_white) { if (t.type == tok_white) rdadd(&rs, ' '); else rdadds(&rs, t.text); } if (wd.type == word_Normal) { time_t thetime = time(NULL); struct tm *broken = localtime(&thetime); wdtext = ustrftime(rs.text, broken); wd.type = style; } else { wdtext = ustrdup(rs.text); } sfree(rs.text); if (t.type != tok_rbrace) { error(err_kwexprbr, &t.pos); } } wd.alt = NULL; wd.aux = 0; if (!indexing || index_visible) { wd.text = ustrdup(wdtext); addword(wd, &whptr); } if (indexing) { wd.text = ustrdup(wdtext); addword(wd, &idximplicit); } sfree(wdtext); if (wd.type == word_FreeTextXref || wd.type == word_HyperLink || wd.type == word_LocalHyperLink) { /* * Hyperlinks are different: they then * expect another left brace, to begin * delimiting the text marked by the link. */ dtor(t), t = get_token(in); /* * Special cases: \W{}\c, \W{}\e, \W{}\cw */ sitem = mknew(struct stack_item); sitem->type = stack_hyper; if (t.type == tok_cmd && (t.cmd == c_e || t.cmd == c_c || t.cmd == c_cw)) { if (style != word_Normal) error(err_nestedstyles, &t.pos); else { style = (t.cmd == c_c ? word_Code : t.cmd == c_cw ? word_WeakCode : word_Emph); spcstyle = tospacestyle(style); sitem->type |= stack_style; } dtor(t), t = get_token(in); } if (t.type != tok_lbrace) { error(err_explbr, &t.pos); sfree(sitem); } else { stk_push(parsestk, sitem); } } break; case c_c: case c_cw: case c_e: type = t.cmd; if (style != word_Normal) { error(err_nestedstyles, &t.pos); /* Error recovery: eat lbrace, push nop. */ dtor(t), t = get_token(in); sitem = mknew(struct stack_item); sitem->type = stack_nop; stk_push(parsestk, sitem); } dtor(t), t = get_token(in); if (t.type != tok_lbrace) { error(err_explbr, &t.pos); } else { style = (type == c_c ? word_Code : type == c_cw ? word_WeakCode : word_Emph); spcstyle = tospacestyle(style); sitem = mknew(struct stack_item); sitem->type = stack_style; stk_push(parsestk, sitem); } break; case c_i: case c_ii: case c_I: type = t.cmd; if (indexing) { error(err_nestedindex, &t.pos); /* Error recovery: eat lbrace, push nop. */ dtor(t), t = get_token(in); sitem = mknew(struct stack_item); sitem->type = stack_nop; stk_push(parsestk, sitem); } sitem = mknew(struct stack_item); sitem->type = stack_idx; dtor(t), t = get_token(in); /* * Special cases: \i\c, \i\e, \i\cw */ wd.fpos = t.pos; if (t.type == tok_cmd && (t.cmd == c_e || t.cmd == c_c || t.cmd == c_cw)) { if (style != word_Normal) error(err_nestedstyles, &t.pos); else { style = (t.cmd == c_c ? word_Code : t.cmd == c_cw ? word_WeakCode : word_Emph); spcstyle = tospacestyle(style); sitem->type |= stack_style; } dtor(t), t = get_token(in); } if (t.type != tok_lbrace) { sfree(sitem); error(err_explbr, &t.pos); } else { /* Add an index-reference word with no text as yet */ wd.type = word_IndexRef; wd.text = NULL; wd.alt = NULL; wd.aux = 0; wd.breaks = FALSE; indexword = addword(wd, &whptr); /* Set up a rdstring to read the index text */ indexstr = nullrs; /* Flags so that we do the Right Things with text */ index_visible = (type != c_I); index_downcase = (type == c_ii); indexing = TRUE; idxwordlist = NULL; idximplicit = &idxwordlist; /* Stack item to close the indexing on exit */ stk_push(parsestk, sitem); } break; case c_u: uchr = t.aux; utext[0] = uchr; utext[1] = 0; wd.type = style; wd.breaks = FALSE; wd.alt = NULL; wd.aux = 0; wd.fpos = t.pos; if (!indexing || index_visible) { wd.text = ustrdup(utext); uword = addword(wd, &whptr); } else uword = NULL; if (indexing) { wd.text = ustrdup(utext); iword = addword(wd, &idximplicit); } else iword = NULL; dtor(t), t = get_token(in); if (t.type == tok_lbrace) { /* * \u with a left brace. Until the brace * closes, all further words go on a * sidetrack from the main thread of the * paragraph. */ sitem = mknew(struct stack_item); sitem->type = stack_ualt; sitem->whptr = whptr; sitem->idximplicit = idximplicit; stk_push(parsestk, sitem); whptr = uword ? &uword->alt : NULL; idximplicit = iword ? &iword->alt : NULL; } else { if (indexing) rdadd(&indexstr, uchr); already = TRUE; } break; default: if (!macrolookup(macros, in, t.text, &t.pos)) error(err_badmidcmd, t.text, &t.pos); break; }