static void addhexfilter(fz_obj *dict)
{
fz_obj *f, *dp, *newf, *newdp;
fz_obj *ahx, *nullobj;
ahx = fz_newname("ASCIIHexDecode");
nullobj = fz_newnull();
newf = newdp = nil;
f = fz_dictgets(dict, "Filter");
dp = fz_dictgets(dict, "DecodeParms");
if (fz_isname(f))
{
newf = fz_newarray(2);
fz_arraypush(newf, ahx);
fz_arraypush(newf, f);
f = newf;
if (fz_isdict(dp))
{
newdp = fz_newarray(2);
fz_arraypush(newdp, nullobj);
fz_arraypush(newdp, dp);
dp = newdp;
}
}
else if (fz_isarray(f))
{
fz_arrayinsert(f, ahx);
if (fz_isarray(dp))
fz_arrayinsert(dp, nullobj);
}
else
f = ahx;
fz_dictputs(dict, "Filter", f);
if (dp)
fz_dictputs(dict, "DecodeParms", dp);
fz_dropobj(ahx);
fz_dropobj(nullobj);
if (newf)
fz_dropobj(newf);
if (newdp)
fz_dropobj(newdp);
}
void
editcopy(int pagenum)
{
fz_error *error;
fz_obj *obj;
fz_obj *ref;
fz_obj *num;
printf("copy %s page %d\n", srcname, pagenum);
ref = srcpages->pref[pagenum - 1];
obj = pdf_getpageobject(srcpages, pagenum - 1);
fz_dictdels(obj, "Parent");
/*
fz_dictdels(obj, "B");
fz_dictdels(obj, "PieceInfo");
fz_dictdels(obj, "Metadata");
fz_dictdels(obj, "Annots");
fz_dictdels(obj, "Tabs");
*/
pdf_updateobject(src, fz_tonum(ref), fz_togen(ref), obj);
error = fz_arraypush(editobjects, ref);
if (error)
die(error);
error = fz_newint(&num, editmode);
if (error)
die(error);
error = fz_arraypush(editmodelist, num);
if (error)
die(error);
fz_dropobj(num);
}
static fz_error parsearray(fz_obj **obj, pdf_xref *xref, char **sp, struct vap *v)
{
fz_error error;
fz_obj *a;
fz_obj *o;
char *s = *sp;
error = fz_newarray(&a, 8);
if (error)
return fz_rethrow(error, "cannot create array");
s ++; /* skip '[' */
while (*s)
{
skipwhite(&s);
if (*s == ']')
{
s ++;
break;
}
error = parseobj(&o, xref, &s, v);
if (error)
{
fz_dropobj(a);
return fz_rethrow(error, "cannot parse item");
}
error = fz_arraypush(a, o);
if (error)
{
fz_dropobj(o);
fz_dropobj(a);
return fz_rethrow(error, "cannot add item to array");
}
fz_dropobj(o);
}
*obj = a;
*sp = s;
return fz_okay;
}
void
editflushobjects(void)
{
fz_error *error;
fz_obj *results;
int i;
error = pdf_transplant(editxref, src, &results, editobjects);
if (error)
die(error);
for (i = 0; i < fz_arraylen(results); i++)
{
error = fz_arraypush(editpagelist, fz_arrayget(results, i));
if (error)
die(error);
}
fz_dropobj(results);
}
static void retainpages(int argc, char **argv)
{
fz_error error;
fz_obj *oldroot, *root, *pages, *kids, *countobj, *parent;
/* Load the old page tree */
error = pdf_loadpagetree(xref);
if (error)
die(fz_rethrow(error, "cannot load page tree"));
/* Keep only pages/type entry to avoid references to unretained pages */
oldroot = fz_dictgets(xref->trailer, "Root");
pages = fz_dictgets(oldroot, "Pages");
root = fz_newdict(2);
fz_dictputs(root, "Type", fz_dictgets(oldroot, "Type"));
fz_dictputs(root, "Pages", fz_dictgets(oldroot, "Pages"));
pdf_updateobject(xref, fz_tonum(oldroot), fz_togen(oldroot), root);
fz_dropobj(root);
/* Create a new kids array with only the pages we want to keep */
parent = fz_newindirect(fz_tonum(pages), fz_togen(pages), xref);
kids = fz_newarray(1);
/* Retain pages specified */
while (argc - fz_optind)
{
int page, spage, epage;
char *spec, *dash;
char *pagelist = argv[fz_optind];
spec = fz_strsep(&pagelist, ",");
while (spec)
{
dash = strchr(spec, '-');
if (dash == spec)
spage = epage = pdf_getpagecount(xref);
else
spage = epage = atoi(spec);
if (dash)
{
if (strlen(dash) > 1)
epage = atoi(dash + 1);
else
epage = pdf_getpagecount(xref);
}
if (spage > epage)
page = spage, spage = epage, epage = page;
if (spage < 1)
spage = 1;
if (epage > pdf_getpagecount(xref))
epage = pdf_getpagecount(xref);
for (page = spage; page <= epage; page++)
{
fz_obj *pageobj = pdf_getpageobject(xref, page);
fz_obj *pageref = pdf_getpageref(xref, page);
fz_dictputs(pageobj, "Parent", parent);
/* Store page object in new kids array */
fz_arraypush(kids, pageref);
}
spec = fz_strsep(&pagelist, ",");
}
fz_optind++;
}
fz_dropobj(parent);
/* Update page count and kids array */
countobj = fz_newint(fz_arraylen(kids));
fz_dictputs(pages, "Count", countobj);
fz_dropobj(countobj);
fz_dictputs(pages, "Kids", kids);
fz_dropobj(kids);
}
int
copyPdfFile(
soPdfFile* inFile,
soPdfFile* outFile
)
{
fz_error *error;
int pageTreeNum, pageTreeGen;
assert(inFile != NULL);
assert(outFile != NULL);
//
// Process every page in the source file
//
{
printf("\nProcessing input page : ");
for (int pageNo = 0; pageNo < pdf_getpagecount(inFile->pageTree); pageNo++)
{
displayPageNumber(pageNo + 1, !pageNo);
// Get the page object from the source
fz_obj *pageRef = inFile->pageTree->pref[pageNo];
fz_obj *pageObj = pdf_getpageobject(inFile->pageTree, pageNo);
//
// Process the page. Each page can be split into up-to 3 pages
//
fz_rect bbRect[3];
error = processPage(inFile, pageNo, bbRect, 3);
if (error)
return soPdfError(error);
for (int ctr = 0; ctr < 3; ctr++)
{
// Check if this was a blank page
if (fz_isemptyrect(bbRect[ctr]))
break;
//
// copy the source page dictionary entry. The way this is done is basically
// by making a copy of the page dict object in the source file, and adding
// the copy in the source file. Then the copied page dict object is
// referenced and added to the destination file.
//
// This convoluted procedure is done because the copy is done by pdf_transplant
// function that accepts a source and destination. Whatever is referenced by
// destination object is deep copied
//
// allocate an object id and generation id in source file
//
// There is a bug in mupdf where the object allocation returns
// 0 oid and 0 gid when the input pdf file has iref stream
// so to work around the issue, we wrap the pdf_allocojbect
// in a for loop 10 times to get the number
//
int sNum, sGen, tries;
for (tries = 0; tries < 10; tries++)
{
error = pdf_allocobject(inFile->xref, &sNum, &sGen);
if (error)
return soPdfError(error);
// If sNum is non zero then the allocation was successful
if (sNum != 0)
break;
pdf_updateobject(inFile->xref, sNum, sGen, pageObj);
}
// If we didn't succeed even after 10 tries then this file
// is not going to work.
if (tries >= 10)
return soPdfError(fz_throw("cannot allocate object because of mupdf bug"));
// make a deep copy of the original page dict
fz_obj *pageObj2;
error = fz_deepcopydict(&pageObj2, pageObj);
if (error)
return soPdfError(error);
// update the source file with the duplicate page object
pdf_updateobject(inFile->xref, sNum, sGen, pageObj2);
fz_dropobj(pageObj2);
// create an indirect reference to the page object
fz_obj *pageRef2;
error = fz_newindirect(&pageRef2, sNum, sGen);
if (error)
return soPdfError(error);
// delete the parent dictionary entry
// Do we need to delete any other dictionary entry
// like annot, tabs, metadata, etc
fz_dictdels(pageObj2, "Parent");
// Set the media box
setPageMediaBox(inFile->xref, pageObj2, bbRect[ctr]);
// Set the rotation based on input
switch(p_mode)
{
// no rotation if fit height
case FitHeight:
case Fit2xHeight:
break;
// rotate -90 deg if fit width
case Fit2xWidth:
case FitWidth:
setPageRotate(pageObj2, p_reverseLandscape ? 90 : -90);
break;
case SmartFitHeight:
case SmartFitWidth:
default:
return soPdfError(fz_throw("Mode(%d) not yet implemented.", p_mode));
break;
}
// push the indirect reference to the destination list for copy by pdf_transplant
error = fz_arraypush(outFile->editobjs, pageRef2);
if (error)
return soPdfError(error);
}
}
}
// flush the objects into destination from source
{
fz_obj *results;
int outPages;
printf("\nCopying output page : ");
error = pdf_transplant(outFile->xref, inFile->xref, &results, outFile->editobjs);
if (error)
return soPdfError(error);
outPages = fz_arraylen(results);
for (int ctr = 0; ctr < outPages; ctr++)
{
displayPageNumber(ctr + 1, !ctr);
error = fz_arraypush(outFile->pagelist, fz_arrayget(results,
p_reverseLandscape ? outPages - 1 - ctr : ctr));
if (error)
return soPdfError(error);
}
fz_dropobj(results);
}
// flush page tree
// Create page tree and add back-links
{
fz_obj *pageTreeObj;
fz_obj *pageTreeRef;
// allocate a new object in out file for pageTree object
error = pdf_allocobject(outFile->xref, &pageTreeNum, &pageTreeGen);
if (error)
return soPdfError(error);
// Create a page tree object
error = fz_packobj(&pageTreeObj, "<</Type/Pages/Count %i/Kids %o>>",
fz_arraylen(outFile->pagelist), outFile->pagelist);
if (error)
return soPdfError(error);
// Update the xref entry with the pageTree object
pdf_updateobject(outFile->xref, pageTreeNum, pageTreeGen, pageTreeObj);
fz_dropobj(pageTreeObj);
// Create a reference to the pageTree object
error = fz_newindirect(&pageTreeRef, pageTreeNum, pageTreeGen);
if (error)
return soPdfError(error);
//
// For every page in the output file, update the parent entry
//
for (int ctr = 0; ctr < fz_arraylen(outFile->pagelist); ctr++)
{
fz_obj *pageObj;
int num = fz_tonum(fz_arrayget(outFile->pagelist, ctr));
int gen = fz_togen(fz_arrayget(outFile->pagelist, ctr));
// Get the page object from xreft
error = pdf_loadobject(&pageObj, outFile->xref, num, gen);
if (error)
return soPdfError(error);
// Update the parent entry in the page dictionary
error = fz_dictputs(pageObj, "Parent", pageTreeRef);
if (error)
return soPdfError(error);
// Update the entry with the updated page object
pdf_updateobject(outFile->xref, num, gen, pageObj);
fz_dropobj(pageObj);
}
}
// Create catalog and root entries
{
fz_obj *catObj, *infoObj;
int rootNum, rootGen;
int infoNum, infoGen;
//
// Copy the info catalog to the destination
// alloc an object id and gen id in destination file
error = pdf_allocobject(outFile->xref, &infoNum, &infoGen);
if (error)
return soPdfError(error);
// make a deep copy of the original page dict
error = fz_deepcopydict(&infoObj, inFile->xref->info);
if (error)
return soPdfError(error);
// update the dest file with object
pdf_updateobject(outFile->xref, infoNum, infoGen, infoObj);
outFile->xref->info = infoObj;
fz_dropobj(infoObj);
//
// root/catalog object creation
error = pdf_allocobject(outFile->xref, &rootNum, &rootGen);
if (error)
return soPdfError(error);
error = fz_packobj(&catObj, "<</Type/Catalog /Pages %r>>", pageTreeNum, pageTreeGen);
if (error)
return soPdfError(error);
pdf_updateobject(outFile->xref, rootNum, rootGen, catObj);
fz_dropobj(catObj);
// Create trailer
error = fz_packobj(&outFile->xref->trailer, "<</Root %r /Info %r>>",
rootNum, rootGen, infoNum, infoGen);
if (error)
return soPdfError(error);
}
// Update the info in the target file and save the xref
printf("\nSaving.\n");
error = setPageInfo(inFile, outFile);
if (error)
return soPdfError(error);
error = pdf_savexref(outFile->xref, outFile->fileName, NULL);
if (error)
return soPdfError(error);
if (g_errorCount != 0)
{
printf("\nFollowing issues encounted were ignored.\n\n");
for (int ctr = g_errorCount - 1; ctr >= 0; ctr--)
soPdfError(g_errorList[ctr]);
}
printf("\nSaved.\n");
return 0;
}
fz_error
pdf_parsearray(fz_obj **op, pdf_xref *xref, fz_stream *file, char *buf, int cap)
{
fz_error error = fz_okay;
fz_obj *ary = nil;
fz_obj *obj = nil;
int a = 0, b = 0, n = 0;
pdf_token_e tok;
int len;
ary = fz_newarray(4);
while (1)
{
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_dropobj(ary);
return fz_rethrow(error, "cannot parse array");
}
if (tok != PDF_TINT && tok != PDF_TR)
{
if (n > 0)
{
obj = fz_newint(a);
fz_arraypush(ary, obj);
fz_dropobj(obj);
}
if (n > 1)
{
obj = fz_newint(b);
fz_arraypush(ary, obj);
fz_dropobj(obj);
}
n = 0;
}
if (tok == PDF_TINT && n == 2)
{
obj = fz_newint(a);
fz_arraypush(ary, obj);
fz_dropobj(obj);
a = b;
n --;
}
switch (tok)
{
case PDF_TCARRAY:
*op = ary;
return fz_okay;
case PDF_TINT:
if (n == 0)
a = atoi(buf);
if (n == 1)
b = atoi(buf);
n ++;
break;
case PDF_TR:
if (n != 2)
{
fz_dropobj(ary);
return fz_throw("cannot parse indirect reference in array");
}
obj = fz_newindirect(a, b, xref);
fz_arraypush(ary, obj);
fz_dropobj(obj);
n = 0;
break;
case PDF_TOARRAY:
error = pdf_parsearray(&obj, xref, file, buf, cap);
if (error)
{
fz_dropobj(ary);
return fz_rethrow(error, "cannot parse array");
}
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TODICT:
error = pdf_parsedict(&obj, xref, file, buf, cap);
if (error)
{
fz_dropobj(ary);
return fz_rethrow(error, "cannot parse array");
}
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TNAME:
obj = fz_newname(buf);
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TREAL:
obj = fz_newreal(atof(buf));
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TSTRING:
obj = fz_newstring(buf, len);
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TTRUE:
obj = fz_newbool(1);
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TFALSE:
obj = fz_newbool(0);
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
case PDF_TNULL:
obj = fz_newnull();
fz_arraypush(ary, obj);
fz_dropobj(obj);
break;
default:
fz_dropobj(ary);
return fz_throw("cannot parse token in array");
}
}
}
fz_error *
pdf_parsearray(fz_obj **op, fz_stream *file, char *buf, int cap)
{
fz_error *error = nil;
fz_obj *ary = nil;
fz_obj *obj = nil;
int a = 0, b = 0, n = 0;
int tok, len;
error = fz_newarray(op, 4);
if (error) return error;
ary = *op;
while (1)
{
tok = pdf_lex(file, buf, cap, &len);
if (tok != PDF_TINT && tok != PDF_TR)
{
if (n > 0)
{
error = fz_newint(&obj, a);
if (error) goto cleanup;
error = fz_arraypush(ary, obj);
if (error) goto cleanup;
fz_dropobj(obj);
obj = nil;
}
if (n > 1)
{
error = fz_newint(&obj, b);
if (error) goto cleanup;
error = fz_arraypush(ary, obj);
if (error) goto cleanup;
fz_dropobj(obj);
obj = nil;
}
n = 0;
}
if (tok == PDF_TINT && n == 2)
{
error = fz_newint(&obj, a);
if (error) goto cleanup;
error = fz_arraypush(ary, obj);
if (error) goto cleanup;
fz_dropobj(obj);
obj = nil;
a = b;
n --;
}
switch (tok)
{
case PDF_TCARRAY:
return nil;
case PDF_TINT:
if (n == 0)
a = atoi(buf);
if (n == 1)
b = atoi(buf);
n ++;
break;
case PDF_TR:
if (n != 2)
goto cleanup;
error = fz_newindirect(&obj, a, b);
if (error) goto cleanup;
n = 0;
break;
case PDF_TOARRAY: error = pdf_parsearray(&obj, file, buf, cap); break;
case PDF_TODICT: error = pdf_parsedict(&obj, file, buf, cap); break;
case PDF_TNAME: error = fz_newname(&obj, buf); break;
case PDF_TREAL: error = fz_newreal(&obj, atof(buf)); break;
case PDF_TSTRING: error = fz_newstring(&obj, buf, len); break;
case PDF_TTRUE: error = fz_newbool(&obj, 1); break;
case PDF_TFALSE: error = fz_newbool(&obj, 0); break;
case PDF_TNULL: error = fz_newnull(&obj); break;
default: goto cleanup;
}
if (error) goto cleanup;
if (obj)
{
error = fz_arraypush(ary, obj);
if (error) goto cleanup;
fz_dropobj(obj);
}
obj = nil;
}
cleanup:
if (obj) fz_dropobj(obj);
if (ary) fz_dropobj(ary);
if (error) return error;
return fz_throw("syntaxerror: corrupt array");
}
