pdf.c

这是一个同样来自贝尔实验室的和UNIX有着渊源的操作系统, 其简洁的设计和实现易于我们学习和理解

/*
 * pdf.c
 * Copyright (C) 2003-2005 A.J. van Os; Released under GNU GPL
 *
 * Description:
 * Functions to deal with the Adobe Portable Document Format (pdf)
 *
 */

#include <stdarg.h>
#include <string.h>
#include "version.h"
#include "antiword.h"


/* Constants for the file positions */
#define INITIAL_LOCATION_SIZE	20
#define INITIAL_PAGEOBJECT_SIZE	 5
#if defined(DEBUG)
#define EXTENSION_ARRAY_SIZE	10
#else
#define EXTENSION_ARRAY_SIZE	30
#endif /* DEBUG */

/* The character set */
static encoding_type	eEncoding = encoding_neutral;
/* Current creator for a PDF header */
static const char	*szProducer = NULL;
/* The height and width of a PDF page (in DrawUnits) */
static long		lPageHeight = LONG_MAX;
static long		lPageWidth = LONG_MAX;
/* The height of the footer on the current page (in DrawUnits) */
static long		lFooterHeight = 0;
/* Inside a footer (to prevent an infinite loop when the footer is too big) */
static BOOL		bInFtrSpace = FALSE;
/* Current font information */
static drawfile_fontref	tFontRefCurr = (drawfile_fontref)-1;
static USHORT		usFontSizeCurr = 0;
static int		iFontColorCurr = -1;
/* Current vertical position information */
static long		lYtopCurr = -1;
/* Image counter */
static int		iImageCount = 0;
/* Section index */
static int		iSectionIndex = 0;
/* Are we on the first page of the section? */
static BOOL		bFirstInSection = TRUE;
/* File positions */
static long		lFilePosition = 0;
static long		*alLocation = NULL;
static size_t		tLocations = 0;
static int		iMaxLocationNumber = 0;
/* File position at the start of a page */
static long		lStreamStart = -1;
/* Page objects */
static int		*aiPageObject = NULL;
static int		iPageCount = 0;
static size_t		tMaxPageObjects = 0;
/* Current object number */
/* 1 = root; 2 = info; 3 = pages; 4 = encoding; 5-16 = fonts; 17 = resources */
static int		iObjectNumberCurr = 17;

static void		vMoveTo(diagram_type *, long);

static const struct {
	const char	*szPDFname;
	const char	*szPSname;
} atFontname[] = {
	{ "Courier",			FONT_MONOSPACED_PLAIN },
	{ "Courier-Bold",		FONT_MONOSPACED_BOLD },
	{ "Courier-Oblique",		FONT_MONOSPACED_ITALIC },
	{ "Courier-BoldOblique",	FONT_MONOSPACED_BOLDITALIC },
	{ "Helvetica",			FONT_SANS_SERIF_PLAIN },
	{ "Helvetica-Bold",		FONT_SANS_SERIF_BOLD },
	{ "Helvetica-Oblique",		FONT_SANS_SERIF_ITALIC },
	{ "Helvetica-BoldOblique",	FONT_SANS_SERIF_BOLDITALIC },
	{ "Times-Roman",		FONT_SERIF_PLAIN },
	{ "Times-Bold",			FONT_SERIF_BOLD },
	{ "Times-Italic",		FONT_SERIF_ITALIC },
	{ "Times-BoldItalic",		FONT_SERIF_BOLDITALIC },
};

static const char *iso_8859_1[] = {
"128 /Euro",
"140 /ellipsis /trademark /perthousand /bullet",
"    /quoteleft /quoteright /guilsinglleft /guilsinglright",
"    /quotedblleft /quotedblright /quotedblbase /endash /emdash",
"    /minus /OE /oe /dagger /daggerdbl /fi /fl",
"160 /space /exclamdown /cent /sterling /currency",
"    /yen /brokenbar /section /dieresis /copyright",
"    /ordfeminine /guillemotleft /logicalnot /hyphen /registered",
"    /macron /degree /plusminus /twosuperior /threesuperior",
"    /acute /mu /paragraph /periodcentered /cedilla",
"    /onesuperior /ordmasculine /guillemotright /onequarter",
"    /onehalf /threequarters /questiondown /Agrave /Aacute",
"    /Acircumflex /Atilde /Adieresis /Aring /AE /Ccedilla",
"    /Egrave /Eacute /Ecircumflex /Edieresis /Igrave /Iacute",
"    /Icircumflex /Idieresis /Eth /Ntilde /Ograve /Oacute",
"    /Ocircumflex /Otilde /Odieresis /multiply /Oslash",
"    /Ugrave /Uacute /Ucircumflex /Udieresis /Yacute /Thorn",
"    /germandbls /agrave /aacute /acircumflex /atilde",
"    /adieresis /aring /ae /ccedilla /egrave /eacute",
"    /ecircumflex /edieresis /igrave /iacute /icircumflex",
"    /idieresis /eth /ntilde /ograve /oacute /ocircumflex",
"    /otilde /odieresis /divide /oslash /ugrave /uacute",
"    /ucircumflex /udieresis /yacute /thorn /ydieresis",
};

static const char *iso_8859_2[] = {
"160 /space /Aogonek /breve /Lslash /currency /Lcaron",
"    /Sacute /section /dieresis /Scaron /Scommaaccent",
"    /Tcaron /Zacute /hyphen /Zcaron /Zdotaccent /degree",
"    /aogonek /ogonek /lslash /acute /lcaron /sacute",
"    /caron /cedilla /scaron /scommaaccent /tcaron",
"    /zacute /hungarumlaut /zcaron /zdotaccent /Racute",
"    /Aacute /Acircumflex /Abreve /Adieresis /Lacute",
"    /Cacute /Ccedilla /Ccaron /Eacute /Eogonek",
"    /Edieresis /Ecaron /Iacute /Icircumflex /Dcaron",
"    /.notdef /Nacute /Ncaron /Oacute /Ocircumflex",
"    /Ohungarumlaut /Odieresis /multiply /Rcaron /Uring",
"    /Uacute /Uhungarumlaut /Udieresis /Yacute /Tcommaaccent",
"    /germandbls /racute /aacute /acircumflex /abreve",
"    /adieresis /lacute /cacute /ccedilla /ccaron /eacute",
"    /eogonek /edieresis /ecaron /iacute /icircumflex",
"    /dcaron /.notdef /nacute /ncaron /oacute /ocircumflex",
"    /ohungarumlaut /odieresis /divide /rcaron /uring",
"    /uacute /uhungarumlaut /udieresis /yacute /tcommaaccent",
"    /dotaccent",
};


/*
 * tGetFontIndex - get the font index
 */
static size_t
tGetFontIndex(drawfile_fontref tFontRef)
{
	const char	*szFontname;
	size_t		tIndex;

	/* Get the font name */
	szFontname = szGetFontname(tFontRef);
	fail(szFontname == NULL);
	if (szFontname == NULL) {
		return 0;
	}

	/* Find the name in the table */
	for (tIndex = 0; tIndex < elementsof(atFontname); tIndex++) {
		if (STRCEQ(atFontname[tIndex].szPSname, szFontname)) {
			return tIndex;
		}
	}
	/* Not found */
	DBG_DEC(tFontRef);
	DBG_MSG(szFontname);
	return 0;
} /* end of tGetFontIndex */

/*
 * vSetLocation - store the location of objects
 */
static void
vSetLocation(int iLocationNumber)
{
	fail(iLocationNumber <= 0);

	if ((size_t)iLocationNumber >= tLocations) {
		/* Extend and set to zero */
		tLocations += EXTENSION_ARRAY_SIZE;
		alLocation = xrealloc(alLocation, tLocations * sizeof(long));
		memset(alLocation + tLocations - EXTENSION_ARRAY_SIZE,
			0,
			EXTENSION_ARRAY_SIZE * sizeof(long));
		DBG_DEC(tLocations);
	}
	if (iLocationNumber > iMaxLocationNumber) {
		iMaxLocationNumber = iLocationNumber;
	}

	DBG_DEC_C((size_t)iLocationNumber >= tLocations, iLocationNumber);
	DBG_DEC_C((size_t)iLocationNumber >= tLocations, tLocations);
	fail((size_t)iLocationNumber >= tLocations);

	alLocation[iLocationNumber] = lFilePosition;
} /* end of vSetLocation */

/*
 * vFillNextPageObject - fil the next page object with the current object number
 */
static void
vFillNextPageObject(void)
{
	iPageCount++;
	if ((size_t)iPageCount >= tMaxPageObjects) {
		/* Extend the array */
		tMaxPageObjects += EXTENSION_ARRAY_SIZE;
		aiPageObject = xrealloc(aiPageObject,
					tMaxPageObjects * sizeof(int));
		DBG_DEC(tMaxPageObjects);
	}
	aiPageObject[iPageCount] = iObjectNumberCurr;
} /* end of vFillNextPageObject */

/*
 * vFPprintf - printf and update the fileposition
 *
 * called with arguments like fprintf(3)
 */
static void
vFPprintf(FILE *pOutFile, const char *szFormat, ...)
{
	va_list	tArg;

	va_start(tArg, szFormat);
	lFilePosition += vfprintf(pOutFile, szFormat, tArg);
	va_end(tArg);
} /* end of vFPprintf */

/*
 * vCreateInfoDictionary - create the document information dictionary
 */
void
vCreateInfoDictionary(diagram_type *pDiag, int iWordVersion)
{
	FILE	*pOutFile;
	const char	*szTitle, *szAuthor, *szSubject, *szCreator;
	const char	*szCreationDate, *szModDate;

	fail(pDiag == NULL);
	fail(pDiag->pOutFile == NULL);
	fail(iWordVersion < 0);
	fail(szProducer == NULL || szProducer[0] == '\0');

	szTitle = szGetTitle();
	szAuthor = szGetAuthor();
	szSubject = szGetSubject();
	szCreationDate = szGetCreationDate();
	szModDate = szGetModDate();

	switch (iWordVersion) {
	case 0: szCreator = "Word for DOS"; break;
	case 1: szCreator = "WinWord 1.x"; break;
	case 2: szCreator = "WinWord 2.0"; break;
	case 4: szCreator = "MacWord 4"; break;
	case 5: szCreator = "MacWord 5"; break;
	case 6: szCreator = "Word 6"; break;
	case 7: szCreator = "Word 7/95"; break;
	case 8: szCreator = "Word 97 or later"; break;
	default: szCreator = NULL; break;
	}

	pOutFile = pDiag->pOutFile;

	vSetLocation(2);
	vFPprintf(pOutFile, "2 0 obj\n");
	vFPprintf(pOutFile, "<<\n");
	if (szTitle != NULL && szTitle[0] != '\0') {
		vFPprintf(pOutFile, "/Title (%s)\n", szTitle);
	}
	if (szAuthor != NULL && szAuthor[0] != '\0') {
		vFPprintf(pOutFile, "/Author (%s)\n", szAuthor);
	}
	if (szSubject != NULL && szSubject[0] != '\0') {
		vFPprintf(pOutFile, "/Subject (%s)\n", szSubject);
	}
	if (szCreator != NULL && szCreator[0] != '\0') {
		vFPprintf(pOutFile, "/Creator (%s)\n", szCreator);
	}
	vFPprintf(pOutFile, "/Producer (%s %s)\n", szProducer, VERSIONSTRING);
	if (szCreationDate != NULL && szCreationDate[0] != '\0') {
		vFPprintf(pOutFile, "/CreationDate (%s)\n", szCreationDate);
	}
	if (szModDate != NULL && szModDate[0] != '\0') {
		vFPprintf(pOutFile, "/ModDate (%s)\n", szModDate);
	}
	vFPprintf(pOutFile, ">>\n");
	vFPprintf(pOutFile, "endobj\n");
} /* end of vCreateInfoDictionary */

/*
 * vAddHdrFtr - add a header or footer
 */
static void
vAddHdrFtr(diagram_type *pDiag, const hdrftr_block_type *pHdrFtrInfo)
{
	output_type	*pStart, *pPrev, *pNext;

	fail(pDiag == NULL);
	fail(pHdrFtrInfo == NULL);

	vStartOfParagraphPDF(pDiag, 0);
	pStart = pHdrFtrInfo->pText;
	while (pStart != NULL) {
		pNext = pStart;
		while (pNext != NULL &&
		       (pNext->tNextFree != 1 ||
		        (pNext->szStorage[0] != PAR_END &&
		         pNext->szStorage[0] != HARD_RETURN))) {
			pNext = pNext->pNext;
		}
		if (pNext == NULL) {
			if (bOutputContainsText(pStart)) {
				vAlign2Window(pDiag, pStart,
					lChar2MilliPoints(DEFAULT_SCREEN_WIDTH),
					ALIGNMENT_LEFT);
			} else {
				vMove2NextLinePDF(pDiag, pStart->usFontSize);
			}
			break;
		}
		fail(pNext->tNextFree != 1);
		fail(pNext->szStorage[0] != PAR_END &&
			pNext->szStorage[0] != HARD_RETURN);

		if (pStart != pNext) {
			/* There is something to print */
			pPrev = pNext->pPrev;
			fail(pPrev->pNext != pNext);
			/* Cut the chain */
			pPrev->pNext = NULL;
			if (bOutputContainsText(pStart)) {
				/* Print it */
				vAlign2Window(pDiag, pStart,
					lChar2MilliPoints(DEFAULT_SCREEN_WIDTH),
					ALIGNMENT_LEFT);
			} else {
				/* Just an empty line */
				vMove2NextLinePDF(pDiag, pStart->usFontSize);
			}
			/* Repair the chain */
			pPrev->pNext = pNext;
		}
		if (pNext->szStorage[0] == PAR_END) {
			vEndOfParagraphPDF(pDiag, pNext->usFontSize,
					(long)pNext->usFontSize * 200);
		}
		pStart = pNext->pNext;
	}
} /* end of vAddHdrFtr */

/*
 * vAddHeader - add a page header
 */
static void
vAddHeader(diagram_type *pDiag)
{
	const hdrftr_block_type *pHdrInfo;
	const hdrftr_block_type *pFtrInfo;

	fail(pDiag == NULL);

	NO_DBG_MSG("vAddHeader");

	pHdrInfo = pGetHdrFtrInfo(iSectionIndex, TRUE,
					odd(iPageCount), bFirstInSection);
	pFtrInfo = pGetHdrFtrInfo(iSectionIndex, FALSE,
					odd(iPageCount), bFirstInSection);
	/* Set the height of the footer of this page */
	lFooterHeight = pFtrInfo == NULL ? 0 : pFtrInfo->lHeight;
	fail(lFooterHeight < 0);

	if (pHdrInfo == NULL ||
	    pHdrInfo->pText == NULL ||
	    pHdrInfo->lHeight <= 0) {
		fail(pHdrInfo != NULL && pHdrInfo->lHeight < 0);
		fail(pHdrInfo != NULL &&
			pHdrInfo->pText != NULL &&
			pHdrInfo->lHeight == 0);
		return;
	}

	vAddHdrFtr(pDiag, pHdrInfo);

	DBG_DEC_C(pHdrInfo->lHeight !=
		lPageHeight - PS_TOP_MARGIN - pDiag->lYtop,
		pHdrInfo->lHeight);
	DBG_DEC_C(pHdrInfo->lHeight !=
		lPageHeight - PS_TOP_MARGIN - pDiag->lYtop,
		lPageHeight - PS_TOP_MARGIN - pDiag->lYtop);
} /* end of vAddHeader */

/*