1 #include <u.h> 2 #include <libc.h> 3 #include <draw.h> 4 #include <ctype.h> 5 #include <html.h> 6 #include "impl.h" 7 8 // A stack for holding integer values 9 enum { 10 Nestmax = 40 // max nesting level of lists, font styles, etc. 11 }; 12 13 struct Stack { 14 int n; // next available slot (top of stack is stack[n-1]) 15 int slots[Nestmax]; // stack entries 16 }; 17 18 // Parsing state 19 struct Pstate 20 { 21 Pstate* next; // in stack of Pstates 22 int skipping; // true when we shouldn't add items 23 int skipwhite; // true when we should strip leading space 24 int curfont; // font index for current font 25 int curfg; // current foreground color 26 Background curbg; // current background 27 int curvoff; // current baseline offset 28 uchar curul; // current underline/strike state 29 uchar curjust; // current justify state 30 int curanchor; // current (href) anchor id (if in one), or 0 31 int curstate; // current value of item state 32 int literal; // current literal state 33 int inpar; // true when in a paragraph-like construct 34 int adjsize; // current font size adjustment 35 Item* items; // dummy head of item list we're building 36 Item* lastit; // tail of item list we're building 37 Item* prelastit; // item before lastit 38 Stack fntstylestk; // style stack 39 Stack fntsizestk; // size stack 40 Stack fgstk; // text color stack 41 Stack ulstk; // underline stack 42 Stack voffstk; // vertical offset stack 43 Stack listtypestk; // list type stack 44 Stack listcntstk; // list counter stack 45 Stack juststk; // justification stack 46 Stack hangstk; // hanging stack 47 }; 48 49 struct ItemSource 50 { 51 Docinfo* doc; 52 Pstate* psstk; 53 int nforms; 54 int ntables; 55 int nanchors; 56 int nframes; 57 Form* curform; 58 Map* curmap; 59 Table* tabstk; 60 Kidinfo* kidstk; 61 }; 62 63 // Some layout parameters 64 enum { 65 FRKIDMARGIN = 6, // default margin around kid frames 66 IMGHSPACE = 0, // default hspace for images (0 matches IE, Netscape) 67 IMGVSPACE = 0, // default vspace for images 68 FLTIMGHSPACE = 2, // default hspace for float images 69 TABSP = 5, // default cellspacing for tables 70 TABPAD = 1, // default cell padding for tables 71 LISTTAB = 1, // number of tabs to indent lists 72 BQTAB = 1, // number of tabs to indent blockquotes 73 HRSZ = 2, // thickness of horizontal rules 74 SUBOFF = 4, // vertical offset for subscripts 75 SUPOFF = 6, // vertical offset for superscripts 76 NBSP = 160 // non-breaking space character 77 }; 78 79 // These tables must be sorted 80 static StringInt align_tab[] = { 81 {L"baseline", ALbaseline}, 82 {L"bottom", ALbottom}, 83 {L"center", ALcenter}, 84 {L"char", ALchar}, 85 {L"justify", ALjustify}, 86 {L"left", ALleft}, 87 {L"middle", ALmiddle}, 88 {L"right", ALright}, 89 {L"top", ALtop} 90 }; 91 #define NALIGNTAB (sizeof(align_tab)/sizeof(StringInt)) 92 93 static StringInt input_tab[] = { 94 {L"button", Fbutton}, 95 {L"checkbox", Fcheckbox}, 96 {L"file", Ffile}, 97 {L"hidden", Fhidden}, 98 {L"image", Fimage}, 99 {L"password", Fpassword}, 100 {L"radio", Fradio}, 101 {L"reset", Freset}, 102 {L"submit", Fsubmit}, 103 {L"text", Ftext} 104 }; 105 #define NINPUTTAB (sizeof(input_tab)/sizeof(StringInt)) 106 107 static StringInt clear_tab[] = { 108 {L"all", IFcleft|IFcright}, 109 {L"left", IFcleft}, 110 {L"right", IFcright} 111 }; 112 #define NCLEARTAB (sizeof(clear_tab)/sizeof(StringInt)) 113 114 static StringInt fscroll_tab[] = { 115 {L"auto", FRhscrollauto|FRvscrollauto}, 116 {L"no", FRnoscroll}, 117 {L"yes", FRhscroll|FRvscroll}, 118 }; 119 #define NFSCROLLTAB (sizeof(fscroll_tab)/sizeof(StringInt)) 120 121 static StringInt shape_tab[] = { 122 {L"circ", SHcircle}, 123 {L"circle", SHcircle}, 124 {L"poly", SHpoly}, 125 {L"polygon", SHpoly}, 126 {L"rect", SHrect}, 127 {L"rectangle", SHrect} 128 }; 129 #define NSHAPETAB (sizeof(shape_tab)/sizeof(StringInt)) 130 131 static StringInt method_tab[] = { 132 {L"get", HGet}, 133 {L"post", HPost} 134 }; 135 #define NMETHODTAB (sizeof(method_tab)/sizeof(StringInt)) 136 137 static Rune* roman[15]= { 138 L"I", L"II", L"III", L"IV", L"V", L"VI", L"VII", L"VIII", L"IX", L"X", 139 L"XI", L"XII", L"XIII", L"XIV", L"XV" 140 }; 141 #define NROMAN 15 142 143 // List number types 144 enum { 145 LTdisc, LTsquare, LTcircle, LT1, LTa, LTA, LTi, LTI 146 }; 147 148 enum { 149 SPBefore = 2, 150 SPAfter = 4, 151 BL = 1, 152 BLBA = (BL|SPBefore|SPAfter) 153 }; 154 155 // blockbrk[tag] is break info for a block level element, or one 156 // of a few others that get the same treatment re ending open paragraphs 157 // and requiring a line break / vertical space before them. 158 // If we want a line of space before the given element, SPBefore is OR'd in. 159 // If we want a line of space after the given element, SPAfter is OR'd in. 160 161 static uchar blockbrk[Numtags]= { 162 [Taddress] BLBA, [Tblockquote] BLBA, [Tcenter] BL, 163 [Tdir] BLBA, [Tdiv] BL, [Tdd] BL, [Tdl] BLBA, 164 [Tdt] BL, [Tform] BLBA, 165 // headings and tables get breaks added manually 166 [Th1] BL, [Th2] BL, [Th3] BL, 167 [Th4] BL, [Th5] BL, [Th6] BL, 168 [Thr] BL, [Tisindex] BLBA, [Tli] BL, [Tmenu] BLBA, 169 [Tol] BLBA, [Tp] BLBA, [Tpre] BLBA, 170 [Tul] BLBA 171 }; 172 173 enum { 174 AGEN = 1 175 }; 176 177 // attrinfo is information about attributes. 178 // The AGEN value means that the attribute is generic (applies to almost all elements) 179 static uchar attrinfo[Numattrs]= { 180 [Aid] AGEN, [Aclass] AGEN, [Astyle] AGEN, [Atitle] AGEN, 181 [Aonblur] AGEN, [Aonchange] AGEN, [Aonclick] AGEN, 182 [Aondblclick] AGEN, [Aonfocus] AGEN, [Aonkeypress] AGEN, 183 [Aonkeyup] AGEN, [Aonload] AGEN, [Aonmousedown] AGEN, 184 [Aonmousemove] AGEN, [Aonmouseout] AGEN, [Aonmouseover] AGEN, 185 [Aonmouseup] AGEN, [Aonreset] AGEN, [Aonselect] AGEN, 186 [Aonsubmit] AGEN, [Aonunload] AGEN 187 }; 188 189 static uchar scriptev[Numattrs]= { 190 [Aonblur] SEonblur, [Aonchange] SEonchange, [Aonclick] SEonclick, 191 [Aondblclick] SEondblclick, [Aonfocus] SEonfocus, [Aonkeypress] SEonkeypress, 192 [Aonkeyup] SEonkeyup, [Aonload] SEonload, [Aonmousedown] SEonmousedown, 193 [Aonmousemove] SEonmousemove, [Aonmouseout] SEonmouseout, [Aonmouseover] SEonmouseover, 194 [Aonmouseup] SEonmouseup, [Aonreset] SEonreset, [Aonselect] SEonselect, 195 [Aonsubmit] SEonsubmit, [Aonunload] SEonunload 196 }; 197 198 // Color lookup table 199 static StringInt color_tab[] = { 200 {L"aqua", 0x00FFFF}, 201 {L"black", 0x000000}, 202 {L"blue", 0x0000CC}, 203 {L"fuchsia", 0xFF00FF}, 204 {L"gray", 0x808080}, 205 {L"green", 0x008000}, 206 {L"lime", 0x00FF00}, 207 {L"maroon", 0x800000}, 208 {L"navy", 0x000080,}, 209 {L"olive", 0x808000}, 210 {L"purple", 0x800080}, 211 {L"red", 0xFF0000}, 212 {L"silver", 0xC0C0C0}, 213 {L"teal", 0x008080}, 214 {L"white", 0xFFFFFF}, 215 {L"yellow", 0xFFFF00} 216 }; 217 #define NCOLORS (sizeof(color_tab)/sizeof(StringInt)) 218 219 static StringInt *targetmap; 220 static int targetmapsize; 221 static int ntargets; 222 223 static int buildinited = 0; 224 225 #define SMALLBUFSIZE 240 226 #define BIGBUFSIZE 2000 227 228 int dbgbuild = 0; 229 int warn = 0; 230 231 static Align aalign(Token* tok); 232 static int acolorval(Token* tok, int attid, int dflt); 233 static void addbrk(Pstate* ps, int sp, int clr); 234 static void additem(Pstate* ps, Item* it, Token* tok); 235 static void addlinebrk(Pstate* ps, int clr); 236 static void addnbsp(Pstate* ps); 237 static void addtext(Pstate* ps, Rune* s); 238 static Dimen adimen(Token* tok, int attid); 239 static int aflagval(Token* tok, int attid); 240 static int aintval(Token* tok, int attid, int dflt); 241 static Rune* astrval(Token* tok, int attid, Rune* dflt); 242 static int atabval(Token* tok, int attid, StringInt* tab, int ntab, int dflt); 243 static int atargval(Token* tok, int dflt); 244 static int auintval(Token* tok, int attid, int dflt); 245 static Rune* aurlval(Token* tok, int attid, Rune* dflt, Rune* base); 246 static Rune* aval(Token* tok, int attid); 247 static void buildinit(void); 248 static Pstate* cell_pstate(Pstate* oldps, int ishead); 249 static void changehang(Pstate* ps, int delta); 250 static void changeindent(Pstate* ps, int delta); 251 static int color(Rune* s, int dflt); 252 static void copystack(Stack* tostk, Stack* fromstk); 253 static int dimprint(char* buf, int nbuf, Dimen d); 254 static Pstate* finishcell(Table* curtab, Pstate* psstk); 255 static void finish_table(Table* t); 256 static void freeanchor(Anchor* a); 257 static void freedestanchor(DestAnchor* da); 258 static void freeform(Form* f); 259 static void freeformfield(Formfield* ff); 260 static void freeitem(Item* it); 261 static void freepstate(Pstate* p); 262 static void freepstatestack(Pstate* pshead); 263 static void freescriptevents(SEvent* ehead); 264 static void freetable(Table* t); 265 static Map* getmap(Docinfo* di, Rune* name); 266 static Rune* getpcdata(Token* toks, int tokslen, int* ptoki); 267 static Pstate* lastps(Pstate* psl); 268 static Rune* listmark(uchar ty, int n); 269 static int listtyval(Token* tok, int dflt); 270 static Align makealign(int halign, int valign); 271 static Background makebackground(Rune* imgurl, int color); 272 static Dimen makedimen(int kind, int spec); 273 static Anchor* newanchor(int index, Rune* name, Rune* href, int target, Anchor* link); 274 static Area* newarea(int shape, Rune* href, int target, Area* link); 275 static DestAnchor* newdestanchor(int index, Rune* name, Item* item, DestAnchor* link); 276 static Docinfo* newdocinfo(void); 277 static Genattr* newgenattr(Rune* id, Rune* class, Rune* style, Rune* title, Attr* events); 278 static Form* newform(int formid, Rune* name, Rune* action, 279 int target, int method, Form* link); 280 static Formfield* newformfield(int ftype, int fieldid, Form* form, Rune* name, 281 Rune* value, int size, int maxlength, Formfield* link); 282 static Item* newifloat(Item* it, int side); 283 static Item* newiformfield(Formfield* ff); 284 static Item* newiimage(Rune* src, Rune* altrep, int align, int width, int height, 285 int hspace, int vspace, int border, int ismap, Map* map); 286 static Item* newirule(int align, int size, int noshade, Dimen wspec); 287 static Item* newispacer(int spkind); 288 static Item* newitable(Table* t); 289 static ItemSource* newitemsource(Docinfo* di); 290 static Item* newitext(Rune* s, int fnt, int fg, int voff, int ul); 291 static Kidinfo* newkidinfo(int isframeset, Kidinfo* link); 292 static Option* newoption(int selected, Rune* value, Rune* display, Option* link); 293 static Pstate* newpstate(Pstate* link); 294 static SEvent* newscriptevent(int type, Rune* script, SEvent* link); 295 static Table* newtable(int tableid, Align align, Dimen width, int border, 296 int cellspacing, int cellpadding, Background bg, Token* tok, Table* link); 297 static Tablecell* newtablecell(int cellid, int rowspan, int colspan, Align align, Dimen wspec, 298 int hspec, Background bg, int flags, Tablecell* link); 299 static Tablerow* newtablerow(Align align, Background bg, int flags, Tablerow* link); 300 static Dimen parsedim(Rune* s, int ns); 301 static void pop(Stack* stk); 302 static void popfontsize(Pstate* ps); 303 static void popfontstyle(Pstate* ps); 304 static void popjust(Pstate* ps); 305 static int popretnewtop(Stack* stk, int dflt); 306 static int push(Stack* stk, int val); 307 static void pushfontsize(Pstate* ps, int sz); 308 static void pushfontstyle(Pstate* ps, int sty); 309 static void pushjust(Pstate* ps, int j); 310 static Item* textit(Pstate* ps, Rune* s); 311 static Rune* removeallwhite(Rune* s); 312 static void resetdocinfo(Docinfo* d); 313 static void setcurfont(Pstate* ps); 314 static void setcurjust(Pstate* ps); 315 static void setdimarray(Token* tok, int attid, Dimen** pans, int* panslen); 316 static Rune* stringalign(int a); 317 static void targetmapinit(void); 318 static int toint(Rune* s); 319 static int top(Stack* stk, int dflt); 320 static void trim_cell(Tablecell* c); 321 static int validalign(Align a); 322 static int validdimen(Dimen d); 323 static int validformfield(Formfield* f); 324 static int validhalign(int a); 325 static int validptr(void* p); 326 static int validStr(Rune* s); 327 static int validtable(Table* t); 328 static int validtablerow(Tablerow* r); 329 static int validtablecol(Tablecol* c); 330 static int validtablecell(Tablecell* c); 331 static int validvalign(int a); 332 static int Iconv(Fmt *f); 333 334 static void 335 buildinit(void) 336 { 337 fmtinstall('I', Iconv); 338 targetmapinit(); 339 buildinited = 1; 340 } 341 342 static ItemSource* 343 newitemsource(Docinfo* di) 344 { 345 ItemSource* is; 346 Pstate* ps; 347 348 ps = newpstate(nil); 349 if(di->mediatype != TextHtml) { 350 ps->curstate &= ~IFwrap; 351 ps->literal = 1; 352 pushfontstyle(ps, FntT); 353 } 354 is = (ItemSource*)emalloc(sizeof(ItemSource)); 355 is->doc = di; 356 is->psstk = ps; 357 is->nforms = 0; 358 is->ntables = 0; 359 is->nanchors = 0; 360 is->nframes = 0; 361 is->curform = nil; 362 is->curmap = nil; 363 is->tabstk = nil; 364 is->kidstk = nil; 365 return is; 366 } 367 368 static Item *getitems(ItemSource* is, uchar* data, int datalen); 369 370 // Parse an html document and create a list of layout items. 371 // Allocate and return document info in *pdi. 372 // When caller is done with the items, it should call 373 // freeitems on the returned result, and then 374 // freedocinfo(*pdi). 375 Item* 376 parsehtml(uchar* data, int datalen, Rune* pagesrc, int mtype, int chset, Docinfo** pdi) 377 { 378 Item *it; 379 Docinfo* di; 380 ItemSource* is; 381 382 di = newdocinfo(); 383 di->src = _Strdup(pagesrc); 384 di->base = _Strdup(pagesrc); 385 di->mediatype = mtype; 386 di->chset = chset; 387 *pdi = di; 388 is = newitemsource(di); 389 it = getitems(is, data, datalen); 390 freepstatestack(is->psstk); 391 free(is); 392 return it; 393 } 394 395 // Get a group of tokens for lexer, parse them, and create 396 // a list of layout items. 397 // When caller is done with the items, it should call 398 // freeitems on the returned result. 399 static Item* 400 getitems(ItemSource* is, uchar* data, int datalen) 401 { 402 int i; 403 int j; 404 int nt; 405 int pt; 406 int doscripts; 407 int tokslen; 408 int toki; 409 int h; 410 int sz; 411 int method; 412 int n; 413 int nblank; 414 int norsz; 415 int bramt; 416 int sty; 417 int nosh; 418 int oldcuranchor; 419 int dfltbd; 420 int v; 421 int hang; 422 int isempty; 423 int tag; 424 int brksp; 425 int target; 426 uchar brk; 427 uchar flags; 428 uchar align; 429 uchar al; 430 uchar ty; 431 uchar ty2; 432 Pstate* ps; 433 Pstate* nextps; 434 Pstate* outerps; 435 Table* curtab; 436 Token* tok; 437 Token* toks; 438 Docinfo* di; 439 Item* ans; 440 Item* img; 441 Item* ffit; 442 Item* tabitem; 443 Rune* s; 444 Rune* t; 445 Rune* name; 446 Rune* enctype; 447 Rune* usemap; 448 Rune* prompt; 449 Rune* equiv; 450 Rune* val; 451 Rune* nsz; 452 Rune* script; 453 Map* map; 454 Form* frm; 455 Iimage* ii; 456 Kidinfo* kd; 457 Kidinfo* ks; 458 Kidinfo* pks; 459 Dimen wd; 460 Option* option; 461 Table* tab; 462 Tablecell* c; 463 Tablerow* tr; 464 Formfield* field; 465 Formfield* ff; 466 Rune* href; 467 Rune* src; 468 Rune* scriptsrc; 469 Rune* bgurl; 470 Rune* action; 471 Background bg; 472 473 if(!buildinited) 474 buildinit(); 475 doscripts = 0; // for now 476 ps = is->psstk; 477 curtab = is->tabstk; 478 di = is->doc; 479 toks = _gettoks(data, datalen, di->chset, di->mediatype, &tokslen); 480 toki = 0; 481 for(; toki < tokslen; toki++) { 482 tok = &toks[toki]; 483 if(dbgbuild > 1) 484 fprint(2, "build: curstate %ux, token %T\n", ps->curstate, tok); 485 tag = tok->tag; 486 brk = 0; 487 brksp = 0; 488 if(tag < Numtags) { 489 brk = blockbrk[tag]; 490 if(brk&SPBefore) 491 brksp = 1; 492 } 493 else if(tag < Numtags + RBRA) { 494 brk = blockbrk[tag - RBRA]; 495 if(brk&SPAfter) 496 brksp = 1; 497 } 498 if(brk) { 499 addbrk(ps, brksp, 0); 500 if(ps->inpar) { 501 popjust(ps); 502 ps->inpar = 0; 503 } 504 } 505 // check common case first (Data), then switch statement on tag 506 if(tag == Data) { 507 // Lexing didn't pay attention to SGML record boundary rules: 508 // \n after start tag or before end tag to be discarded. 509 // (Lex has already discarded all \r's). 510 // Some pages assume this doesn't happen in <PRE> text, 511 // so we won't do it if literal is true. 512 // BUG: won't discard \n before a start tag that begins 513 // the next bufferful of tokens. 514 s = tok->text; 515 n = _Strlen(s); 516 if(!ps->literal) { 517 i = 0; 518 j = n; 519 if(toki > 0) { 520 pt = toks[toki - 1].tag; 521 // IE and Netscape both ignore this rule (contrary to spec) 522 // if previous tag was img 523 if(pt < Numtags && pt != Timg && j > 0 && s[0] == '\n') 524 i++; 525 } 526 if(toki < tokslen - 1) { 527 nt = toks[toki + 1].tag; 528 if(nt >= RBRA && nt < Numtags + RBRA && j > i && s[j - 1] == '\n') 529 j--; 530 } 531 if(i > 0 || j < n) { 532 t = s; 533 s = _Strsubstr(s, i, j); 534 free(t); 535 n = j-i; 536 } 537 } 538 if(ps->skipwhite) { 539 _trimwhite(s, n, &t, &nt); 540 if(t == nil) { 541 free(s); 542 s = nil; 543 } 544 else if(t != s) { 545 t = _Strndup(t, nt); 546 free(s); 547 s = t; 548 } 549 if(s != nil) 550 ps->skipwhite = 0; 551 } 552 tok->text = nil; // token doesn't own string anymore 553 if(s != nil) 554 addtext(ps, s); 555 } 556 else 557 switch(tag) { 558 // Some abbrevs used in following DTD comments 559 // %text = #PCDATA 560 // | TT | I | B | U | STRIKE | BIG | SMALL | SUB | SUP 561 // | EM | STRONG | DFN | CODE | SAMP | KBD | VAR | CITE 562 // | A | IMG | APPLET | FONT | BASEFONT | BR | SCRIPT | MAP 563 // | INPUT | SELECT | TEXTAREA 564 // %block = P | UL | OL | DIR | MENU | DL | PRE | DL | DIV | CENTER 565 // | BLOCKQUOTE | FORM | ISINDEX | HR | TABLE 566 // %flow = (%text | %block)* 567 // %body.content = (%heading | %text | %block | ADDRESS)* 568 569 // <!ELEMENT A - - (%text) -(A)> 570 // Anchors are not supposed to be nested, but you sometimes see 571 // href anchors inside destination anchors. 572 case Ta: 573 if(ps->curanchor != 0) { 574 if(warn) 575 fprint(2, "warning: nested <A> or missing </A>\n"); 576 ps->curanchor = 0; 577 } 578 name = aval(tok, Aname); 579 href = aurlval(tok, Ahref, nil, di->base); 580 // ignore rel, rev, and title attrs 581 if(href != nil) { 582 target = atargval(tok, di->target); 583 di->anchors = newanchor(++is->nanchors, name, href, target, di->anchors); 584 if(name != nil) 585 name = _Strdup(name); // for DestAnchor construction, below 586 ps->curanchor = is->nanchors; 587 ps->curfg = push(&ps->fgstk, di->link); 588 ps->curul = push(&ps->ulstk, ULunder); 589 } 590 if(name != nil) { 591 // add a null item to be destination 592 additem(ps, newispacer(ISPnull), tok); 593 di->dests = newdestanchor(++is->nanchors, name, ps->lastit, di->dests); 594 } 595 break; 596 597 case Ta+RBRA : 598 if(ps->curanchor != 0) { 599 ps->curfg = popretnewtop(&ps->fgstk, di->text); 600 ps->curul = popretnewtop(&ps->ulstk, ULnone); 601 ps->curanchor = 0; 602 } 603 break; 604 605 // <!ELEMENT APPLET - - (PARAM | %text)* > 606 // We can't do applets, so ignore PARAMS, and let 607 // the %text contents appear for the alternative rep 608 case Tapplet: 609 case Tapplet+RBRA: 610 if(warn && tag == Tapplet) 611 fprint(2, "warning: <APPLET> ignored\n"); 612 break; 613 614 // <!ELEMENT AREA - O EMPTY> 615 case Tarea: 616 map = di->maps; 617 if(map == nil) { 618 if(warn) 619 fprint(2, "warning: <AREA> not inside <MAP>\n"); 620 continue; 621 } 622 map->areas = newarea(atabval(tok, Ashape, shape_tab, NSHAPETAB, SHrect), 623 aurlval(tok, Ahref, nil, di->base), 624 atargval(tok, di->target), 625 map->areas); 626 setdimarray(tok, Acoords, &map->areas->coords, &map->areas->ncoords); 627 break; 628 629 // <!ELEMENT (B|STRONG) - - (%text)*> 630 case Tb: 631 case Tstrong: 632 pushfontstyle(ps, FntB); 633 break; 634 635 case Tb+RBRA: 636 case Tcite+RBRA: 637 case Tcode+RBRA: 638 case Tdfn+RBRA: 639 case Tem+RBRA: 640 case Tkbd+RBRA: 641 case Ti+RBRA: 642 case Tsamp+RBRA: 643 case Tstrong+RBRA: 644 case Ttt+RBRA: 645 case Tvar+RBRA : 646 case Taddress+RBRA: 647 popfontstyle(ps); 648 break; 649 650 // <!ELEMENT BASE - O EMPTY> 651 case Tbase: 652 t = di->base; 653 di->base = aurlval(tok, Ahref, di->base, di->base); 654 if(t != nil) 655 free(t); 656 di->target = atargval(tok, di->target); 657 break; 658 659 // <!ELEMENT BASEFONT - O EMPTY> 660 case Tbasefont: 661 ps->adjsize = aintval(tok, Asize, 3) - 3; 662 break; 663 664 // <!ELEMENT (BIG|SMALL) - - (%text)*> 665 case Tbig: 666 case Tsmall: 667 sz = ps->adjsize; 668 if(tag == Tbig) 669 sz += Large; 670 else 671 sz += Small; 672 pushfontsize(ps, sz); 673 break; 674 675 case Tbig+RBRA: 676 case Tsmall+RBRA: 677 popfontsize(ps); 678 break; 679 680 // <!ELEMENT BLOCKQUOTE - - %body.content> 681 case Tblockquote: 682 changeindent(ps, BQTAB); 683 break; 684 685 case Tblockquote+RBRA: 686 changeindent(ps, -BQTAB); 687 break; 688 689 // <!ELEMENT BODY O O %body.content> 690 case Tbody: 691 ps->skipping = 0; 692 bg = makebackground(nil, acolorval(tok, Abgcolor, di->background.color)); 693 bgurl = aurlval(tok, Abackground, nil, di->base); 694 if(bgurl != nil) { 695 if(di->backgrounditem != nil) 696 freeitem((Item*)di->backgrounditem); 697 // really should remove old item from di->images list, 698 // but there should only be one BODY element ... 699 di->backgrounditem = (Iimage*)newiimage(bgurl, nil, ALnone, 0, 0, 0, 0, 0, 0, nil); 700 di->backgrounditem->nextimage = di->images; 701 di->images = di->backgrounditem; 702 } 703 ps->curbg = bg; 704 di->background = bg; 705 di->text = acolorval(tok, Atext, di->text); 706 di->link = acolorval(tok, Alink, di->link); 707 di->vlink = acolorval(tok, Avlink, di->vlink); 708 di->alink = acolorval(tok, Aalink, di->alink); 709 if(di->text != ps->curfg) { 710 ps->curfg = di->text; 711 ps->fgstk.n = 0; 712 } 713 break; 714 715 case Tbody+RBRA: 716 // HTML spec says ignore things after </body>, 717 // but IE and Netscape don't 718 // ps.skipping = 1; 719 break; 720 721 // <!ELEMENT BR - O EMPTY> 722 case Tbr: 723 addlinebrk(ps, atabval(tok, Aclear, clear_tab, NCLEARTAB, 0)); 724 break; 725 726 // <!ELEMENT CAPTION - - (%text;)*> 727 case Tcaption: 728 if(curtab == nil) { 729 if(warn) 730 fprint(2, "warning: <CAPTION> outside <TABLE>\n"); 731 continue; 732 } 733 if(curtab->caption != nil) { 734 if(warn) 735 fprint(2, "warning: more than one <CAPTION> in <TABLE>\n"); 736 continue; 737 } 738 ps = newpstate(ps); 739 curtab->caption_place = atabval(tok, Aalign, align_tab, NALIGNTAB, ALtop); 740 break; 741 742 case Tcaption+RBRA: 743 nextps = ps->next; 744 if(curtab == nil || nextps == nil) { 745 if(warn) 746 fprint(2, "warning: unexpected </CAPTION>\n"); 747 continue; 748 } 749 curtab->caption = ps->items->next; 750 free(ps); 751 ps = nextps; 752 break; 753 754 case Tcenter: 755 case Tdiv: 756 if(tag == Tcenter) 757 al = ALcenter; 758 else 759 al = atabval(tok, Aalign, align_tab, NALIGNTAB, ps->curjust); 760 pushjust(ps, al); 761 break; 762 763 case Tcenter+RBRA: 764 case Tdiv+RBRA: 765 popjust(ps); 766 break; 767 768 // <!ELEMENT DD - O %flow > 769 case Tdd: 770 if(ps->hangstk.n == 0) { 771 if(warn) 772 fprint(2, "warning: <DD> not inside <DL\n"); 773 continue; 774 } 775 h = top(&ps->hangstk, 0); 776 if(h != 0) 777 changehang(ps, -10*LISTTAB); 778 else 779 addbrk(ps, 0, 0); 780 push(&ps->hangstk, 0); 781 break; 782 783 //<!ELEMENT (DIR|MENU) - - (LI)+ -(%block) > 784 //<!ELEMENT (OL|UL) - - (LI)+> 785 case Tdir: 786 case Tmenu: 787 case Tol: 788 case Tul: 789 changeindent(ps, LISTTAB); 790 push(&ps->listtypestk, listtyval(tok, (tag==Tol)? LT1 : LTdisc)); 791 push(&ps->listcntstk, aintval(tok, Astart, 1)); 792 break; 793 794 case Tdir+RBRA: 795 case Tmenu+RBRA: 796 case Tol+RBRA: 797 case Tul+RBRA: 798 if(ps->listtypestk.n == 0) { 799 if(warn) 800 fprint(2, "warning: %T ended no list\n", tok); 801 continue; 802 } 803 addbrk(ps, 0, 0); 804 pop(&ps->listtypestk); 805 pop(&ps->listcntstk); 806 changeindent(ps, -LISTTAB); 807 break; 808 809 // <!ELEMENT DL - - (DT|DD)+ > 810 case Tdl: 811 changeindent(ps, LISTTAB); 812 push(&ps->hangstk, 0); 813 break; 814 815 case Tdl+RBRA: 816 if(ps->hangstk.n == 0) { 817 if(warn) 818 fprint(2, "warning: unexpected </DL>\n"); 819 continue; 820 } 821 changeindent(ps, -LISTTAB); 822 if(top(&ps->hangstk, 0) != 0) 823 changehang(ps, -10*LISTTAB); 824 pop(&ps->hangstk); 825 break; 826 827 // <!ELEMENT DT - O (%text)* > 828 case Tdt: 829 if(ps->hangstk.n == 0) { 830 if(warn) 831 fprint(2, "warning: <DT> not inside <DL>\n"); 832 continue; 833 } 834 h = top(&ps->hangstk, 0); 835 pop(&ps->hangstk); 836 if(h != 0) 837 changehang(ps, -10*LISTTAB); 838 changehang(ps, 10*LISTTAB); 839 push(&ps->hangstk, 1); 840 break; 841 842 // <!ELEMENT FONT - - (%text)*> 843 case Tfont: 844 sz = top(&ps->fntsizestk, Normal); 845 if(_tokaval(tok, Asize, &nsz, 0)) { 846 if(_prefix(L"+", nsz)) 847 sz = Normal + _Strtol(nsz+1, nil, 10) + ps->adjsize; 848 else if(_prefix(L"-", nsz)) 849 sz = Normal - _Strtol(nsz+1, nil, 10) + ps->adjsize; 850 else if(nsz != nil) 851 sz = Normal + (_Strtol(nsz, nil, 10) - 3); 852 } 853 ps->curfg = push(&ps->fgstk, acolorval(tok, Acolor, ps->curfg)); 854 pushfontsize(ps, sz); 855 break; 856 857 case Tfont+RBRA: 858 if(ps->fgstk.n == 0) { 859 if(warn) 860 fprint(2, "warning: unexpected </FONT>\n"); 861 continue; 862 } 863 ps->curfg = popretnewtop(&ps->fgstk, di->text); 864 popfontsize(ps); 865 break; 866 867 // <!ELEMENT FORM - - %body.content -(FORM) > 868 case Tform: 869 if(is->curform != nil) { 870 if(warn) 871 fprint(2, "warning: <FORM> nested inside another\n"); 872 continue; 873 } 874 action = aurlval(tok, Aaction, di->base, di->base); 875 s = aval(tok, Aid); 876 name = astrval(tok, Aname, s); 877 if(s) 878 free(s); 879 target = atargval(tok, di->target); 880 method = atabval(tok, Amethod, method_tab, NMETHODTAB, HGet); 881 if(warn && _tokaval(tok, Aenctype, &enctype, 0) && 882 _Strcmp(enctype, L"application/x-www-form-urlencoded")) 883 fprint(2, "form enctype %S not handled\n", enctype); 884 frm = newform(++is->nforms, name, action, target, method, di->forms); 885 di->forms = frm; 886 is->curform = frm; 887 break; 888 889 case Tform+RBRA: 890 if(is->curform == nil) { 891 if(warn) 892 fprint(2, "warning: unexpected </FORM>\n"); 893 continue; 894 } 895 // put fields back in input order 896 is->curform->fields = (Formfield*)_revlist((List*)is->curform->fields); 897 is->curform = nil; 898 break; 899 900 // <!ELEMENT FRAME - O EMPTY> 901 case Tframe: 902 ks = is->kidstk; 903 if(ks == nil) { 904 if(warn) 905 fprint(2, "warning: <FRAME> not in <FRAMESET>\n"); 906 continue; 907 } 908 ks->kidinfos = kd = newkidinfo(0, ks->kidinfos); 909 kd->src = aurlval(tok, Asrc, nil, di->base); 910 kd->name = aval(tok, Aname); 911 if(kd->name == nil) { 912 s = _ltoStr(++is->nframes); 913 kd->name = _Strdup2(L"_fr", s); 914 free(s); 915 } 916 kd->marginw = auintval(tok, Amarginwidth, 0); 917 kd->marginh = auintval(tok, Amarginheight, 0); 918 kd->framebd = auintval(tok, Aframeborder, 1); 919 kd->flags = atabval(tok, Ascrolling, fscroll_tab, NFSCROLLTAB, kd->flags); 920 norsz = aflagval(tok, Anoresize); 921 if(norsz) 922 kd->flags |= FRnoresize; 923 break; 924 925 // <!ELEMENT FRAMESET - - (FRAME|FRAMESET)+> 926 case Tframeset: 927 ks = newkidinfo(1, nil); 928 pks = is->kidstk; 929 if(pks == nil) 930 di->kidinfo = ks; 931 else { 932 ks->next = pks->kidinfos; 933 pks->kidinfos = ks; 934 } 935 ks->nextframeset = pks; 936 is->kidstk = ks; 937 setdimarray(tok, Arows, &ks->rows, &ks->nrows); 938 if(ks->nrows == 0) { 939 ks->rows = (Dimen*)emalloc(sizeof(Dimen)); 940 ks->nrows = 1; 941 ks->rows[0] = makedimen(Dpercent, 100); 942 } 943 setdimarray(tok, Acols, &ks->cols, &ks->ncols); 944 if(ks->ncols == 0) { 945 ks->cols = (Dimen*)emalloc(sizeof(Dimen)); 946 ks->ncols = 1; 947 ks->cols[0] = makedimen(Dpercent, 100); 948 } 949 break; 950 951 case Tframeset+RBRA: 952 if(is->kidstk == nil) { 953 if(warn) 954 fprint(2, "warning: unexpected </FRAMESET>\n"); 955 continue; 956 } 957 ks = is->kidstk; 958 // put kids back in original order 959 // and add blank frames to fill out cells 960 n = ks->nrows*ks->ncols; 961 nblank = n - _listlen((List*)ks->kidinfos); 962 while(nblank-- > 0) 963 ks->kidinfos = newkidinfo(0, ks->kidinfos); 964 ks->kidinfos = (Kidinfo*)_revlist((List*)ks->kidinfos); 965 is->kidstk = is->kidstk->nextframeset; 966 if(is->kidstk == nil) { 967 // end input 968 ans = nil; 969 goto return_ans; 970 } 971 break; 972 973 // <!ELEMENT H1 - - (%text;)*>, etc. 974 case Th1: 975 case Th2: 976 case Th3: 977 case Th4: 978 case Th5: 979 case Th6: 980 bramt = 1; 981 if(ps->items == ps->lastit) 982 bramt = 0; 983 addbrk(ps, bramt, IFcleft|IFcright); 984 sz = Verylarge - (tag - Th1); 985 if(sz < Tiny) 986 sz = Tiny; 987 pushfontsize(ps, sz); 988 sty = top(&ps->fntstylestk, FntR); 989 if(tag == Th1) 990 sty = FntB; 991 pushfontstyle(ps, sty); 992 pushjust(ps, atabval(tok, Aalign, align_tab, NALIGNTAB, ps->curjust)); 993 ps->skipwhite = 1; 994 break; 995 996 case Th1+RBRA: 997 case Th2+RBRA: 998 case Th3+RBRA: 999 case Th4+RBRA: 1000 case Th5+RBRA: 1001 case Th6+RBRA: 1002 addbrk(ps, 1, IFcleft|IFcright); 1003 popfontsize(ps); 1004 popfontstyle(ps); 1005 popjust(ps); 1006 break; 1007 1008 case Thead: 1009 // HTML spec says ignore regular markup in head, 1010 // but Netscape and IE don't 1011 // ps.skipping = 1; 1012 break; 1013 1014 case Thead+RBRA: 1015 ps->skipping = 0; 1016 break; 1017 1018 // <!ELEMENT HR - O EMPTY> 1019 case Thr: 1020 al = atabval(tok, Aalign, align_tab, NALIGNTAB, ALcenter); 1021 sz = auintval(tok, Asize, HRSZ); 1022 wd = adimen(tok, Awidth); 1023 if(dimenkind(wd) == Dnone) 1024 wd = makedimen(Dpercent, 100); 1025 nosh = aflagval(tok, Anoshade); 1026 additem(ps, newirule(al, sz, nosh, wd), tok); 1027 addbrk(ps, 0, 0); 1028 break; 1029 1030 case Ti: 1031 case Tcite: 1032 case Tdfn: 1033 case Tem: 1034 case Tvar: 1035 case Taddress: 1036 pushfontstyle(ps, FntI); 1037 break; 1038 1039 // <!ELEMENT IMG - O EMPTY> 1040 case Timg: 1041 map = nil; 1042 oldcuranchor = ps->curanchor; 1043 if(_tokaval(tok, Ausemap, &usemap, 0)) { 1044 if(!_prefix(L"#", usemap)) { 1045 if(warn) 1046 fprint(2, "warning: can't handle non-local map %S\n", usemap); 1047 } 1048 else { 1049 map = getmap(di, usemap+1); 1050 if(ps->curanchor == 0) { 1051 di->anchors = newanchor(++is->nanchors, nil, nil, di->target, di->anchors); 1052 ps->curanchor = is->nanchors; 1053 } 1054 } 1055 } 1056 align = atabval(tok, Aalign, align_tab, NALIGNTAB, ALbottom); 1057 dfltbd = 0; 1058 if(ps->curanchor != 0) 1059 dfltbd = 2; 1060 src = aurlval(tok, Asrc, nil, di->base); 1061 if(src == nil) { 1062 if(warn) 1063 fprint(2, "warning: <img> has no src attribute\n"); 1064 ps->curanchor = oldcuranchor; 1065 continue; 1066 } 1067 img = newiimage(src, 1068 aval(tok, Aalt), 1069 align, 1070 auintval(tok, Awidth, 0), 1071 auintval(tok, Aheight, 0), 1072 auintval(tok, Ahspace, IMGHSPACE), 1073 auintval(tok, Avspace, IMGVSPACE), 1074 auintval(tok, Aborder, dfltbd), 1075 aflagval(tok, Aismap), 1076 map); 1077 if(align == ALleft || align == ALright) { 1078 additem(ps, newifloat(img, align), tok); 1079 // if no hspace specified, use FLTIMGHSPACE 1080 if(!_tokaval(tok, Ahspace, &val, 0)) 1081 ((Iimage*)img)->hspace = FLTIMGHSPACE; 1082 } 1083 else { 1084 ps->skipwhite = 0; 1085 additem(ps, img, tok); 1086 } 1087 if(!ps->skipping) { 1088 ((Iimage*)img)->nextimage = di->images; 1089 di->images = (Iimage*)img; 1090 } 1091 ps->curanchor = oldcuranchor; 1092 break; 1093 1094 // <!ELEMENT INPUT - O EMPTY> 1095 case Tinput: 1096 ps->skipwhite = 0; 1097 if(is->curform == nil) { 1098 if(warn) 1099 fprint(2, "<INPUT> not inside <FORM>\n"); 1100 continue; 1101 } 1102 is->curform->fields = field = newformfield( 1103 atabval(tok, Atype, input_tab, NINPUTTAB, Ftext), 1104 ++is->curform->nfields, 1105 is->curform, 1106 aval(tok, Aname), 1107 aval(tok, Avalue), 1108 auintval(tok, Asize, 0), 1109 auintval(tok, Amaxlength, 1000), 1110 is->curform->fields); 1111 if(aflagval(tok, Achecked)) 1112 field->flags = FFchecked; 1113 1114 switch(field->ftype) { 1115 case Ftext: 1116 case Fpassword: 1117 case Ffile: 1118 if(field->size == 0) 1119 field->size = 20; 1120 break; 1121 1122 case Fcheckbox: 1123 if(field->name == nil) { 1124 if(warn) 1125 fprint(2, "warning: checkbox form field missing name\n"); 1126 continue; 1127 } 1128 if(field->value == nil) 1129 field->value = _Strdup(L"1"); 1130 break; 1131 1132 case Fradio: 1133 if(field->name == nil || field->value == nil) { 1134 if(warn) 1135 fprint(2, "warning: radio form field missing name or value\n"); 1136 continue; 1137 } 1138 break; 1139 1140 case Fsubmit: 1141 if(field->value == nil) 1142 field->value = _Strdup(L"Submit"); 1143 if(field->name == nil) 1144 field->name = _Strdup(L"_no_name_submit_"); 1145 break; 1146 1147 case Fimage: 1148 src = aurlval(tok, Asrc, nil, di->base); 1149 if(src == nil) { 1150 if(warn) 1151 fprint(2, "warning: image form field missing src\n"); 1152 continue; 1153 } 1154 // width and height attrs aren't specified in HTML 3.2, 1155 // but some people provide them and they help avoid 1156 // a relayout 1157 field->image = newiimage(src, 1158 astrval(tok, Aalt, L"Submit"), 1159 atabval(tok, Aalign, align_tab, NALIGNTAB, ALbottom), 1160 auintval(tok, Awidth, 0), auintval(tok, Aheight, 0), 1161 0, 0, 0, 0, nil); 1162 ii = (Iimage*)field->image; 1163 ii->nextimage = di->images; 1164 di->images = ii; 1165 break; 1166 1167 case Freset: 1168 if(field->value == nil) 1169 field->value = _Strdup(L"Reset"); 1170 break; 1171 1172 case Fbutton: 1173 if(field->value == nil) 1174 field->value = _Strdup(L" "); 1175 break; 1176 } 1177 ffit = newiformfield(field); 1178 additem(ps, ffit, tok); 1179 if(ffit->genattr != nil) 1180 field->events = ffit->genattr->events; 1181 break; 1182 1183 // <!ENTITY ISINDEX - O EMPTY> 1184 case Tisindex: 1185 ps->skipwhite = 0; 1186 prompt = astrval(tok, Aprompt, L"Index search terms:"); 1187 target = atargval(tok, di->target); 1188 additem(ps, textit(ps, prompt), tok); 1189 frm = newform(++is->nforms, 1190 nil, 1191 di->base, 1192 target, 1193 HGet, 1194 di->forms); 1195 di->forms = frm; 1196 ff = newformfield(Ftext, 1197 1, 1198 frm, 1199 _Strdup(L"_ISINDEX_"), 1200 nil, 1201 50, 1202 1000, 1203 nil); 1204 frm->fields = ff; 1205 frm->nfields = 1; 1206 additem(ps, newiformfield(ff), tok); 1207 addbrk(ps, 1, 0); 1208 break; 1209 1210 // <!ELEMENT LI - O %flow> 1211 case Tli: 1212 if(ps->listtypestk.n == 0) { 1213 if(warn) 1214 fprint(2, "<LI> not in list\n"); 1215 continue; 1216 } 1217 ty = top(&ps->listtypestk, 0); 1218 ty2 = listtyval(tok, ty); 1219 if(ty != ty2) { 1220 ty = ty2; 1221 push(&ps->listtypestk, ty2); 1222 } 1223 v = aintval(tok, Avalue, top(&ps->listcntstk, 1)); 1224 if(ty == LTdisc || ty == LTsquare || ty == LTcircle) 1225 hang = 10*LISTTAB - 3; 1226 else 1227 hang = 10*LISTTAB - 1; 1228 changehang(ps, hang); 1229 addtext(ps, listmark(ty, v)); 1230 push(&ps->listcntstk, v + 1); 1231 changehang(ps, -hang); 1232 ps->skipwhite = 1; 1233 break; 1234 1235 // <!ELEMENT MAP - - (AREA)+> 1236 case Tmap: 1237 if(_tokaval(tok, Aname, &name, 0)) 1238 is->curmap = getmap(di, name); 1239 break; 1240 1241 case Tmap+RBRA: 1242 map = is->curmap; 1243 if(map == nil) { 1244 if(warn) 1245 fprint(2, "warning: unexpected </MAP>\n"); 1246 continue; 1247 } 1248 map->areas = (Area*)_revlist((List*)map->areas); 1249 break; 1250 1251 case Tmeta: 1252 if(ps->skipping) 1253 continue; 1254 if(_tokaval(tok, Ahttp_equiv, &equiv, 0)) { 1255 val = aval(tok, Acontent); 1256 n = _Strlen(equiv); 1257 if(!_Strncmpci(equiv, n, L"refresh")) 1258 di->refresh = val; 1259 else if(!_Strncmpci(equiv, n, L"content-script-type")) { 1260 n = _Strlen(val); 1261 if(!_Strncmpci(val, n, L"javascript") 1262 || !_Strncmpci(val, n, L"jscript1.1") 1263 || !_Strncmpci(val, n, L"jscript")) 1264 di->scripttype = TextJavascript; 1265 else { 1266 if(warn) 1267 fprint(2, "unimplemented script type %S\n", val); 1268 di->scripttype = UnknownType; 1269 } 1270 } 1271 } 1272 break; 1273 1274 // Nobr is NOT in HMTL 4.0, but it is ubiquitous on the web 1275 case Tnobr: 1276 ps->skipwhite = 0; 1277 ps->curstate &= ~IFwrap; 1278 break; 1279 1280 case Tnobr+RBRA: 1281 ps->curstate |= IFwrap; 1282 break; 1283 1284 // We do frames, so skip stuff in noframes 1285 case Tnoframes: 1286 ps->skipping = 1; 1287 break; 1288 1289 case Tnoframes+RBRA: 1290 ps->skipping = 0; 1291 break; 1292 1293 // We do scripts (if enabled), so skip stuff in noscripts 1294 case Tnoscript: 1295 if(doscripts) 1296 ps->skipping = 1; 1297 break; 1298 1299 case Tnoscript+RBRA: 1300 if(doscripts) 1301 ps->skipping = 0; 1302 break; 1303 1304 // <!ELEMENT OPTION - O ( //PCDATA)> 1305 case Toption: 1306 if(is->curform == nil || is->curform->fields == nil) { 1307 if(warn) 1308 fprint(2, "warning: <OPTION> not in <SELECT>\n"); 1309 continue; 1310 } 1311 field = is->curform->fields; 1312 if(field->ftype != Fselect) { 1313 if(warn) 1314 fprint(2, "warning: <OPTION> not in <SELECT>\n"); 1315 continue; 1316 } 1317 val = aval(tok, Avalue); 1318 option = newoption(aflagval(tok, Aselected), val, nil, field->options); 1319 field->options = option; 1320 option->display = getpcdata(toks, tokslen, &toki); 1321 if(val == nil) 1322 option->value = _Strdup(option->display); 1323 break; 1324 1325 // <!ELEMENT P - O (%text)* > 1326 case Tp: 1327 pushjust(ps, atabval(tok, Aalign, align_tab, NALIGNTAB, ps->curjust)); 1328 ps->inpar = 1; 1329 ps->skipwhite = 1; 1330 break; 1331 1332 case Tp+RBRA: 1333 break; 1334 1335 // <!ELEMENT PARAM - O EMPTY> 1336 // Do something when we do applets... 1337 case Tparam: 1338 break; 1339 1340 // <!ELEMENT PRE - - (%text)* -(IMG|BIG|SMALL|SUB|SUP|FONT) > 1341 case Tpre: 1342 ps->curstate &= ~IFwrap; 1343 ps->literal = 1; 1344 ps->skipwhite = 0; 1345 pushfontstyle(ps, FntT); 1346 break; 1347 1348 case Tpre+RBRA: 1349 ps->curstate |= IFwrap; 1350 if(ps->literal) { 1351 popfontstyle(ps); 1352 ps->literal = 0; 1353 } 1354 break; 1355 1356 // <!ELEMENT SCRIPT - - CDATA> 1357 case Tscript: 1358 if(doscripts) { 1359 if(!di->hasscripts) { 1360 if(di->scripttype == TextJavascript) { 1361 // TODO: initialize script if nec. 1362 // initjscript(di); 1363 di->hasscripts = 1; 1364 } 1365 } 1366 } 1367 if(!di->hasscripts) { 1368 if(warn) 1369 fprint(2, "warning: <SCRIPT> ignored\n"); 1370 ps->skipping = 1; 1371 } 1372 else { 1373 scriptsrc = aurlval(tok, Asrc, nil, di->base); 1374 script = nil; 1375 if(scriptsrc != nil) { 1376 if(warn) 1377 fprint(2, "warning: non-local <SCRIPT> ignored\n"); 1378 free(scriptsrc); 1379 } 1380 else { 1381 script = getpcdata(toks, tokslen, &toki); 1382 } 1383 if(script != nil) { 1384 if(warn) 1385 fprint(2, "script ignored\n"); 1386 free(script); 1387 } 1388 } 1389 break; 1390 1391 case Tscript+RBRA: 1392 ps->skipping = 0; 1393 break; 1394 1395 // <!ELEMENT SELECT - - (OPTION+)> 1396 case Tselect: 1397 if(is->curform == nil) { 1398 if(warn) 1399 fprint(2, "<SELECT> not inside <FORM>\n"); 1400 continue; 1401 } 1402 field = newformfield(Fselect, 1403 ++is->curform->nfields, 1404 is->curform, 1405 aval(tok, Aname), 1406 nil, 1407 auintval(tok, Asize, 0), 1408 0, 1409 is->curform->fields); 1410 is->curform->fields = field; 1411 if(aflagval(tok, Amultiple)) 1412 field->flags = FFmultiple; 1413 ffit = newiformfield(field); 1414 additem(ps, ffit, tok); 1415 if(ffit->genattr != nil) 1416 field->events = ffit->genattr->events; 1417 // throw away stuff until next tag (should be <OPTION>) 1418 s = getpcdata(toks, tokslen, &toki); 1419 if(s != nil) 1420 free(s); 1421 break; 1422 1423 case Tselect+RBRA: 1424 if(is->curform == nil || is->curform->fields == nil) { 1425 if(warn) 1426 fprint(2, "warning: unexpected </SELECT>\n"); 1427 continue; 1428 } 1429 field = is->curform->fields; 1430 if(field->ftype != Fselect) 1431 continue; 1432 // put options back in input order 1433 field->options = (Option*)_revlist((List*)field->options); 1434 break; 1435 1436 // <!ELEMENT (STRIKE|U) - - (%text)*> 1437 case Tstrike: 1438 case Tu: 1439 ps->curul = push(&ps->ulstk, (tag==Tstrike)? ULmid : ULunder); 1440 break; 1441 1442 case Tstrike+RBRA: 1443 case Tu+RBRA: 1444 if(ps->ulstk.n == 0) { 1445 if(warn) 1446 fprint(2, "warning: unexpected %T\n", tok); 1447 continue; 1448 } 1449 ps->curul = popretnewtop(&ps->ulstk, ULnone); 1450 break; 1451 1452 // <!ELEMENT STYLE - - CDATA> 1453 case Tstyle: 1454 if(warn) 1455 fprint(2, "warning: unimplemented <STYLE>\n"); 1456 ps->skipping = 1; 1457 break; 1458 1459 case Tstyle+RBRA: 1460 ps->skipping = 0; 1461 break; 1462 1463 // <!ELEMENT (SUB|SUP) - - (%text)*> 1464 case Tsub: 1465 case Tsup: 1466 if(tag == Tsub) 1467 ps->curvoff += SUBOFF; 1468 else 1469 ps->curvoff -= SUPOFF; 1470 push(&ps->voffstk, ps->curvoff); 1471 sz = top(&ps->fntsizestk, Normal); 1472 pushfontsize(ps, sz - 1); 1473 break; 1474 1475 case Tsub+RBRA: 1476 case Tsup+RBRA: 1477 if(ps->voffstk.n == 0) { 1478 if(warn) 1479 fprint(2, "warning: unexpected %T\n", tok); 1480 continue; 1481 } 1482 ps->curvoff = popretnewtop(&ps->voffstk, 0); 1483 popfontsize(ps); 1484 break; 1485 1486 // <!ELEMENT TABLE - - (CAPTION?, TR+)> 1487 case Ttable: 1488 ps->skipwhite = 0; 1489 tab = newtable(++is->ntables, 1490 aalign(tok), 1491 adimen(tok, Awidth), 1492 aflagval(tok, Aborder), 1493 auintval(tok, Acellspacing, TABSP), 1494 auintval(tok, Acellpadding, TABPAD), 1495 makebackground(nil, acolorval(tok, Abgcolor, ps->curbg.color)), 1496 tok, 1497 is->tabstk); 1498 is->tabstk = tab; 1499 curtab = tab; 1500 break; 1501 1502 case Ttable+RBRA: 1503 if(curtab == nil) { 1504 if(warn) 1505 fprint(2, "warning: unexpected </TABLE>\n"); 1506 continue; 1507 } 1508 isempty = (curtab->cells == nil); 1509 if(isempty) { 1510 if(warn) 1511 fprint(2, "warning: <TABLE> has no cells\n"); 1512 } 1513 else { 1514 ps = finishcell(curtab, ps); 1515 if(curtab->rows != nil) 1516 curtab->rows->flags = 0; 1517 finish_table(curtab); 1518 } 1519 ps->skipping = 0; 1520 if(!isempty) { 1521 tabitem = newitable(curtab); 1522 al = curtab->align.halign; 1523 switch(al) { 1524 case ALleft: 1525 case ALright: 1526 additem(ps, newifloat(tabitem, al), tok); 1527 break; 1528 default: 1529 if(al == ALcenter) 1530 pushjust(ps, ALcenter); 1531 addbrk(ps, 0, 0); 1532 if(ps->inpar) { 1533 popjust(ps); 1534 ps->inpar = 0; 1535 } 1536 additem(ps, tabitem, curtab->tabletok); 1537 if(al == ALcenter) 1538 popjust(ps); 1539 break; 1540 } 1541 } 1542 if(is->tabstk == nil) { 1543 if(warn) 1544 fprint(2, "warning: table stack is wrong\n"); 1545 } 1546 else 1547 is->tabstk = is->tabstk->next; 1548 curtab->next = di->tables; 1549 di->tables = curtab; 1550 curtab = is->tabstk; 1551 if(!isempty) 1552 addbrk(ps, 0, 0); 1553 break; 1554 1555 // <!ELEMENT (TH|TD) - O %body.content> 1556 // Cells for a row are accumulated in reverse order. 1557 // We push ps on a stack, and use a new one to accumulate 1558 // the contents of the cell. 1559 case Ttd: 1560 case Tth: 1561 if(curtab == nil) { 1562 if(warn) 1563 fprint(2, "%T outside <TABLE>\n", tok); 1564 continue; 1565 } 1566 if(ps->inpar) { 1567 popjust(ps); 1568 ps->inpar = 0; 1569 } 1570 ps = finishcell(curtab, ps); 1571 tr = nil; 1572 if(curtab->rows != nil) 1573 tr = curtab->rows; 1574 if(tr == nil || !tr->flags) { 1575 if(warn) 1576 fprint(2, "%T outside row\n", tok); 1577 tr = newtablerow(makealign(ALnone, ALnone), 1578 makebackground(nil, curtab->background.color), 1579 TFparsing, 1580 curtab->rows); 1581 curtab->rows = tr; 1582 } 1583 ps = cell_pstate(ps, tag == Tth); 1584 flags = TFparsing; 1585 if(aflagval(tok, Anowrap)) { 1586 flags |= TFnowrap; 1587 ps->curstate &= ~IFwrap; 1588 } 1589 if(tag == Tth) 1590 flags |= TFisth; 1591 c = newtablecell(curtab->cells==nil? 1 : curtab->cells->cellid+1, 1592 auintval(tok, Arowspan, 1), 1593 auintval(tok, Acolspan, 1), 1594 aalign(tok), 1595 adimen(tok, Awidth), 1596 auintval(tok, Aheight, 0), 1597 makebackground(nil, acolorval(tok, Abgcolor, tr->background.color)), 1598 flags, 1599 curtab->cells); 1600 curtab->cells = c; 1601 ps->curbg = c->background; 1602 if(c->align.halign == ALnone) { 1603 if(tr->align.halign != ALnone) 1604 c->align.halign = tr->align.halign; 1605 else if(tag == Tth) 1606 c->align.halign = ALcenter; 1607 else 1608 c->align.halign = ALleft; 1609 } 1610 if(c->align.valign == ALnone) { 1611 if(tr->align.valign != ALnone) 1612 c->align.valign = tr->align.valign; 1613 else 1614 c->align.valign = ALmiddle; 1615 } 1616 c->nextinrow = tr->cells; 1617 tr->cells = c; 1618 break; 1619 1620 case Ttd+RBRA: 1621 case Tth+RBRA: 1622 if(curtab == nil || curtab->cells == nil) { 1623 if(warn) 1624 fprint(2, "unexpected %T\n", tok); 1625 continue; 1626 } 1627 ps = finishcell(curtab, ps); 1628 break; 1629 1630 // <!ELEMENT TEXTAREA - - ( //PCDATA)> 1631 case Ttextarea: 1632 if(is->curform == nil) { 1633 if(warn) 1634 fprint(2, "<TEXTAREA> not inside <FORM>\n"); 1635 continue; 1636 } 1637 field = newformfield(Ftextarea, 1638 ++is->curform->nfields, 1639 is->curform, 1640 aval(tok, Aname), 1641 nil, 1642 0, 1643 0, 1644 is->curform->fields); 1645 is->curform->fields = field; 1646 field->rows = auintval(tok, Arows, 3); 1647 field->cols = auintval(tok, Acols, 50); 1648 field->value = getpcdata(toks, tokslen, &toki); 1649 if(warn && toki < tokslen - 1 && toks[toki + 1].tag != Ttextarea + RBRA) 1650 fprint(2, "warning: <TEXTAREA> data ended by %T\n", &toks[toki + 1]); 1651 ffit = newiformfield(field); 1652 additem(ps, ffit, tok); 1653 if(ffit->genattr != nil) 1654 field->events = ffit->genattr->events; 1655 break; 1656 1657 // <!ELEMENT TITLE - - ( //PCDATA)* -(%head.misc)> 1658 case Ttitle: 1659 di->doctitle = getpcdata(toks, tokslen, &toki); 1660 if(warn && toki < tokslen - 1 && toks[toki + 1].tag != Ttitle + RBRA) 1661 fprint(2, "warning: <TITLE> data ended by %T\n", &toks[toki + 1]); 1662 break; 1663 1664 // <!ELEMENT TR - O (TH|TD)+> 1665 // rows are accumulated in reverse order in curtab->rows 1666 case Ttr: 1667 if(curtab == nil) { 1668 if(warn) 1669 fprint(2, "warning: <TR> outside <TABLE>\n"); 1670 continue; 1671 } 1672 if(ps->inpar) { 1673 popjust(ps); 1674 ps->inpar = 0; 1675 } 1676 ps = finishcell(curtab, ps); 1677 if(curtab->rows != nil) 1678 curtab->rows->flags = 0; 1679 curtab->rows = newtablerow(aalign(tok), 1680 makebackground(nil, acolorval(tok, Abgcolor, curtab->background.color)), 1681 TFparsing, 1682 curtab->rows); 1683 break; 1684 1685 case Ttr+RBRA: 1686 if(curtab == nil || curtab->rows == nil) { 1687 if(warn) 1688 fprint(2, "warning: unexpected </TR>\n"); 1689 continue; 1690 } 1691 ps = finishcell(curtab, ps); 1692 tr = curtab->rows; 1693 if(tr->cells == nil) { 1694 if(warn) 1695 fprint(2, "warning: empty row\n"); 1696 curtab->rows = tr->next; 1697 tr->next = nil; 1698 } 1699 else 1700 tr->flags = 0; 1701 break; 1702 1703 // <!ELEMENT (TT|CODE|KBD|SAMP) - - (%text)*> 1704 case Ttt: 1705 case Tcode: 1706 case Tkbd: 1707 case Tsamp: 1708 pushfontstyle(ps, FntT); 1709 break; 1710 1711 // Tags that have empty action 1712 case Tabbr: 1713 case Tabbr+RBRA: 1714 case Tacronym: 1715 case Tacronym+RBRA: 1716 case Tarea+RBRA: 1717 case Tbase+RBRA: 1718 case Tbasefont+RBRA: 1719 case Tbr+RBRA: 1720 case Tdd+RBRA: 1721 case Tdt+RBRA: 1722 case Tframe+RBRA: 1723 case Thr+RBRA: 1724 case Thtml: 1725 case Thtml+RBRA: 1726 case Timg+RBRA: 1727 case Tinput+RBRA: 1728 case Tisindex+RBRA: 1729 case Tli+RBRA: 1730 case Tlink: 1731 case Tlink+RBRA: 1732 case Tmeta+RBRA: 1733 case Toption+RBRA: 1734 case Tparam+RBRA: 1735 case Ttextarea+RBRA: 1736 case Ttitle+RBRA: 1737 break; 1738 1739 1740 // Tags not implemented 1741 case Tbdo: 1742 case Tbdo+RBRA: 1743 case Tbutton: 1744 case Tbutton+RBRA: 1745 case Tdel: 1746 case Tdel+RBRA: 1747 case Tfieldset: 1748 case Tfieldset+RBRA: 1749 case Tiframe: 1750 case Tiframe+RBRA: 1751 case Tins: 1752 case Tins+RBRA: 1753 case Tlabel: 1754 case Tlabel+RBRA: 1755 case Tlegend: 1756 case Tlegend+RBRA: 1757 case Tobject: 1758 case Tobject+RBRA: 1759 case Toptgroup: 1760 case Toptgroup+RBRA: 1761 case Tspan: 1762 case Tspan+RBRA: 1763 if(warn) { 1764 if(tag > RBRA) 1765 tag -= RBRA; 1766 fprint(2, "warning: unimplemented HTML tag: %S\n", tagnames[tag]); 1767 } 1768 break; 1769 1770 default: 1771 if(warn) 1772 fprint(2, "warning: unknown HTML tag: %S\n", tok->text); 1773 break; 1774 } 1775 } 1776 // some pages omit trailing </table> 1777 while(curtab != nil) { 1778 if(warn) 1779 fprint(2, "warning: <TABLE> not closed\n"); 1780 if(curtab->cells != nil) { 1781 ps = finishcell(curtab, ps); 1782 if(curtab->cells == nil) { 1783 if(warn) 1784 fprint(2, "warning: empty table\n"); 1785 } 1786 else { 1787 if(curtab->rows != nil) 1788 curtab->rows->flags = 0; 1789 finish_table(curtab); 1790 ps->skipping = 0; 1791 additem(ps, newitable(curtab), curtab->tabletok); 1792 addbrk(ps, 0, 0); 1793 } 1794 } 1795 if(is->tabstk != nil) 1796 is->tabstk = is->tabstk->next; 1797 curtab->next = di->tables; 1798 di->tables = curtab; 1799 curtab = is->tabstk; 1800 } 1801 outerps = lastps(ps); 1802 ans = outerps->items->next; 1803 // note: ans may be nil and di->kids not nil, if there's a frameset! 1804 outerps->items = newispacer(ISPnull); 1805 outerps->lastit = outerps->items; 1806 is->psstk = ps; 1807 if(ans != nil && di->hasscripts) { 1808 // TODO evalscript(nil); 1809 ; 1810 } 1811 1812 return_ans: 1813 if(dbgbuild) { 1814 assert(validitems(ans)); 1815 if(ans == nil) 1816 fprint(2, "getitems returning nil\n"); 1817 else 1818 printitems(ans, "getitems returning:"); 1819 } 1820 return ans; 1821 } 1822 1823 // Concatenate together maximal set of Data tokens, starting at toks[toki+1]. 1824 // Lexer has ensured that there will either be a following non-data token or 1825 // we will be at eof. 1826 // Return emallocd trimmed concatenation, and update *ptoki to last used toki 1827 static Rune* 1828 getpcdata(Token* toks, int tokslen, int* ptoki) 1829 { 1830 Rune* ans; 1831 Rune* p; 1832 Rune* trimans; 1833 int anslen; 1834 int trimanslen; 1835 int toki; 1836 Token* tok; 1837 1838 ans = nil; 1839 anslen = 0; 1840 // first find length of answer 1841 toki = (*ptoki) + 1; 1842 while(toki < tokslen) { 1843 tok = &toks[toki]; 1844 if(tok->tag == Data) { 1845 toki++; 1846 anslen += _Strlen(tok->text); 1847 } 1848 else 1849 break; 1850 } 1851 // now make up the initial answer 1852 if(anslen > 0) { 1853 ans = _newstr(anslen); 1854 p = ans; 1855 toki = (*ptoki) + 1; 1856 while(toki < tokslen) { 1857 tok = &toks[toki]; 1858 if(tok->tag == Data) { 1859 toki++; 1860 p = _Stradd(p, tok->text, _Strlen(tok->text)); 1861 } 1862 else 1863 break; 1864 } 1865 *p = 0; 1866 _trimwhite(ans, anslen, &trimans, &trimanslen); 1867 if(trimanslen != anslen) { 1868 p = ans; 1869 ans = _Strndup(trimans, trimanslen); 1870 free(p); 1871 } 1872 } 1873 *ptoki = toki-1; 1874 return ans; 1875 } 1876 1877 // If still parsing head of curtab->cells list, finish it off 1878 // by transferring the items on the head of psstk to the cell. 1879 // Then pop the psstk and return the new psstk. 1880 static Pstate* 1881 finishcell(Table* curtab, Pstate* psstk) 1882 { 1883 Tablecell* c; 1884 Pstate* psstknext; 1885 1886 c = curtab->cells; 1887 if(c != nil) { 1888 if((c->flags&TFparsing)) { 1889 psstknext = psstk->next; 1890 if(psstknext == nil) { 1891 if(warn) 1892 fprint(2, "warning: parse state stack is wrong\n"); 1893 } 1894 else { 1895 c->content = psstk->items->next; 1896 c->flags &= ~TFparsing; 1897 freepstate(psstk); 1898 psstk = psstknext; 1899 } 1900 } 1901 } 1902 return psstk; 1903 } 1904 1905 // Make a new Pstate for a cell, based on the old pstate, oldps. 1906 // Also, put the new ps on the head of the oldps stack. 1907 static Pstate* 1908 cell_pstate(Pstate* oldps, int ishead) 1909 { 1910 Pstate* ps; 1911 int sty; 1912 1913 ps = newpstate(oldps); 1914 ps->skipwhite = 1; 1915 ps->curanchor = oldps->curanchor; 1916 copystack(&ps->fntstylestk, &oldps->fntstylestk); 1917 copystack(&ps->fntsizestk, &oldps->fntsizestk); 1918 ps->curfont = oldps->curfont; 1919 ps->curfg = oldps->curfg; 1920 ps->curbg = oldps->curbg; 1921 copystack(&ps->fgstk, &oldps->fgstk); 1922 ps->adjsize = oldps->adjsize; 1923 if(ishead) { 1924 sty = ps->curfont%NumSize; 1925 ps->curfont = FntB*NumSize + sty; 1926 } 1927 return ps; 1928 } 1929 1930 // Return a new Pstate with default starting state. 1931 // Use link to add it to head of a list, if any. 1932 static Pstate* 1933 newpstate(Pstate* link) 1934 { 1935 Pstate* ps; 1936 1937 ps = (Pstate*)emalloc(sizeof(Pstate)); 1938 ps->curfont = DefFnt; 1939 ps->curfg = Black; 1940 ps->curbg.image = nil; 1941 ps->curbg.color = White; 1942 ps->curul = ULnone; 1943 ps->curjust = ALleft; 1944 ps->curstate = IFwrap; 1945 ps->items = newispacer(ISPnull); 1946 ps->lastit = ps->items; 1947 ps->prelastit = nil; 1948 ps->next = link; 1949 return ps; 1950 } 1951 1952 // Return last Pstate on psl list 1953 static Pstate* 1954 lastps(Pstate* psl) 1955 { 1956 assert(psl != nil); 1957 while(psl->next != nil) 1958 psl = psl->next; 1959 return psl; 1960 } 1961 1962 // Add it to end of ps item chain, adding in current state from ps. 1963 // Also, if tok is not nil, scan it for generic attributes and assign 1964 // the genattr field of the item accordingly. 1965 static void 1966 additem(Pstate* ps, Item* it, Token* tok) 1967 { 1968 int aid; 1969 int any; 1970 Rune* i; 1971 Rune* c; 1972 Rune* s; 1973 Rune* t; 1974 Attr* a; 1975 SEvent* e; 1976 1977 if(ps->skipping) { 1978 if(warn) 1979 fprint(2, "warning: skipping item: %I\n", it); 1980 return; 1981 } 1982 it->anchorid = ps->curanchor; 1983 it->state |= ps->curstate; 1984 if(tok != nil) { 1985 any = 0; 1986 i = nil; 1987 c = nil; 1988 s = nil; 1989 t = nil; 1990 e = nil; 1991 for(a = tok->attr; a != nil; a = a->next) { 1992 aid = a->attid; 1993 if(!attrinfo[aid]) 1994 continue; 1995 switch(aid) { 1996 case Aid: 1997 i = a->value; 1998 break; 1999 2000 case Aclass: 2001 c = a->value; 2002 break; 2003 2004 case Astyle: 2005 s = a->value; 2006 break; 2007 2008 case Atitle: 2009 t = a->value; 2010 break; 2011 2012 default: 2013 assert(aid >= Aonblur && aid <= Aonunload); 2014 e = newscriptevent(scriptev[a->attid], a->value, e); 2015 break; 2016 } 2017 a->value = nil; 2018 any = 1; 2019 } 2020 if(any) 2021 it->genattr = newgenattr(i, c, s, t, e); 2022 } 2023 ps->curstate &= ~(IFbrk|IFbrksp|IFnobrk|IFcleft|IFcright); 2024 ps->prelastit = ps->lastit; 2025 ps->lastit->next = it; 2026 ps->lastit = it; 2027 } 2028 2029 // Make a text item out of s, 2030 // using current font, foreground, vertical offset and underline state. 2031 static Item* 2032 textit(Pstate* ps, Rune* s) 2033 { 2034 assert(s != nil); 2035 return newitext(s, ps->curfont, ps->curfg, ps->curvoff + Voffbias, ps->curul); 2036 } 2037 2038 // Add text item or items for s, paying attention to 2039 // current font, foreground, baseline offset, underline state, 2040 // and literal mode. Unless we're in literal mode, compress 2041 // whitespace to single blank, and, if curstate has a break, 2042 // trim any leading whitespace. Whether in literal mode or not, 2043 // turn nonbreaking spaces into spacer items with IFnobrk set. 2044 // 2045 // In literal mode, break up s at newlines and add breaks instead. 2046 // Also replace tabs appropriate number of spaces. 2047 // In nonliteral mode, break up the items every 100 or so characters 2048 // just to make the layout algorithm not go quadratic. 2049 // 2050 // addtext assumes ownership of s. 2051 static void 2052 addtext(Pstate* ps, Rune* s) 2053 { 2054 int n; 2055 int i; 2056 int j; 2057 int k; 2058 int col; 2059 int c; 2060 int nsp; 2061 Item* it; 2062 Rune* ss; 2063 Rune* p; 2064 Rune buf[SMALLBUFSIZE]; 2065 2066 assert(s != nil); 2067 n = runestrlen(s); 2068 i = 0; 2069 j = 0; 2070 if(ps->literal) { 2071 col = 0; 2072 while(i < n) { 2073 if(s[i] == '\n') { 2074 if(i > j) { 2075 // trim trailing blanks from line 2076 for(k = i; k > j; k--) 2077 if(s[k - 1] != ' ') 2078 break; 2079 if(k > j) 2080 additem(ps, textit(ps, _Strndup(s+j, k-j)), nil); 2081 } 2082 addlinebrk(ps, 0); 2083 j = i + 1; 2084 col = 0; 2085 } 2086 else { 2087 if(s[i] == '\t') { 2088 col += i - j; 2089 nsp = 8 - (col%8); 2090 // make ss = s[j:i] + nsp spaces 2091 ss = _newstr(i-j+nsp); 2092 p = _Stradd(ss, s+j, i-j); 2093 p = _Stradd(p, L" ", nsp); 2094 *p = 0; 2095 additem(ps, textit(ps, ss), nil); 2096 col += nsp; 2097 j = i + 1; 2098 } 2099 else if(s[i] == NBSP) { 2100 if(i > j) 2101 additem(ps, textit(ps, _Strndup(s+j, i-j)), nil); 2102 addnbsp(ps); 2103 col += (i - j) + 1; 2104 j = i + 1; 2105 } 2106 } 2107 i++; 2108 } 2109 if(i > j) { 2110 if(j == 0 && i == n) { 2111 // just transfer s over 2112 additem(ps, textit(ps, s), nil); 2113 } 2114 else { 2115 additem(ps, textit(ps, _Strndup(s+j, i-j)), nil); 2116 free(s); 2117 } 2118 } 2119 } 2120 else { // not literal mode 2121 if((ps->curstate&IFbrk) || ps->lastit == ps->items) 2122 while(i < n) { 2123 c = s[i]; 2124 if(c >= 256 || !isspace(c)) 2125 break; 2126 i++; 2127 } 2128 p = buf; 2129 for(j = i; i < n; i++) { 2130 assert(p+i-j < buf+SMALLBUFSIZE-1); 2131 c = s[i]; 2132 if(c == NBSP) { 2133 if(i > j) 2134 p = _Stradd(p, s+j, i-j); 2135 if(p > buf) 2136 additem(ps, textit(ps, _Strndup(buf, p-buf)), nil); 2137 p = buf; 2138 addnbsp(ps); 2139 j = i + 1; 2140 continue; 2141 } 2142 if(c < 256 && isspace(c)) { 2143 if(i > j) 2144 p = _Stradd(p, s+j, i-j); 2145 *p++ = ' '; 2146 while(i < n - 1) { 2147 c = s[i + 1]; 2148 if(c >= 256 || !isspace(c)) 2149 break; 2150 i++; 2151 } 2152 j = i + 1; 2153 } 2154 if(i - j >= 100) { 2155 p = _Stradd(p, s+j, i+1-j); 2156 j = i + 1; 2157 } 2158 if(p-buf >= 100) { 2159 additem(ps, textit(ps, _Strndup(buf, p-buf)), nil); 2160 p = buf; 2161 } 2162 } 2163 if(i > j && j < n) { 2164 assert(p+i-j < buf+SMALLBUFSIZE-1); 2165 p = _Stradd(p, s+j, i-j); 2166 } 2167 // don't add a space if previous item ended in a space 2168 if(p-buf == 1 && buf[0] == ' ' && ps->lastit != nil) { 2169 it = ps->lastit; 2170 if(it->tag == Itexttag) { 2171 ss = ((Itext*)it)->s; 2172 k = _Strlen(ss); 2173 if(k > 0 && ss[k] == ' ') 2174 p = buf; 2175 } 2176 } 2177 if(p > buf) 2178 additem(ps, textit(ps, _Strndup(buf, p-buf)), nil); 2179 free(s); 2180 } 2181 } 2182 2183 // Add a break to ps->curstate, with extra space if sp is true. 2184 // If there was a previous break, combine this one's parameters 2185 // with that to make the amt be the max of the two and the clr 2186 // be the most general. (amt will be 0 or 1) 2187 // Also, if the immediately preceding item was a text item, 2188 // trim any whitespace from the end of it, if not in literal mode. 2189 // Finally, if this is at the very beginning of the item list 2190 // (the only thing there is a null spacer), then don't add the space. 2191 static void 2192 addbrk(Pstate* ps, int sp, int clr) 2193 { 2194 int state; 2195 Rune* l; 2196 int nl; 2197 Rune* r; 2198 int nr; 2199 Itext* t; 2200 Rune* s; 2201 2202 state = ps->curstate; 2203 clr = clr|(state&(IFcleft|IFcright)); 2204 if(sp && !(ps->lastit == ps->items)) 2205 sp = IFbrksp; 2206 else 2207 sp = 0; 2208 ps->curstate = IFbrk|sp|(state&~(IFcleft|IFcright))|clr; 2209 if(ps->lastit != ps->items) { 2210 if(!ps->literal && ps->lastit->tag == Itexttag) { 2211 t = (Itext*)ps->lastit; 2212 _splitr(t->s, _Strlen(t->s), notwhitespace, &l, &nl, &r, &nr); 2213 // try to avoid making empty items 2214 // but not crucial f the occasional one gets through 2215 if(nl == 0 && ps->prelastit != nil) { 2216 ps->lastit = ps->prelastit; 2217 ps->lastit->next = nil; 2218 ps->prelastit = nil; 2219 } 2220 else { 2221 s = t->s; 2222 if(nl == 0) { 2223 // need a non-nil pointer to empty string 2224 // (_Strdup(L"") returns nil) 2225 t->s = emalloc(sizeof(Rune)); 2226 t->s[0] = 0; 2227 } 2228 else 2229 t->s = _Strndup(l, nl); 2230 if(s) 2231 free(s); 2232 } 2233 } 2234 } 2235 } 2236 2237 // Add break due to a <br> or a newline within a preformatted section. 2238 // We add a null item first, with current font's height and ascent, to make 2239 // sure that the current line takes up at least that amount of vertical space. 2240 // This ensures that <br>s on empty lines cause blank lines, and that 2241 // multiple <br>s in a row give multiple blank lines. 2242 // However don't add the spacer if the previous item was something that 2243 // takes up space itself. 2244 static void 2245 addlinebrk(Pstate* ps, int clr) 2246 { 2247 int obrkstate; 2248 int b; 2249 int addit; 2250 2251 // don't want break before our null item unless the previous item 2252 // was also a null item for the purposes of line breaking 2253 obrkstate = ps->curstate&(IFbrk|IFbrksp); 2254 b = IFnobrk; 2255 addit = 0; 2256 if(ps->lastit != nil) { 2257 if(ps->lastit->tag == Ispacertag) { 2258 if(((Ispacer*)ps->lastit)->spkind == ISPvline) 2259 b = IFbrk; 2260 addit = 1; 2261 } 2262 else if(ps->lastit->tag == Ifloattag) 2263 addit = 1; 2264 } 2265 if(addit) { 2266 ps->curstate = (ps->curstate&~(IFbrk|IFbrksp))|b; 2267 additem(ps, newispacer(ISPvline), nil); 2268 ps->curstate = (ps->curstate&~(IFbrk|IFbrksp))|obrkstate; 2269 } 2270 addbrk(ps, 0, clr); 2271 } 2272 2273 // Add a nonbreakable space 2274 static void 2275 addnbsp(Pstate* ps) 2276 { 2277 // if nbsp comes right where a break was specified, 2278 // do the break anyway (nbsp is being used to generate undiscardable 2279 // space rather than to prevent a break) 2280 if((ps->curstate&IFbrk) == 0) 2281 ps->curstate |= IFnobrk; 2282 additem(ps, newispacer(ISPhspace), nil); 2283 // but definitely no break on next item 2284 ps->curstate |= IFnobrk; 2285 } 2286 2287 // Change hang in ps.curstate by delta. 2288 // The amount is in 1/10ths of tabs, and is the amount that 2289 // the current contiguous set of items with a hang value set 2290 // is to be shifted left from its normal (indented) place. 2291 static void 2292 changehang(Pstate* ps, int delta) 2293 { 2294 int amt; 2295 2296 amt = (ps->curstate&IFhangmask) + delta; 2297 if(amt < 0) { 2298 if(warn) 2299 fprint(2, "warning: hang went negative\n"); 2300 amt = 0; 2301 } 2302 ps->curstate = (ps->curstate&~IFhangmask)|amt; 2303 } 2304 2305 // Change indent in ps.curstate by delta. 2306 static void 2307 changeindent(Pstate* ps, int delta) 2308 { 2309 int amt; 2310 2311 amt = ((ps->curstate&IFindentmask) >> IFindentshift) + delta; 2312 if(amt < 0) { 2313 if(warn) 2314 fprint(2, "warning: indent went negative\n"); 2315 amt = 0; 2316 } 2317 ps->curstate = (ps->curstate&~IFindentmask)|(amt << IFindentshift); 2318 } 2319 2320 // Push val on top of stack, and also return value pushed 2321 static int 2322 push(Stack* stk, int val) 2323 { 2324 if(stk->n == Nestmax) { 2325 if(warn) 2326 fprint(2, "warning: build stack overflow\n"); 2327 } 2328 else 2329 stk->slots[stk->n++] = val; 2330 return val; 2331 } 2332 2333 // Pop top of stack 2334 static void 2335 pop(Stack* stk) 2336 { 2337 if(stk->n > 0) 2338 --stk->n; 2339 } 2340 2341 //Return top of stack, using dflt if stack is empty 2342 static int 2343 top(Stack* stk, int dflt) 2344 { 2345 if(stk->n == 0) 2346 return dflt; 2347 return stk->slots[stk->n-1]; 2348 } 2349 2350 // pop, then return new top, with dflt if empty 2351 static int 2352 popretnewtop(Stack* stk, int dflt) 2353 { 2354 if(stk->n == 0) 2355 return dflt; 2356 stk->n--; 2357 if(stk->n == 0) 2358 return dflt; 2359 return stk->slots[stk->n-1]; 2360 } 2361 2362 // Copy fromstk entries into tostk 2363 static void 2364 copystack(Stack* tostk, Stack* fromstk) 2365 { 2366 int n; 2367 2368 n = fromstk->n; 2369 tostk->n = n; 2370 memmove(tostk->slots, fromstk->slots, n*sizeof(int)); 2371 } 2372 2373 static void 2374 popfontstyle(Pstate* ps) 2375 { 2376 pop(&ps->fntstylestk); 2377 setcurfont(ps); 2378 } 2379 2380 static void 2381 pushfontstyle(Pstate* ps, int sty) 2382 { 2383 push(&ps->fntstylestk, sty); 2384 setcurfont(ps); 2385 } 2386 2387 static void 2388 popfontsize(Pstate* ps) 2389 { 2390 pop(&ps->fntsizestk); 2391 setcurfont(ps); 2392 } 2393 2394 static void 2395 pushfontsize(Pstate* ps, int sz) 2396 { 2397 push(&ps->fntsizestk, sz); 2398 setcurfont(ps); 2399 } 2400 2401 static void 2402 setcurfont(Pstate* ps) 2403 { 2404 int sty; 2405 int sz; 2406 2407 sty = top(&ps->fntstylestk, FntR); 2408 sz = top(&ps->fntsizestk, Normal); 2409 if(sz < Tiny) 2410 sz = Tiny; 2411 if(sz > Verylarge) 2412 sz = Verylarge; 2413 ps->curfont = sty*NumSize + sz; 2414 } 2415 2416 static void 2417 popjust(Pstate* ps) 2418 { 2419 pop(&ps->juststk); 2420 setcurjust(ps); 2421 } 2422 2423 static void 2424 pushjust(Pstate* ps, int j) 2425 { 2426 push(&ps->juststk, j); 2427 setcurjust(ps); 2428 } 2429 2430 static void 2431 setcurjust(Pstate* ps) 2432 { 2433 int j; 2434 int state; 2435 2436 j = top(&ps->juststk, ALleft); 2437 if(j != ps->curjust) { 2438 ps->curjust = j; 2439 state = ps->curstate; 2440 state &= ~(IFrjust|IFcjust); 2441 if(j == ALcenter) 2442 state |= IFcjust; 2443 else if(j == ALright) 2444 state |= IFrjust; 2445 ps->curstate = state; 2446 } 2447 } 2448 2449 // Do final rearrangement after table parsing is finished 2450 // and assign cells to grid points 2451 static void 2452 finish_table(Table* t) 2453 { 2454 int ncol; 2455 int nrow; 2456 int r; 2457 Tablerow* rl; 2458 Tablecell* cl; 2459 int* rowspancnt; 2460 Tablecell** rowspancell; 2461 int ri; 2462 int ci; 2463 Tablecell* c; 2464 Tablecell* cnext; 2465 Tablerow* row; 2466 Tablerow* rownext; 2467 int rcols; 2468 int newncol; 2469 int k; 2470 int j; 2471 int cspan; 2472 int rspan; 2473 int i; 2474 2475 rl = t->rows; 2476 t->nrow = nrow = _listlen((List*)rl); 2477 t->rows = (Tablerow*)emalloc(nrow * sizeof(Tablerow)); 2478 ncol = 0; 2479 r = nrow - 1; 2480 for(row = rl; row != nil; row = rownext) { 2481 // copy the data from the allocated Tablerow into the array slot 2482 t->rows[r] = *row; 2483 rownext = row->next; 2484 row = &t->rows[r]; 2485 r--; 2486 rcols = 0; 2487 c = row->cells; 2488 2489 // If rowspan is > 1 but this is the last row, 2490 // reset the rowspan 2491 if(c != nil && c->rowspan > 1 && r == nrow-2) 2492 c->rowspan = 1; 2493 2494 // reverse row->cells list (along nextinrow pointers) 2495 row->cells = nil; 2496 while(c != nil) { 2497 cnext = c->nextinrow; 2498 c->nextinrow = row->cells; 2499 row->cells = c; 2500 rcols += c->colspan; 2501 c = cnext; 2502 } 2503 if(rcols > ncol) 2504 ncol = rcols; 2505 } 2506 t->ncol = ncol; 2507 t->cols = (Tablecol*)emalloc(ncol * sizeof(Tablecol)); 2508 2509 // Reverse cells just so they are drawn in source order. 2510 // Also, trim their contents so they don't end in whitespace. 2511 t->cells = (Tablecell*)_revlist((List*)t->cells); 2512 for(c = t->cells; c != nil; c= c->next) 2513 trim_cell(c); 2514 t->grid = (Tablecell***)emalloc(nrow * sizeof(Tablecell**)); 2515 for(i = 0; i < nrow; i++) 2516 t->grid[i] = (Tablecell**)emalloc(ncol * sizeof(Tablecell*)); 2517 2518 // The following arrays keep track of cells that are spanning 2519 // multiple rows; rowspancnt[i] is the number of rows left 2520 // to be spanned in column i. 2521 // When done, cell's (row,col) is upper left grid point. 2522 rowspancnt = (int*)emalloc(ncol * sizeof(int)); 2523 rowspancell = (Tablecell**)emalloc(ncol * sizeof(Tablecell*)); 2524 for(ri = 0; ri < nrow; ri++) { 2525 row = &t->rows[ri]; 2526 cl = row->cells; 2527 ci = 0; 2528 while(ci < ncol || cl != nil) { 2529 if(ci < ncol && rowspancnt[ci] > 0) { 2530 t->grid[ri][ci] = rowspancell[ci]; 2531 rowspancnt[ci]--; 2532 ci++; 2533 } 2534 else { 2535 if(cl == nil) { 2536 ci++; 2537 continue; 2538 } 2539 c = cl; 2540 cl = cl->nextinrow; 2541 cspan = c->colspan; 2542 rspan = c->rowspan; 2543 if(ci + cspan > ncol) { 2544 // because of row spanning, we calculated 2545 // ncol incorrectly; adjust it 2546 newncol = ci + cspan; 2547 t->cols = (Tablecol*)erealloc(t->cols, newncol * sizeof(Tablecol)); 2548 rowspancnt = (int*)erealloc(rowspancnt, newncol * sizeof(int)); 2549 rowspancell = (Tablecell**)erealloc(rowspancell, newncol * sizeof(Tablecell*)); 2550 k = newncol-ncol; 2551 memset(t->cols+ncol, 0, k*sizeof(Tablecol)); 2552 memset(rowspancnt+ncol, 0, k*sizeof(int)); 2553 memset(rowspancell+ncol, 0, k*sizeof(Tablecell*)); 2554 for(j = 0; j < nrow; j++) { 2555 t->grid[j] = (Tablecell**)erealloc(t->grid[j], newncol * sizeof(Tablecell*)); 2556 memset(t->grid[j], 0, k*sizeof(Tablecell*)); 2557 } 2558 t->ncol = ncol = newncol; 2559 } 2560 c->row = ri; 2561 c->col = ci; 2562 for(i = 0; i < cspan; i++) { 2563 t->grid[ri][ci] = c; 2564 if(rspan > 1) { 2565 rowspancnt[ci] = rspan - 1; 2566 rowspancell[ci] = c; 2567 } 2568 ci++; 2569 } 2570 } 2571 } 2572 } 2573 } 2574 2575 // Remove tail of cell content until it isn't whitespace. 2576 static void 2577 trim_cell(Tablecell* c) 2578 { 2579 int dropping; 2580 Rune* s; 2581 Rune* x; 2582 Rune* y; 2583 int nx; 2584 int ny; 2585 Item* p; 2586 Itext* q; 2587 Item* pprev; 2588 2589 dropping = 1; 2590 while(c->content != nil && dropping) { 2591 p = c->content; 2592 pprev = nil; 2593 while(p->next != nil) { 2594 pprev = p; 2595 p = p->next; 2596 } 2597 dropping = 0; 2598 if(!(p->state&IFnobrk)) { 2599 if(p->tag == Itexttag) { 2600 q = (Itext*)p; 2601 s = q->s; 2602 _splitr(s, _Strlen(s), notwhitespace, &x, &nx, &y, &ny); 2603 if(nx != 0 && ny != 0) { 2604 q->s = _Strndup(x, nx); 2605 free(s); 2606 } 2607 break; 2608 } 2609 } 2610 if(dropping) { 2611 if(pprev == nil) 2612 c->content = nil; 2613 else 2614 pprev->next = nil; 2615 freeitem(p); 2616 } 2617 } 2618 } 2619 2620 // Caller must free answer (eventually). 2621 static Rune* 2622 listmark(uchar ty, int n) 2623 { 2624 Rune* s; 2625 Rune* t; 2626 int n2; 2627 int i; 2628 2629 s = nil; 2630 switch(ty) { 2631 case LTdisc: 2632 case LTsquare: 2633 case LTcircle: 2634 s = _newstr(1); 2635 s[0] = (ty == LTdisc)? 0x2022 // bullet 2636 : ((ty == LTsquare)? 0x220e // filled square 2637 : 0x2218); // degree 2638 s[1] = 0; 2639 break; 2640 2641 case LT1: 2642 t = _ltoStr(n); 2643 n2 = _Strlen(t); 2644 s = _newstr(n2+1); 2645 t = _Stradd(s, t, n2); 2646 *t++ = '.'; 2647 *t = 0; 2648 break; 2649 2650 case LTa: 2651 case LTA: 2652 n--; 2653 i = 0; 2654 if(n < 0) 2655 n = 0; 2656 s = _newstr((n <= 25)? 2 : 3); 2657 if(n > 25) { 2658 n2 = n%26; 2659 n /= 26; 2660 if(n2 > 25) 2661 n2 = 25; 2662 s[i++] = n2 + (ty == LTa)? 'a' : 'A'; 2663 } 2664 s[i++] = n + (ty == LTa)? 'a' : 'A'; 2665 s[i++] = '.'; 2666 s[i] = 0; 2667 break; 2668 2669 case LTi: 2670 case LTI: 2671 if(n >= NROMAN) { 2672 if(warn) 2673 fprint(2, "warning: unimplemented roman number > %d\n", NROMAN); 2674 n = NROMAN; 2675 } 2676 t = roman[n - 1]; 2677 n2 = _Strlen(t); 2678 s = _newstr(n2+1); 2679 for(i = 0; i < n2; i++) 2680 s[i] = (ty == LTi)? tolower(t[i]) : t[i]; 2681 s[i++] = '.'; 2682 s[i] = 0; 2683 break; 2684 } 2685 return s; 2686 } 2687 2688 // Find map with given name in di.maps. 2689 // If not there, add one, copying name. 2690 // Ownership of map remains with di->maps list. 2691 static Map* 2692 getmap(Docinfo* di, Rune* name) 2693 { 2694 Map* m; 2695 2696 for(m = di->maps; m != nil; m = m->next) { 2697 if(!_Strcmp(name, m->name)) 2698 return m; 2699 } 2700 m = (Map*)emalloc(sizeof(Map)); 2701 m->name = _Strdup(name); 2702 m->areas = nil; 2703 m->next = di->maps; 2704 di->maps = m; 2705 return m; 2706 } 2707 2708 // Transfers ownership of href to Area 2709 static Area* 2710 newarea(int shape, Rune* href, int target, Area* link) 2711 { 2712 Area* a; 2713 2714 a = (Area*)emalloc(sizeof(Area)); 2715 a->shape = shape; 2716 a->href = href; 2717 a->target = target; 2718 a->next = link; 2719 return a; 2720 } 2721 2722 // Return string value associated with attid in tok, nil if none. 2723 // Caller must free the result (eventually). 2724 static Rune* 2725 aval(Token* tok, int attid) 2726 { 2727 Rune* ans; 2728 2729 _tokaval(tok, attid, &ans, 1); // transfers string ownership from token to ans 2730 return ans; 2731 } 2732 2733 // Like aval, but use dflt if there was no such attribute in tok. 2734 // Caller must free the result (eventually). 2735 static Rune* 2736 astrval(Token* tok, int attid, Rune* dflt) 2737 { 2738 Rune* ans; 2739 2740 if(_tokaval(tok, attid, &ans, 1)) 2741 return ans; // transfers string ownership from token to ans 2742 else 2743 return _Strdup(dflt); 2744 } 2745 2746 // Here we're supposed to convert to an int, 2747 // and have a default when not found 2748 static int 2749 aintval(Token* tok, int attid, int dflt) 2750 { 2751 Rune* ans; 2752 2753 if(!_tokaval(tok, attid, &ans, 0) || ans == nil) 2754 return dflt; 2755 else 2756 return toint(ans); 2757 } 2758 2759 // Like aintval, but result should be >= 0 2760 static int 2761 auintval(Token* tok, int attid, int dflt) 2762 { 2763 Rune* ans; 2764 int v; 2765 2766 if(!_tokaval(tok, attid, &ans, 0) || ans == nil) 2767 return dflt; 2768 else { 2769 v = toint(ans); 2770 return v >= 0? v : 0; 2771 } 2772 } 2773 2774 // int conversion, but with possible error check (if warning) 2775 static int 2776 toint(Rune* s) 2777 { 2778 int ans; 2779 Rune* eptr; 2780 2781 ans = _Strtol(s, &eptr, 10); 2782 if(warn) { 2783 if(*eptr != 0) { 2784 eptr = _Strclass(eptr, notwhitespace); 2785 if(eptr != nil) 2786 fprint(2, "warning: expected integer, got %S\n", s); 2787 } 2788 } 2789 return ans; 2790 } 2791 2792 // Attribute value when need a table to convert strings to ints 2793 static int 2794 atabval(Token* tok, int attid, StringInt* tab, int ntab, int dflt) 2795 { 2796 Rune* aval; 2797 int ans; 2798 2799 ans = dflt; 2800 if(_tokaval(tok, attid, &aval, 0)) { 2801 if(!_lookup(tab, ntab, aval, _Strlen(aval), &ans)) { 2802 ans = dflt; 2803 if(warn) 2804 fprint(2, "warning: name not found in table lookup: %S\n", aval); 2805 } 2806 } 2807 return ans; 2808 } 2809 2810 // Attribute value when supposed to be a color 2811 static int 2812 acolorval(Token* tok, int attid, int dflt) 2813 { 2814 Rune* aval; 2815 int ans; 2816 2817 ans = dflt; 2818 if(_tokaval(tok, attid, &aval, 0)) 2819 ans = color(aval, dflt); 2820 return ans; 2821 } 2822 2823 // Attribute value when supposed to be a target frame name 2824 static int 2825 atargval(Token* tok, int dflt) 2826 { 2827 int ans; 2828 Rune* aval; 2829 2830 ans = dflt; 2831 if(_tokaval(tok, Atarget, &aval, 0)){ 2832 ans = targetid(aval); 2833 } 2834 return ans; 2835 } 2836 2837 // special for list types, where "i" and "I" are different, 2838 // but "square" and "SQUARE" are the same 2839 static int 2840 listtyval(Token* tok, int dflt) 2841 { 2842 Rune* aval; 2843 int ans; 2844 int n; 2845 2846 ans = dflt; 2847 if(_tokaval(tok, Atype, &aval, 0)) { 2848 n = _Strlen(aval); 2849 if(n == 1) { 2850 switch(aval[0]) { 2851 case '1': 2852 ans = LT1; 2853 break; 2854 case 'A': 2855 ans = LTA; 2856 break; 2857 case 'I': 2858 ans = LTI; 2859 break; 2860 case 'a': 2861 ans = LTa; 2862 break; 2863 case 'i': 2864 ans = LTi; 2865 default: 2866 if(warn) 2867 fprint(2, "warning: unknown list element type %c\n", aval[0]); 2868 } 2869 } 2870 else { 2871 if(!_Strncmpci(aval, n, L"circle")) 2872 ans = LTcircle; 2873 else if(!_Strncmpci(aval, n, L"disc")) 2874 ans = LTdisc; 2875 else if(!_Strncmpci(aval, n, L"square")) 2876 ans = LTsquare; 2877 else { 2878 if(warn) 2879 fprint(2, "warning: unknown list element type %S\n", aval); 2880 } 2881 } 2882 } 2883 return ans; 2884 } 2885 2886 // Attribute value when value is a URL, possibly relative to base. 2887 // FOR NOW: leave the url relative. 2888 // Caller must free the result (eventually). 2889 static Rune* 2890 aurlval(Token* tok, int attid, Rune* dflt, Rune* base) 2891 { 2892 Rune* ans; 2893 Rune* url; 2894 2895 USED(base); 2896 ans = nil; 2897 if(_tokaval(tok, attid, &url, 0) && url != nil) 2898 ans = removeallwhite(url); 2899 if(ans == nil) 2900 ans = _Strdup(dflt); 2901 return ans; 2902 } 2903 2904 // Return copy of s but with all whitespace (even internal) removed. 2905 // This fixes some buggy URL specification strings. 2906 static Rune* 2907 removeallwhite(Rune* s) 2908 { 2909 int j; 2910 int n; 2911 int i; 2912 int c; 2913 Rune* ans; 2914 2915 j = 0; 2916 n = _Strlen(s); 2917 for(i = 0; i < n; i++) { 2918 c = s[i]; 2919 if(c >= 256 || !isspace(c)) 2920 j++; 2921 } 2922 if(j < n) { 2923 ans = _newstr(j); 2924 j = 0; 2925 for(i = 0; i < n; i++) { 2926 c = s[i]; 2927 if(c >= 256 || !isspace(c)) 2928 ans[j++] = c; 2929 } 2930 ans[j] = 0; 2931 } 2932 else 2933 ans = _Strdup(s); 2934 return ans; 2935 } 2936 2937 // Attribute value when mere presence of attr implies value of 1, 2938 // but if there is an integer there, return it as the value. 2939 static int 2940 aflagval(Token* tok, int attid) 2941 { 2942 int val; 2943 Rune* sval; 2944 2945 val = 0; 2946 if(_tokaval(tok, attid, &sval, 0)) { 2947 val = 1; 2948 if(sval != nil) 2949 val = toint(sval); 2950 } 2951 return val; 2952 } 2953 2954 static Align 2955 makealign(int halign, int valign) 2956 { 2957 Align al; 2958 2959 al.halign = halign; 2960 al.valign = valign; 2961 return al; 2962 } 2963 2964 // Make an Align (two alignments, horizontal and vertical) 2965 static Align 2966 aalign(Token* tok) 2967 { 2968 return makealign( 2969 atabval(tok, Aalign, align_tab, NALIGNTAB, ALnone), 2970 atabval(tok, Avalign, align_tab, NALIGNTAB, ALnone)); 2971 } 2972 2973 // Make a Dimen, based on value of attid attr 2974 static Dimen 2975 adimen(Token* tok, int attid) 2976 { 2977 Rune* wd; 2978 2979 if(_tokaval(tok, attid, &wd, 0)) 2980 return parsedim(wd, _Strlen(wd)); 2981 else 2982 return makedimen(Dnone, 0); 2983 } 2984 2985 // Parse s[0:n] as num[.[num]][unit][%|*] 2986 static Dimen 2987 parsedim(Rune* s, int ns) 2988 { 2989 int kind; 2990 int spec; 2991 Rune* l; 2992 int nl; 2993 Rune* r; 2994 int nr; 2995 int mul; 2996 int i; 2997 Rune* f; 2998 int nf; 2999 int Tkdpi; 3000 Rune* units; 3001 3002 kind = Dnone; 3003 spec = 0; 3004 _splitl(s, ns, L"^0-9", &l, &nl, &r, &nr); 3005 if(nl != 0) { 3006 spec = 1000*_Strtol(l, nil, 10); 3007 if(nr > 0 && r[0] == '.') { 3008 _splitl(r+1, nr-1, L"^0-9", &f, &nf, &r, &nr); 3009 if(nf != 0) { 3010 mul = 100; 3011 for(i = 0; i < nf; i++) { 3012 spec = spec + mul*(f[i]-'0'); 3013 mul = mul/10; 3014 } 3015 } 3016 } 3017 kind = Dpixels; 3018 if(nr != 0) { 3019 if(nr >= 2) { 3020 Tkdpi = 100; 3021 units = r; 3022 r = r+2; 3023 nr -= 2; 3024 if(!_Strncmpci(units, 2, L"pt")) 3025 spec = (spec*Tkdpi)/72; 3026 else if(!_Strncmpci(units, 2, L"pi")) 3027 spec = (spec*12*Tkdpi)/72; 3028 else if(!_Strncmpci(units, 2, L"in")) 3029 spec = spec*Tkdpi; 3030 else if(!_Strncmpci(units, 2, L"cm")) 3031 spec = (spec*100*Tkdpi)/254; 3032 else if(!_Strncmpci(units, 2, L"mm")) 3033 spec = (spec*10*Tkdpi)/254; 3034 else if(!_Strncmpci(units, 2, L"em")) 3035 spec = spec*15; 3036 else { 3037 if(warn) 3038 fprint(2, "warning: unknown units %C%Cs\n", units[0], units[1]); 3039 } 3040 } 3041 if(nr >= 1) { 3042 if(r[0] == '%') 3043 kind = Dpercent; 3044 else if(r[0] == '*') 3045 kind = Drelative; 3046 } 3047 } 3048 spec = spec/1000; 3049 } 3050 else if(nr == 1 && r[0] == '*') { 3051 spec = 1; 3052 kind = Drelative; 3053 } 3054 return makedimen(kind, spec); 3055 } 3056 3057 static void 3058 setdimarray(Token* tok, int attid, Dimen** pans, int* panslen) 3059 { 3060 Rune* s; 3061 Dimen* d; 3062 int k; 3063 int nc; 3064 Rune* a[SMALLBUFSIZE]; 3065 int an[SMALLBUFSIZE]; 3066 3067 if(_tokaval(tok, attid, &s, 0)) { 3068 nc = _splitall(s, _Strlen(s), L", ", a, an, SMALLBUFSIZE); 3069 if(nc > 0) { 3070 d = (Dimen*)emalloc(nc * sizeof(Dimen)); 3071 for(k = 0; k < nc; k++) { 3072 d[k] = parsedim(a[k], an[k]); 3073 } 3074 *pans = d; 3075 *panslen = nc; 3076 return; 3077 } 3078 } 3079 *pans = nil; 3080 *panslen = 0; 3081 } 3082 3083 static Background 3084 makebackground(Rune* imageurl, int color) 3085 { 3086 Background bg; 3087 3088 bg.image = imageurl; 3089 bg.color = color; 3090 return bg; 3091 } 3092 3093 static Item* 3094 newitext(Rune* s, int fnt, int fg, int voff, int ul) 3095 { 3096 Itext* t; 3097 3098 assert(s != nil); 3099 t = (Itext*)emalloc(sizeof(Itext)); 3100 t->tag = Itexttag; 3101 t->s = s; 3102 t->fnt = fnt; 3103 t->fg = fg; 3104 t->voff = voff; 3105 t->ul = ul; 3106 return (Item*)t; 3107 } 3108 3109 static Item* 3110 newirule(int align, int size, int noshade, Dimen wspec) 3111 { 3112 Irule* r; 3113 3114 r = (Irule*)emalloc(sizeof(Irule)); 3115 r->tag = Iruletag; 3116 r->align = align; 3117 r->size = size; 3118 r->noshade = noshade; 3119 r->wspec = wspec; 3120 return (Item*)r; 3121 } 3122 3123 // Map is owned elsewhere. 3124 static Item* 3125 newiimage(Rune* src, Rune* altrep, int align, int width, int height, 3126 int hspace, int vspace, int border, int ismap, Map* map) 3127 { 3128 Iimage* i; 3129 int state; 3130 3131 state = 0; 3132 if(ismap) 3133 state = IFsmap; 3134 i = (Iimage*)emalloc(sizeof(Iimage)); 3135 i->tag = Iimagetag; 3136 i->state = state; 3137 i->imsrc = src; 3138 i->altrep = altrep; 3139 i->align = align; 3140 i->imwidth = width; 3141 i->imheight = height; 3142 i->hspace = hspace; 3143 i->vspace = vspace; 3144 i->border = border; 3145 i->map = map; 3146 i->ctlid = -1; 3147 return (Item*)i; 3148 } 3149 3150 static Item* 3151 newiformfield(Formfield* ff) 3152 { 3153 Iformfield* f; 3154 3155 f = (Iformfield*)emalloc(sizeof(Iformfield)); 3156 f->tag = Iformfieldtag; 3157 f->formfield = ff; 3158 return (Item*)f; 3159 } 3160 3161 static Item* 3162 newitable(Table* tab) 3163 { 3164 Itable* t; 3165 3166 t = (Itable*)emalloc(sizeof(Itable)); 3167 t->tag = Itabletag; 3168 t->table = tab; 3169 return (Item*)t; 3170 } 3171 3172 static Item* 3173 newifloat(Item* it, int side) 3174 { 3175 Ifloat* f; 3176 3177 f = (Ifloat*)emalloc(sizeof(Ifloat)); 3178 f->tag = Ifloattag; 3179 f->state = IFwrap; 3180 f->item = it; 3181 f->side = side; 3182 return (Item*)f; 3183 } 3184 3185 static Item* 3186 newispacer(int spkind) 3187 { 3188 Ispacer* s; 3189 3190 s = (Ispacer*)emalloc(sizeof(Ispacer)); 3191 s->tag = Ispacertag; 3192 s->spkind = spkind; 3193 return (Item*)s; 3194 } 3195 3196 // Free one item (caller must deal with next pointer) 3197 static void 3198 freeitem(Item* it) 3199 { 3200 Iimage* ii; 3201 Genattr* ga; 3202 3203 if(it == nil) 3204 return; 3205 3206 switch(it->tag) { 3207 case Itexttag: 3208 free(((Itext*)it)->s); 3209 break; 3210 case Iimagetag: 3211 ii = (Iimage*)it; 3212 free(ii->imsrc); 3213 free(ii->altrep); 3214 break; 3215 case Iformfieldtag: 3216 freeformfield(((Iformfield*)it)->formfield); 3217 break; 3218 case Itabletag: 3219 freetable(((Itable*)it)->table); 3220 break; 3221 case Ifloattag: 3222 freeitem(((Ifloat*)it)->item); 3223 break; 3224 } 3225 ga = it->genattr; 3226 if(ga != nil) { 3227 free(ga->id); 3228 free(ga->class); 3229 free(ga->style); 3230 free(ga->title); 3231 freescriptevents(ga->events); 3232 } 3233 free(it); 3234 } 3235 3236 // Free list of items chained through next pointer 3237 void 3238 freeitems(Item* ithead) 3239 { 3240 Item* it; 3241 Item* itnext; 3242 3243 it = ithead; 3244 while(it != nil) { 3245 itnext = it->next; 3246 freeitem(it); 3247 it = itnext; 3248 } 3249 } 3250 3251 static void 3252 freeformfield(Formfield* ff) 3253 { 3254 Option* o; 3255 Option* onext; 3256 3257 if(ff == nil) 3258 return; 3259 3260 free(ff->name); 3261 free(ff->value); 3262 for(o = ff->options; o != nil; o = onext) { 3263 onext = o->next; 3264 free(o->value); 3265 free(o->display); 3266 } 3267 free(ff); 3268 } 3269 3270 static void 3271 freetable(Table* t) 3272 { 3273 int i; 3274 Tablecell* c; 3275 Tablecell* cnext; 3276 3277 if(t == nil) 3278 return; 3279 3280 // We'll find all the unique cells via t->cells and next pointers. 3281 // (Other pointers to cells in the table are duplicates of these) 3282 for(c = t->cells; c != nil; c = cnext) { 3283 cnext = c->next; 3284 freeitems(c->content); 3285 } 3286 if(t->grid != nil) { 3287 for(i = 0; i < t->nrow; i++) 3288 free(t->grid[i]); 3289 free(t->grid); 3290 } 3291 free(t->rows); 3292 free(t->cols); 3293 freeitems(t->caption); 3294 free(t); 3295 } 3296 3297 static void 3298 freeform(Form* f) 3299 { 3300 if(f == nil) 3301 return; 3302 3303 free(f->name); 3304 free(f->action); 3305 // Form doesn't own its fields (Iformfield items do) 3306 free(f); 3307 } 3308 3309 static void 3310 freeforms(Form* fhead) 3311 { 3312 Form* f; 3313 Form* fnext; 3314 3315 for(f = fhead; f != nil; f = fnext) { 3316 fnext = f->next; 3317 freeform(f); 3318 } 3319 } 3320 3321 static void 3322 freeanchor(Anchor* a) 3323 { 3324 if(a == nil) 3325 return; 3326 3327 free(a->name); 3328 free(a->href); 3329 free(a); 3330 } 3331 3332 static void 3333 freeanchors(Anchor* ahead) 3334 { 3335 Anchor* a; 3336 Anchor* anext; 3337 3338 for(a = ahead; a != nil; a = anext) { 3339 anext = a->next; 3340 freeanchor(a); 3341 } 3342 } 3343 3344 static void 3345 freedestanchor(DestAnchor* da) 3346 { 3347 if(da == nil) 3348 return; 3349 3350 free(da->name); 3351 free(da); 3352 } 3353 3354 static void 3355 freedestanchors(DestAnchor* dahead) 3356 { 3357 DestAnchor* da; 3358 DestAnchor* danext; 3359 3360 for(da = dahead; da != nil; da = danext) { 3361 danext = da->next; 3362 freedestanchor(da); 3363 } 3364 } 3365 3366 static void 3367 freearea(Area* a) 3368 { 3369 if(a == nil) 3370 return; 3371 free(a->href); 3372 free(a->coords); 3373 } 3374 3375 static void freekidinfos(Kidinfo* khead); 3376 3377 static void 3378 freekidinfo(Kidinfo* k) 3379 { 3380 if(k->isframeset) { 3381 free(k->rows); 3382 free(k->cols); 3383 freekidinfos(k->kidinfos); 3384 } 3385 else { 3386 free(k->src); 3387 free(k->name); 3388 } 3389 free(k); 3390 } 3391 3392 static void 3393 freekidinfos(Kidinfo* khead) 3394 { 3395 Kidinfo* k; 3396 Kidinfo* knext; 3397 3398 for(k = khead; k != nil; k = knext) { 3399 knext = k->next; 3400 freekidinfo(k); 3401 } 3402 } 3403 3404 static void 3405 freemap(Map* m) 3406 { 3407 Area* a; 3408 Area* anext; 3409 3410 if(m == nil) 3411 return; 3412 3413 free(m->name); 3414 for(a = m->areas; a != nil; a = anext) { 3415 anext = a->next; 3416 freearea(a); 3417 } 3418 free(m); 3419 } 3420 3421 static void 3422 freemaps(Map* mhead) 3423 { 3424 Map* m; 3425 Map* mnext; 3426 3427 for(m = mhead; m != nil; m = mnext) { 3428 mnext = m->next; 3429 freemap(m); 3430 } 3431 } 3432 3433 void 3434 freedocinfo(Docinfo* d) 3435 { 3436 if(d == nil) 3437 return; 3438 free(d->src); 3439 free(d->base); 3440 freeitem((Item*)d->backgrounditem); 3441 free(d->refresh); 3442 freekidinfos(d->kidinfo); 3443 freeanchors(d->anchors); 3444 freedestanchors(d->dests); 3445 freeforms(d->forms); 3446 freemaps(d->maps); 3447 // tables, images, and formfields are freed when 3448 // the items pointing at them are freed 3449 free(d); 3450 } 3451 3452 // Currently, someone else owns all the memory 3453 // pointed to by things in a Pstate. 3454 static void 3455 freepstate(Pstate* p) 3456 { 3457 free(p); 3458 } 3459 3460 static void 3461 freepstatestack(Pstate* pshead) 3462 { 3463 Pstate* p; 3464 Pstate* pnext; 3465 3466 for(p = pshead; p != nil; p = pnext) { 3467 pnext = p->next; 3468 free(p); 3469 } 3470 } 3471 3472 static int 3473 Iconv(Fmt *f) 3474 { 3475 Item* it; 3476 Itext* t; 3477 Irule* r; 3478 Iimage* i; 3479 Ifloat* fl; 3480 int state; 3481 Formfield* ff; 3482 Rune* ty; 3483 Tablecell* c; 3484 Table* tab; 3485 char* p; 3486 int cl; 3487 int hang; 3488 int indent; 3489 int bi; 3490 int nbuf; 3491 char buf[BIGBUFSIZE]; 3492 3493 it = va_arg(f->args, Item*); 3494 bi = 0; 3495 nbuf = sizeof(buf); 3496 state = it->state; 3497 nbuf = nbuf-1; 3498 if(state&IFbrk) { 3499 cl = state&(IFcleft|IFcright); 3500 p = ""; 3501 if(cl) { 3502 if(cl == (IFcleft|IFcright)) 3503 p = " both"; 3504 else if(cl == IFcleft) 3505 p = " left"; 3506 else 3507 p = " right"; 3508 } 3509 bi = snprint(buf, nbuf, "brk(%d%s)", (state&IFbrksp)? 1 : 0, p); 3510 } 3511 if(state&IFnobrk) 3512 bi += snprint(buf+bi, nbuf-bi, " nobrk"); 3513 if(!(state&IFwrap)) 3514 bi += snprint(buf+bi, nbuf-bi, " nowrap"); 3515 if(state&IFrjust) 3516 bi += snprint(buf+bi, nbuf-bi, " rjust"); 3517 if(state&IFcjust) 3518 bi += snprint(buf+bi, nbuf-bi, " cjust"); 3519 if(state&IFsmap) 3520 bi += snprint(buf+bi, nbuf-bi, " smap"); 3521 indent = (state&IFindentmask) >> IFindentshift; 3522 if(indent > 0) 3523 bi += snprint(buf+bi, nbuf-bi, " indent=%d", indent); 3524 hang = state&IFhangmask; 3525 if(hang > 0) 3526 bi += snprint(buf+bi, nbuf-bi, " hang=%d", hang); 3527 3528 switch(it->tag) { 3529 case Itexttag: 3530 t = (Itext*)it; 3531 bi += snprint(buf+bi, nbuf-bi, " Text '%S', fnt=%d, fg=%x", t->s, t->fnt, t->fg); 3532 break; 3533 3534 case Iruletag: 3535 r = (Irule*)it; 3536 bi += snprint(buf+bi, nbuf-bi, "Rule size=%d, al=%S, wspec=", r->size, stringalign(r->align)); 3537 bi += dimprint(buf+bi, nbuf-bi, r->wspec); 3538 break; 3539 3540 case Iimagetag: 3541 i = (Iimage*)it; 3542 bi += snprint(buf+bi, nbuf-bi, 3543 "Image src=%S, alt=%S, al=%S, w=%d, h=%d hsp=%d, vsp=%d, bd=%d, map=%S", 3544 i->imsrc, i->altrep? i->altrep : L"", stringalign(i->align), i->imwidth, i->imheight, 3545 i->hspace, i->vspace, i->border, i->map? i->map->name : L""); 3546 break; 3547 3548 case Iformfieldtag: 3549 ff = ((Iformfield*)it)->formfield; 3550 if(ff->ftype == Ftextarea) 3551 ty = L"textarea"; 3552 else if(ff->ftype == Fselect) 3553 ty = L"select"; 3554 else { 3555 ty = _revlookup(input_tab, NINPUTTAB, ff->ftype); 3556 if(ty == nil) 3557 ty = L"none"; 3558 } 3559 bi += snprint(buf+bi, nbuf-bi, "Formfield %S, fieldid=%d, formid=%d, name=%S, value=%S", 3560 ty, ff->fieldid, ff->form->formid, ff->name? ff->name : L"", 3561 ff->value? ff->value : L""); 3562 break; 3563 3564 case Itabletag: 3565 tab = ((Itable*)it)->table; 3566 bi += snprint(buf+bi, nbuf-bi, "Table tableid=%d, width=", tab->tableid); 3567 bi += dimprint(buf+bi, nbuf-bi, tab->width); 3568 bi += snprint(buf+bi, nbuf-bi, ", nrow=%d, ncol=%d, ncell=%d, totw=%d, toth=%d\n", 3569 tab->nrow, tab->ncol, tab->ncell, tab->totw, tab->toth); 3570 for(c = tab->cells; c != nil; c = c->next) 3571 bi += snprint(buf+bi, nbuf-bi, "Cell %d.%d, at (%d,%d) ", 3572 tab->tableid, c->cellid, c->row, c->col); 3573 bi += snprint(buf+bi, nbuf-bi, "End of Table %d", tab->tableid); 3574 break; 3575 3576 case Ifloattag: 3577 fl = (Ifloat*)it; 3578 bi += snprint(buf+bi, nbuf-bi, "Float, x=%d y=%d, side=%S, it=%I", 3579 fl->x, fl->y, stringalign(fl->side), fl->item); 3580 bi += snprint(buf+bi, nbuf-bi, "\n\t"); 3581 break; 3582 3583 case Ispacertag: 3584 p = ""; 3585 switch(((Ispacer*)it)->spkind) { 3586 case ISPnull: 3587 p = "null"; 3588 break; 3589 case ISPvline: 3590 p = "vline"; 3591 break; 3592 case ISPhspace: 3593 p = "hspace"; 3594 break; 3595 } 3596 bi += snprint(buf+bi, nbuf-bi, "Spacer %s ", p); 3597 break; 3598 } 3599 bi += snprint(buf+bi, nbuf-bi, " w=%d, h=%d, a=%d, anchor=%d\n", 3600 it->width, it->height, it->ascent, it->anchorid); 3601 buf[bi] = 0; 3602 return fmtstrcpy(f, buf); 3603 } 3604 3605 // String version of alignment 'a' 3606 static Rune* 3607 stringalign(int a) 3608 { 3609 Rune* s; 3610 3611 s = _revlookup(align_tab, NALIGNTAB, a); 3612 if(s == nil) 3613 s = L"none"; 3614 return s; 3615 } 3616 3617 // Put at most nbuf chars of representation of d into buf, 3618 // and return number of characters put 3619 static int 3620 dimprint(char* buf, int nbuf, Dimen d) 3621 { 3622 int n; 3623 int k; 3624 3625 n = 0; 3626 n += snprint(buf, nbuf, "%d", dimenspec(d)); 3627 k = dimenkind(d); 3628 if(k == Dpercent) 3629 buf[n++] = '%'; 3630 if(k == Drelative) 3631 buf[n++] = '*'; 3632 return n; 3633 } 3634 3635 void 3636 printitems(Item* items, char* msg) 3637 { 3638 Item* il; 3639 3640 fprint(2, "%s\n", msg); 3641 il = items; 3642 while(il != nil) { 3643 fprint(2, "%I", il); 3644 il = il->next; 3645 } 3646 } 3647 3648 static Genattr* 3649 newgenattr(Rune* id, Rune* class, Rune* style, Rune* title, SEvent* events) 3650 { 3651 Genattr* g; 3652 3653 g = (Genattr*)emalloc(sizeof(Genattr)); 3654 g->id = id; 3655 g->class = class; 3656 g->style = style; 3657 g->title = title; 3658 g->events = events; 3659 return g; 3660 } 3661 3662 static Formfield* 3663 newformfield(int ftype, int fieldid, Form* form, Rune* name, 3664 Rune* value, int size, int maxlength, Formfield* link) 3665 { 3666 Formfield* ff; 3667 3668 ff = (Formfield*)emalloc(sizeof(Formfield)); 3669 ff->ftype = ftype; 3670 ff->fieldid = fieldid; 3671 ff->form = form; 3672 ff->name = name; 3673 ff->value = value; 3674 ff->size = size; 3675 ff->maxlength = maxlength; 3676 ff->ctlid = -1; 3677 ff->next = link; 3678 return ff; 3679 } 3680 3681 // Transfers ownership of value and display to Option. 3682 static Option* 3683 newoption(int selected, Rune* value, Rune* display, Option* link) 3684 { 3685 Option *o; 3686 3687 o = (Option*)emalloc(sizeof(Option)); 3688 o->selected = selected; 3689 o->value = value; 3690 o->display = display; 3691 o->next = link; 3692 return o; 3693 } 3694 3695 static Form* 3696 newform(int formid, Rune* name, Rune* action, int target, int method, Form* link) 3697 { 3698 Form* f; 3699 3700 f = (Form*)emalloc(sizeof(Form)); 3701 f->formid = formid; 3702 f->name = name; 3703 f->action = action; 3704 f->target = target; 3705 f->method = method; 3706 f->nfields = 0; 3707 f->fields = nil; 3708 f->next = link; 3709 return f; 3710 } 3711 3712 static Table* 3713 newtable(int tableid, Align align, Dimen width, int border, 3714 int cellspacing, int cellpadding, Background bg, Token* tok, Table* link) 3715 { 3716 Table* t; 3717 3718 t = (Table*)emalloc(sizeof(Table)); 3719 t->tableid = tableid; 3720 t->align = align; 3721 t->width = width; 3722 t->border = border; 3723 t->cellspacing = cellspacing; 3724 t->cellpadding = cellpadding; 3725 t->background = bg; 3726 t->caption_place = ALbottom; 3727 t->caption_lay = nil; 3728 t->tabletok = tok; 3729 t->tabletok = nil; 3730 t->next = link; 3731 return t; 3732 } 3733 3734 static Tablerow* 3735 newtablerow(Align align, Background bg, int flags, Tablerow* link) 3736 { 3737 Tablerow* tr; 3738 3739 tr = (Tablerow*)emalloc(sizeof(Tablerow)); 3740 tr->align = align; 3741 tr->background = bg; 3742 tr->flags = flags; 3743 tr->next = link; 3744 return tr; 3745 } 3746 3747 static Tablecell* 3748 newtablecell(int cellid, int rowspan, int colspan, Align align, Dimen wspec, int hspec, 3749 Background bg, int flags, Tablecell* link) 3750 { 3751 Tablecell* c; 3752 3753 c = (Tablecell*)emalloc(sizeof(Tablecell)); 3754 c->cellid = cellid; 3755 c->lay = nil; 3756 c->rowspan = rowspan; 3757 c->colspan = colspan; 3758 c->align = align; 3759 c->flags = flags; 3760 c->wspec = wspec; 3761 c->hspec = hspec; 3762 c->background = bg; 3763 c->next = link; 3764 return c; 3765 } 3766 3767 static Anchor* 3768 newanchor(int index, Rune* name, Rune* href, int target, Anchor* link) 3769 { 3770 Anchor* a; 3771 3772 a = (Anchor*)emalloc(sizeof(Anchor)); 3773 a->index = index; 3774 a->name = name; 3775 a->href = href; 3776 a->target = target; 3777 a->next = link; 3778 return a; 3779 } 3780 3781 static DestAnchor* 3782 newdestanchor(int index, Rune* name, Item* item, DestAnchor* link) 3783 { 3784 DestAnchor* d; 3785 3786 d = (DestAnchor*)emalloc(sizeof(DestAnchor)); 3787 d->index = index; 3788 d->name = name; 3789 d->item = item; 3790 d->next = link; 3791 return d; 3792 } 3793 3794 static SEvent* 3795 newscriptevent(int type, Rune* script, SEvent* link) 3796 { 3797 SEvent* ans; 3798 3799 ans = (SEvent*)emalloc(sizeof(SEvent)); 3800 ans->type = type; 3801 ans->script = script; 3802 ans->next = link; 3803 return ans; 3804 } 3805 3806 static void 3807 freescriptevents(SEvent* ehead) 3808 { 3809 SEvent* e; 3810 SEvent* nexte; 3811 3812 e = ehead; 3813 while(e != nil) { 3814 nexte = e->next; 3815 free(e->script); 3816 free(e); 3817 e = nexte; 3818 } 3819 } 3820 3821 static Dimen 3822 makedimen(int kind, int spec) 3823 { 3824 Dimen d; 3825 3826 if(spec&Dkindmask) { 3827 if(warn) 3828 fprint(2, "warning: dimension spec too big: %d\n", spec); 3829 spec = 0; 3830 } 3831 d.kindspec = kind|spec; 3832 return d; 3833 } 3834 3835 int 3836 dimenkind(Dimen d) 3837 { 3838 return (d.kindspec&Dkindmask); 3839 } 3840 3841 int 3842 dimenspec(Dimen d) 3843 { 3844 return (d.kindspec&Dspecmask); 3845 } 3846 3847 static Kidinfo* 3848 newkidinfo(int isframeset, Kidinfo* link) 3849 { 3850 Kidinfo* ki; 3851 3852 ki = (Kidinfo*)emalloc(sizeof(Kidinfo)); 3853 ki->isframeset = isframeset; 3854 if(!isframeset) { 3855 ki->flags = FRhscrollauto|FRvscrollauto; 3856 ki->marginw = FRKIDMARGIN; 3857 ki->marginh = FRKIDMARGIN; 3858 ki->framebd = 1; 3859 } 3860 ki->next = link; 3861 return ki; 3862 } 3863 3864 static Docinfo* 3865 newdocinfo(void) 3866 { 3867 Docinfo* d; 3868 3869 d = (Docinfo*)emalloc(sizeof(Docinfo)); 3870 resetdocinfo(d); 3871 return d; 3872 } 3873 3874 static void 3875 resetdocinfo(Docinfo* d) 3876 { 3877 memset(d, 0, sizeof(Docinfo)); 3878 d->background = makebackground(nil, White); 3879 d->text = Black; 3880 d->link = Blue; 3881 d->vlink = Blue; 3882 d->alink = Blue; 3883 d->target = FTself; 3884 d->chset = ISO_8859_1; 3885 d->scripttype = TextJavascript; 3886 d->frameid = -1; 3887 } 3888 3889 // Use targetmap array to keep track of name <-> targetid mapping. 3890 // Use real malloc(), and never free 3891 static void 3892 targetmapinit(void) 3893 { 3894 targetmapsize = 10; 3895 targetmap = (StringInt*)emalloc(targetmapsize*sizeof(StringInt)); 3896 memset(targetmap, 0, targetmapsize*sizeof(StringInt)); 3897 targetmap[0].key = _Strdup(L"_top"); 3898 targetmap[0].val = FTtop; 3899 targetmap[1].key = _Strdup(L"_self"); 3900 targetmap[1].val = FTself; 3901 targetmap[2].key = _Strdup(L"_parent"); 3902 targetmap[2].val = FTparent; 3903 targetmap[3].key = _Strdup(L"_blank"); 3904 targetmap[3].val = FTblank; 3905 ntargets = 4; 3906 } 3907 3908 int 3909 targetid(Rune* s) 3910 { 3911 int i; 3912 int n; 3913 3914 n = _Strlen(s); 3915 if(n == 0) 3916 return FTself; 3917 for(i = 0; i < ntargets; i++) 3918 if(_Strcmp(s, targetmap[i].key) == 0) 3919 return targetmap[i].val; 3920 if(i >= targetmapsize) { 3921 targetmapsize += 10; 3922 targetmap = (StringInt*)erealloc(targetmap, targetmapsize*sizeof(StringInt)); 3923 } 3924 targetmap[i].key = (Rune*)emalloc((n+1)*sizeof(Rune)); 3925 memmove(targetmap[i].key, s, (n+1)*sizeof(Rune)); 3926 targetmap[i].val = i; 3927 ntargets++; 3928 return i; 3929 } 3930 3931 Rune* 3932 targetname(int targid) 3933 { 3934 int i; 3935 3936 for(i = 0; i < ntargets; i++) 3937 if(targetmap[i].val == targid) 3938 return targetmap[i].key; 3939 return L"?"; 3940 } 3941 3942 // Convert HTML color spec to RGB value, returning dflt if can't. 3943 // Argument is supposed to be a valid HTML color, or "". 3944 // Return the RGB value of the color, using dflt if s 3945 // is nil or an invalid color. 3946 static int 3947 color(Rune* s, int dflt) 3948 { 3949 int v; 3950 Rune* rest; 3951 3952 if(s == nil) 3953 return dflt; 3954 if(_lookup(color_tab, NCOLORS, s, _Strlen(s), &v)) 3955 return v; 3956 if(s[0] == '#') 3957 s++; 3958 v = _Strtol(s, &rest, 16); 3959 if(*rest == 0) 3960 return v; 3961 return dflt; 3962 } 3963 3964 // Debugging 3965 3966 #define HUGEPIX 10000 3967 3968 // A "shallow" validitem, that doesn't follow next links 3969 // or descend into tables. 3970 static int 3971 validitem(Item* i) 3972 { 3973 int ok; 3974 Itext* ti; 3975 Irule* ri; 3976 Iimage* ii; 3977 Ifloat* fi; 3978 int a; 3979 3980 ok = (i->tag >= Itexttag && i->tag <= Ispacertag) && 3981 (i->next == nil || validptr(i->next)) && 3982 (i->width >= 0 && i->width < HUGEPIX) && 3983 (i->height >= 0 && i->height < HUGEPIX) && 3984 (i->ascent > -HUGEPIX && i->ascent < HUGEPIX) && 3985 (i->anchorid >= 0) && 3986 (i->genattr == nil || validptr(i->genattr)); 3987 // also, could check state for ridiculous combinations 3988 // also, could check anchorid for within-doc-range 3989 if(ok) 3990 switch(i->tag) { 3991 case Itexttag: 3992 ti = (Itext*)i; 3993 ok = validStr(ti->s) && 3994 (ti->fnt >= 0 && ti->fnt < NumStyle*NumSize) && 3995 (ti->ul == ULnone || ti->ul == ULunder || ti->ul == ULmid); 3996 break; 3997 case Iruletag: 3998 ri = (Irule*)i; 3999 ok = (validvalign(ri->align) || validhalign(ri->align)) && 4000 (ri->size >=0 && ri->size < HUGEPIX); 4001 break; 4002 case Iimagetag: 4003 ii = (Iimage*)i; 4004 ok = (ii->imsrc == nil || validptr(ii->imsrc)) && 4005 (ii->width >= 0 && ii->width < HUGEPIX) && 4006 (ii->height >= 0 && ii->height < HUGEPIX) && 4007 (ii->imwidth >= 0 && ii->imwidth < HUGEPIX) && 4008 (ii->imheight >= 0 && ii->imheight < HUGEPIX) && 4009 (ii->altrep == nil || validStr(ii->altrep)) && 4010 (ii->map == nil || validptr(ii->map)) && 4011 (validvalign(ii->align) || validhalign(ii->align)) && 4012 (ii->nextimage == nil || validptr(ii->nextimage)); 4013 break; 4014 case Iformfieldtag: 4015 ok = validformfield(((Iformfield*)i)->formfield); 4016 break; 4017 case Itabletag: 4018 ok = validptr((Itable*)i); 4019 break; 4020 case Ifloattag: 4021 fi = (Ifloat*)i; 4022 ok = (fi->side == ALleft || fi->side == ALright) && 4023 validitem(fi->item) && 4024 (fi->item->tag == Iimagetag || fi->item->tag == Itabletag); 4025 break; 4026 case Ispacertag: 4027 a = ((Ispacer*)i)->spkind; 4028 ok = a==ISPnull || a==ISPvline || a==ISPhspace || a==ISPgeneral; 4029 break; 4030 default: 4031 ok = 0; 4032 } 4033 return ok; 4034 } 4035 4036 // "deep" validation, that checks whole list of items, 4037 // and descends into tables and floated tables. 4038 // nil is ok for argument. 4039 int 4040 validitems(Item* i) 4041 { 4042 int ok; 4043 Item* ii; 4044 4045 ok = 1; 4046 while(i != nil && ok) { 4047 ok = validitem(i); 4048 if(ok) { 4049 if(i->tag == Itabletag) { 4050 ok = validtable(((Itable*)i)->table); 4051 } 4052 else if(i->tag == Ifloattag) { 4053 ii = ((Ifloat*)i)->item; 4054 if(ii->tag == Itabletag) 4055 ok = validtable(((Itable*)ii)->table); 4056 } 4057 } 4058 if(!ok) { 4059 fprint(2, "invalid item: %I\n", i); 4060 } 4061 i = i->next; 4062 } 4063 return ok; 4064 } 4065 4066 static int 4067 validformfield(Formfield* f) 4068 { 4069 int ok; 4070 4071 ok = (f->next == nil || validptr(f->next)) && 4072 (f->ftype >= 0 && f->ftype <= Ftextarea) && 4073 f->fieldid >= 0 && 4074 (f->form == nil || validptr(f->form)) && 4075 (f->name == nil || validStr(f->name)) && 4076 (f->value == nil || validStr(f->value)) && 4077 (f->options == nil || validptr(f->options)) && 4078 (f->image == nil || validitem(f->image)) && 4079 (f->events == nil || validptr(f->events)); 4080 // when all built, should have f->fieldid < f->form->nfields, 4081 // but this may be called during build... 4082 return ok; 4083 } 4084 4085 // "deep" validation -- checks cell contents too 4086 static int 4087 validtable(Table* t) 4088 { 4089 int ok; 4090 int i, j; 4091 Tablecell* c; 4092 4093 ok = (t->next == nil || validptr(t->next)) && 4094 t->nrow >= 0 && 4095 t->ncol >= 0 && 4096 t->ncell >= 0 && 4097 validalign(t->align) && 4098 validdimen(t->width) && 4099 (t->border >= 0 && t->border < HUGEPIX) && 4100 (t->cellspacing >= 0 && t->cellspacing < HUGEPIX) && 4101 (t->cellpadding >= 0 && t->cellpadding < HUGEPIX) && 4102 validitems(t->caption) && 4103 (t->caption_place == ALtop || t->caption_place == ALbottom) && 4104 (t->totw >= 0 && t->totw < HUGEPIX) && 4105 (t->toth >= 0 && t->toth < HUGEPIX) && 4106 (t->tabletok == nil || validptr(t->tabletok)); 4107 // during parsing, t->rows has list; 4108 // only when parsing is done is t->nrow set > 0 4109 if(ok && t->nrow > 0 && t->ncol > 0) { 4110 // table is "finished" 4111 for(i = 0; i < t->nrow && ok; i++) 4112 ok = validtablerow(t->rows+i); 4113 for(j = 0; j < t->ncol && ok; j++) 4114 ok = validtablecol(t->cols+j); 4115 for(c = t->cells; c != nil && ok; c = c->next) 4116 ok = validtablecell(c); 4117 for(i = 0; i < t->nrow && ok; i++) 4118 for(j = 0; j < t->ncol && ok; j++) 4119 ok = validptr(t->grid[i][j]); 4120 } 4121 return ok; 4122 } 4123 4124 static int 4125 validvalign(int a) 4126 { 4127 return a == ALnone || a == ALmiddle || a == ALbottom || a == ALtop || a == ALbaseline; 4128 } 4129 4130 static int 4131 validhalign(int a) 4132 { 4133 return a == ALnone || a == ALleft || a == ALcenter || a == ALright || 4134 a == ALjustify || a == ALchar; 4135 } 4136 4137 static int 4138 validalign(Align a) 4139 { 4140 return validhalign(a.halign) && validvalign(a.valign); 4141 } 4142 4143 static int 4144 validdimen(Dimen d) 4145 { 4146 int ok; 4147 int s; 4148 4149 ok = 0; 4150 s = d.kindspec&Dspecmask; 4151 switch(d.kindspec&Dkindmask) { 4152 case Dnone: 4153 ok = s==0; 4154 break; 4155 case Dpixels: 4156 ok = s < HUGEPIX; 4157 break; 4158 case Dpercent: 4159 case Drelative: 4160 ok = 1; 4161 break; 4162 } 4163 return ok; 4164 } 4165 4166 static int 4167 validtablerow(Tablerow* r) 4168 { 4169 return (r->cells == nil || validptr(r->cells)) && 4170 (r->height >= 0 && r->height < HUGEPIX) && 4171 (r->ascent > -HUGEPIX && r->ascent < HUGEPIX) && 4172 validalign(r->align); 4173 } 4174 4175 static int 4176 validtablecol(Tablecol* c) 4177 { 4178 return c->width >= 0 && c->width < HUGEPIX 4179 && validalign(c->align); 4180 } 4181 4182 static int 4183 validtablecell(Tablecell* c) 4184 { 4185 int ok; 4186 4187 ok = (c->next == nil || validptr(c->next)) && 4188 (c->nextinrow == nil || validptr(c->nextinrow)) && 4189 (c->content == nil || validptr(c->content)) && 4190 (c->lay == nil || validptr(c->lay)) && 4191 c->rowspan >= 0 && 4192 c->colspan >= 0 && 4193 validalign(c->align) && 4194 validdimen(c->wspec) && 4195 c->row >= 0 && 4196 c->col >= 0; 4197 if(ok) { 4198 if(c->content != nil) 4199 ok = validitems(c->content); 4200 } 4201 return ok; 4202 } 4203 4204 static int 4205 validptr(void* p) 4206 { 4207 // TODO: a better job of this. 4208 // For now, just dereference, which cause a bomb 4209 // if not valid 4210 static char c; 4211 4212 c = *((char*)p); 4213 return 1; 4214 } 4215 4216 static int 4217 validStr(Rune* s) 4218 { 4219 return s != nil && validptr(s); 4220 } 4221