Update pdf_write_document to support incremental update

12 files changed, 220 insertions, 60 deletions

diff --git a/include/mupdf/fitz/write-document.h b/include/mupdf/fitz/write-document.h
index f345cc56..9fe27f5f 100644
--- a/include/mupdf/fitz/write-document.h
+++ b/include/mupdf/fitz/write-document.h
@@ -12,6 +12,7 @@
 */
 struct fz_write_options_s
 {
+	int do_incremental; /* Write just the changed objects */
 	int do_ascii; /* If non-zero then attempt (where possible) to make
 				the output ascii. */
 	int do_expand; /* Bitflags; each non zero bit indicates an aspect
diff --git a/include/mupdf/pdf/xref.h b/include/mupdf/pdf/xref.h
index c0544717..5d4ecde9 100644
--- a/include/mupdf/pdf/xref.h
+++ b/include/mupdf/pdf/xref.h
@@ -76,6 +76,7 @@ pdf_xref_entry *pdf_get_populating_xref_entry(pdf_document *doc, int i);
 pdf_xref_entry *pdf_get_xref_entry(pdf_document *doc, int i);
 void pdf_replace_xref(pdf_document *doc, pdf_xref_entry *entries, int n);
 void pdf_xref_ensure_incremental_object(pdf_document *doc, int num);
+int pdf_xref_is_incremental(pdf_document *doc, int num);
 
 void pdf_repair_xref(pdf_document *doc, pdf_lexbuf *buf);
 void pdf_repair_obj_stms(pdf_document *doc);
diff --git a/platform/android/jni/mupdf.c b/platform/android/jni/mupdf.c
index bb827778..889abe07 100644
--- a/platform/android/jni/mupdf.c
+++ b/platform/android/jni/mupdf.c
@@ -2342,21 +2342,42 @@ JNI_FN(MuPDFCore_saveInternal)(JNIEnv * env, jobject thiz)
 	{
 		char *tmp;
 		fz_write_options opts;
-		opts.do_ascii = 1;
+		opts.do_incremental = 1;
+		opts.do_ascii = 0;
 		opts.do_expand = 0;
-		opts.do_garbage = 1;
+		opts.do_garbage = 0;
 		opts.do_linear = 0;
 
 		tmp = tmp_path(glo->current_path);
 		if (tmp)
 		{
-			int written;
+			int written = 0;
 
 			fz_var(written);
 			fz_try(ctx)
 			{
-				fz_write_document(glo->doc, tmp, &opts);
-				written = 1;
+				FILE *fin = fopen(glo->current_path, "rb");
+				FILE *fout = fopen(tmp, "wb");
+				char buf[256];
+				int n, err = 1;
+
+				if (fin && fout)
+				{
+					while ((n = fread(buf, 1, sizeof(buf), fin)) > 0)
+						fwrite(buf, 1, n, fout);
+					err = (ferror(fin) || ferror(fout));
+				}
+
+				if (fin)
+					fclose(fin);
+				if (fout)
+					fclose(fout);
+
+				if (!err)
+				{
+					fz_write_document(glo->doc, tmp, &opts);
+					written = 1;
+				}
 			}
 			fz_catch(ctx)
 			{
diff --git a/platform/x11/jstest_main.c b/platform/x11/jstest_main.c
index 2003ad55..a07385bc 100644
--- a/platform/x11/jstest_main.c
+++ b/platform/x11/jstest_main.c
@@ -108,6 +108,10 @@ void winreplacefile(char *source, char *target)
 {
 }
 
+void wincopyfile(char *source, char *target)
+{
+}
+
 void wincursor(pdfapp_t *app, int curs)
 {
 }
diff --git a/platform/x11/pdfapp.c b/platform/x11/pdfapp.c
index e76c6c7c..71561f14 100644
--- a/platform/x11/pdfapp.c
+++ b/platform/x11/pdfapp.c
@@ -299,13 +299,15 @@ static int pdfapp_save(pdfapp_t *app)
 	{
 		fz_write_options opts;
 
-		opts.do_ascii = 1;
+		opts.do_incremental = 1;
+		opts.do_ascii = 0;
 		opts.do_expand = 0;
-		opts.do_garbage = 1;
+		opts.do_garbage = 0;
 		opts.do_linear = 0;
 
 		if (strcmp(buf, app->docpath) != 0)
 		{
+			wincopyfile(app->docpath, buf);
 			fz_write_document(app->doc, buf, &opts);
 			return 1;
 		}
@@ -316,6 +318,7 @@ static int pdfapp_save(pdfapp_t *app)
 
 			fz_try(app->ctx)
 			{
+				wincopyfile(app->docpath, buf);
 				fz_write_document(app->doc, buf, &opts);
 				written = 1;
 			}
diff --git a/platform/x11/pdfapp.h b/platform/x11/pdfapp.h
index 3e40e0c4..0b89923e 100644
--- a/platform/x11/pdfapp.h
+++ b/platform/x11/pdfapp.h
@@ -42,6 +42,7 @@ extern void winalert(pdfapp_t *, pdf_alert_event *alert);
 extern void winprint(pdfapp_t *);
 extern void winadvancetimer(pdfapp_t *, float duration);
 extern void winreplacefile(char *source, char *target);
+extern void wincopyfile(char *source, char *target);
 
 struct pdfapp_s
 {
diff --git a/platform/x11/win_main.c b/platform/x11/win_main.c
index cc01d539..a22e7fc0 100644
--- a/platform/x11/win_main.c
+++ b/platform/x11/win_main.c
@@ -248,6 +248,28 @@ void winreplacefile(char *source, char *target)
 #endif
 }
 
+void wincopyfile(char *source, char *target)
+{
+	wchar_t wsource[PATH_MAX];
+	wchar_t wtarget[PATH_MAX];
+
+	int sz = MultiByteToWideChar(CP_UTF8, 0, source, -1, wsource, PATH_MAX);
+	if (sz == 0)
+	{
+		winerror(&gapp, "cannot convert filename to Unicode");
+		return;
+	}
+
+	sz = MultiByteToWideChar(CP_UTF8, 0, target, -1, wtarget, PATH_MAX);
+	if (sz == 0)
+	{
+		winerror(&gapp, "cannot convert filename to Unicode");
+		return;
+	}
+
+	CopyFile(wsource, wtarget, FALSE);
+}
+
 static char pd_filename[256] = "The file is encrypted.";
 static char pd_password[256] = "";
 static wchar_t pd_passwordw[256] = {0};
diff --git a/platform/x11/x11_main.c b/platform/x11/x11_main.c
index 481adce1..0bcb98f4 100644
--- a/platform/x11/x11_main.c
+++ b/platform/x11/x11_main.c
@@ -281,6 +281,17 @@ void winreplacefile(char *source, char *target)
 	rename(source, target);
 }
 
+void wincopyfile(char *source, char *target)
+{
+	char *buf = malloc(strlen(source)+strlen(target)+4);
+	if (buf)
+	{
+		sprintf(buf, "cp %s %s", source, target);
+		system(buf);
+		free(buf);
+	}
+}
+
 void cleanup(pdfapp_t *app)
 {
 	fz_context *ctx = app->ctx;
diff --git a/source/pdf/pdf-write.c b/source/pdf/pdf-write.c
index 4e6ddb21..63fab4f1 100644
--- a/source/pdf/pdf-write.c
+++ b/source/pdf/pdf-write.c
@@ -46,6 +46,7 @@ typedef struct {
 struct pdf_write_options_s
 {
 	FILE *out;
+	int do_incremental;
 	int do_ascii;
 	int do_expand;
 	int do_garbage;
@@ -1750,16 +1751,11 @@ static void writeobject(pdf_document *doc, pdf_write_options *opts, int num, int
 	pdf_drop_obj(obj);
 }
 
-static void writexref(pdf_document *doc, pdf_write_options *opts, int from, int to, int first, int main_xref_offset, int startxref)
+static void writexrefsubsect(pdf_write_options *opts, int from, int to)
 {
-	pdf_obj *trailer = NULL;
-	pdf_obj *obj;
-	pdf_obj *nobj = NULL;
 	int num;
-	fz_context *ctx = doc->ctx;
 
-	fprintf(opts->out, "xref\n%d %d\n", from, to - from);
-	opts->first_xref_entry_offset = ftell(opts->out);
+	fprintf(opts->out, "%d %d\n", from, to - from);
 	for (num = from; num < to; num++)
 	{
 		if (opts->use_list[num])
@@ -1767,6 +1763,43 @@ static void writexref(pdf_document *doc, pdf_write_options *opts, int from, int
 		else
 			fprintf(opts->out, "%010d %05d f \n", opts->ofs_list[num], opts->gen_list[num]);
 	}
+}
+
+static void writexref(pdf_document *doc, pdf_write_options *opts, int from, int to, int first, int main_xref_offset, int startxref)
+{
+	pdf_obj *trailer = NULL;
+	pdf_obj *obj;
+	pdf_obj *nobj = NULL;
+	fz_context *ctx = doc->ctx;
+
+	fprintf(opts->out, "xref\n");
+	opts->first_xref_entry_offset = ftell(opts->out);
+
+	if (opts->do_incremental)
+	{
+		int subfrom = from;
+		int subto;
+
+		while (subfrom < to)
+		{
+			while (subfrom < to && !pdf_xref_is_incremental(doc, subfrom))
+				subfrom++;
+
+			subto = subfrom;
+			while (subto < to && pdf_xref_is_incremental(doc, subto))
+				subto++;
+
+			if (subfrom < subto)
+				writexrefsubsect(opts, subfrom, subto);
+
+			subfrom = subto;
+		}
+	}
+	else
+	{
+		writexrefsubsect(opts, from, to);
+	}
+
 	fprintf(opts->out, "\n");
 
 	fz_var(trailer);
@@ -1774,33 +1807,43 @@ static void writexref(pdf_document *doc, pdf_write_options *opts, int from, int
 
 	fz_try(ctx)
 	{
-		trailer = pdf_new_dict(doc, 5);
-
-		nobj = pdf_new_int(doc, to);
-		pdf_dict_puts(trailer, "Size", nobj);
-		pdf_drop_obj(nobj);
-		nobj = NULL;
-
-		if (first)
+		if (opts->do_incremental)
 		{
-			obj = pdf_dict_gets(pdf_trailer(doc), "Info");
-			if (obj)
-				pdf_dict_puts(trailer, "Info", obj);
-
-			obj = pdf_dict_gets(pdf_trailer(doc), "Root");
-			if (obj)
-				pdf_dict_puts(trailer, "Root", obj);
-
-			obj = pdf_dict_gets(pdf_trailer(doc), "ID");
-			if (obj)
-				pdf_dict_puts(trailer, "ID", obj);
+			trailer = pdf_keep_obj(pdf_trailer(doc));
+			pdf_dict_puts_drop(trailer, "Size", pdf_new_int(doc, pdf_xref_len(doc)));
+			pdf_dict_puts_drop(trailer, "Prev", pdf_new_int(doc, doc->startxref));
+			doc->startxref = startxref;
 		}
-		if (main_xref_offset != 0)
+		else
 		{
-			nobj = pdf_new_int(doc, main_xref_offset);
-			pdf_dict_puts(trailer, "Prev", nobj);
+			trailer = pdf_new_dict(doc, 5);
+
+			nobj = pdf_new_int(doc, to);
+			pdf_dict_puts(trailer, "Size", nobj);
 			pdf_drop_obj(nobj);
 			nobj = NULL;
+
+			if (first)
+			{
+				obj = pdf_dict_gets(pdf_trailer(doc), "Info");
+				if (obj)
+					pdf_dict_puts(trailer, "Info", obj);
+
+				obj = pdf_dict_gets(pdf_trailer(doc), "Root");
+				if (obj)
+					pdf_dict_puts(trailer, "Root", obj);
+
+				obj = pdf_dict_gets(pdf_trailer(doc), "ID");
+				if (obj)
+					pdf_dict_puts(trailer, "ID", obj);
+			}
+			if (main_xref_offset != 0)
+			{
+				nobj = pdf_new_int(doc, main_xref_offset);
+				pdf_dict_puts(trailer, "Prev", nobj);
+				pdf_drop_obj(nobj);
+				nobj = NULL;
+			}
 		}
 	}
 	fz_always(ctx)
@@ -1861,7 +1904,8 @@ dowriteobject(pdf_document *doc, pdf_write_options *opts, int num, int pass)
 		if (pass > 0)
 			padto(opts->out, opts->ofs_list[num]);
 		opts->ofs_list[num] = ftell(opts->out);
-		writeobject(doc, opts, num, opts->gen_list[num]);
+		if (!opts->do_incremental || pdf_xref_is_incremental(doc, num))
+			writeobject(doc, opts, num, opts->gen_list[num]);
 	}
 	else
 		opts->use_list[num] = 0;
@@ -1873,8 +1917,11 @@ writeobjects(pdf_document *doc, pdf_write_options *opts, int pass)
 	int num;
 	int xref_len = pdf_xref_len(doc);
 
-	fprintf(opts->out, "%%PDF-%d.%d\n", doc->version / 10, doc->version % 10);
-	fprintf(opts->out, "%%\316\274\341\277\246\n\n");
+	if (!opts->do_incremental)
+	{
+		fprintf(opts->out, "%%PDF-%d.%d\n", doc->version / 10, doc->version % 10);
+		fprintf(opts->out, "%%\316\274\341\277\246\n\n");
+	}
 
 	dowriteobject(doc, opts, opts->start, pass);
 
@@ -2241,12 +2288,23 @@ void pdf_write_document(pdf_document *doc, char *filename, fz_write_options *fz_
 
 	xref_len = pdf_xref_len(doc);
 
-	opts.out = fopen(filename, "wb");
+	if (fz_opts->do_incremental)
+	{
+		opts.out = fopen(filename, "ab");
+		if (opts.out)
+			fseek(opts.out, 0, SEEK_END);
+	}
+	else
+	{
+		opts.out = fopen(filename, "wb");
+	}
+
 	if (!opts.out)
 		fz_throw(ctx, FZ_ERROR_GENERIC, "cannot open output file '%s'", filename);
 
 	fz_try(ctx)
 	{
+		opts.do_incremental = fz_opts ? fz_opts->do_incremental : 0;
 		opts.do_expand = fz_opts ? fz_opts->do_expand : 0;
 		opts.do_garbage = fz_opts ? fz_opts->do_garbage : 0;
 		opts.do_ascii = fz_opts ? fz_opts->do_ascii: 0;
@@ -2275,7 +2333,8 @@ void pdf_write_document(pdf_document *doc, char *filename, fz_write_options *fz_
 		}
 
 		/* Make sure any objects hidden in compressed streams have been loaded */
-		preloadobjstms(doc);
+		if (!opts.do_incremental)
+			preloadobjstms(doc);
 
 		/* Sweep & mark objects from the trailer */
 		if (opts.do_garbage >= 1)
@@ -2307,15 +2366,30 @@ void pdf_write_document(pdf_document *doc, char *filename, fz_write_options *fz_
 		dump_object_details(doc, &opts);
 #endif
 
-		/* Construct linked list of free object slots */
-		lastfree = 0;
-		for (num = 0; num < xref_len; num++)
+		if (opts.do_incremental)
 		{
-			if (!opts.use_list[num])
+			for (num = 0; num < xref_len; num++)
 			{
-				opts.gen_list[num]++;
-				opts.ofs_list[lastfree] = num;
-				lastfree = num;
+				if (!opts.use_list[num] && pdf_xref_is_incremental(doc, num))
+				{
+					/* Make unreusable. FIXME: would be better to link to existing free list */
+					opts.gen_list[num] = 65535;
+					opts.ofs_list[num] = 0;
+				}
+			}
+		}
+		else
+		{
+			/* Construct linked list of free object slots */
+			lastfree = 0;
+			for (num = 0; num < xref_len; num++)
+			{
+				if (!opts.use_list[num])
+				{
+					opts.gen_list[num]++;
+					opts.ofs_list[lastfree] = num;
+					lastfree = num;
+				}
 			}
 		}
 
diff --git a/source/pdf/pdf-xref.c b/source/pdf/pdf-xref.c
index 0337bb16..71213fd7 100644
--- a/source/pdf/pdf-xref.c
+++ b/source/pdf/pdf-xref.c
@@ -154,21 +154,34 @@ pdf_xref_entry *pdf_get_xref_entry(pdf_document *doc, int i)
 static void ensure_incremental_xref(pdf_document *doc)
 {
 	fz_context *ctx = doc->ctx;
-	pdf_xref *xref, *pxref;
 
 	if (!doc->xref_altered)
 	{
-		pdf_xref_entry *new_table;
-		doc->xref_sections = fz_resize_array(ctx, doc->xref_sections, doc->num_xref_sections + 1, sizeof(pdf_xref));
-		xref = &doc->xref_sections[0];
-		pxref = &doc->xref_sections[1];
-		new_table = fz_calloc(ctx, xref->len, sizeof(pdf_xref_entry));
-		memmove(pxref, xref, doc->num_xref_sections * sizeof(pdf_xref));
-		doc->num_xref_sections++;
-		/* xref->len is already correct */
-		xref->table = new_table;
-		xref->trailer = pdf_keep_obj(pxref->trailer);
-		doc->xref_altered = 1;
+		pdf_xref *xref = &doc->xref_sections[0];
+		pdf_xref *pxref;
+		pdf_xref_entry *new_table = fz_calloc(ctx, xref->len, sizeof(pdf_xref_entry));
+		pdf_obj *trailer = NULL;
+
+		fz_var(trailer);
+		fz_try(ctx)
+		{
+			trailer = pdf_copy_dict(xref->trailer);
+			doc->xref_sections = fz_resize_array(ctx, doc->xref_sections, doc->num_xref_sections + 1, sizeof(pdf_xref));
+			xref = &doc->xref_sections[0];
+			pxref = &doc->xref_sections[1];
+			memmove(pxref, xref, doc->num_xref_sections * sizeof(pdf_xref));
+			/* xref->len is already correct */
+			xref->table = new_table;
+			xref->trailer = trailer;
+			doc->num_xref_sections++;
+			doc->xref_altered = 1;
+		}
+		fz_catch(ctx)
+		{
+			fz_free(ctx, new_table);
+			pdf_drop_obj(trailer);
+			fz_rethrow(ctx);
+		}
 	}
 }
 
@@ -188,6 +201,13 @@ static pdf_xref_entry *pdf_get_incremental_xref_entry(pdf_document *doc, int i)
 	return &xref->table[i];
 }
 
+int pdf_xref_is_incremental(pdf_document *doc, int num)
+{
+	pdf_xref *xref = &doc->xref_sections[0];
+
+	return doc->xref_altered && num < xref->len && xref->table[num].type;
+}
+
 /* Ensure that an object has been cloned into the incremental xref section */
 void pdf_xref_ensure_incremental_object(pdf_document *doc, int num)
 {
diff --git a/source/tools/pdfclean.c b/source/tools/pdfclean.c
index aac3a888..f42269a8 100644
--- a/source/tools/pdfclean.c
+++ b/source/tools/pdfclean.c
@@ -163,6 +163,7 @@ int pdfclean_main(int argc, char **argv)
 	int write_failed = 0;
 	int errors = 0;
 
+	opts.do_incremental = 0;
 	opts.do_garbage = 0;
 	opts.do_expand = 0;
 	opts.do_ascii = 0;
diff --git a/source/tools/pdfposter.c b/source/tools/pdfposter.c
index 981bf0c0..4702c2aa 100644
--- a/source/tools/pdfposter.c
+++ b/source/tools/pdfposter.c
@@ -133,6 +133,7 @@ int pdfposter_main(int argc, char **argv)
 	pdf_document *doc;
 	fz_context *ctx;
 
+	opts.do_incremental = 0;
 	opts.do_garbage = 0;
 	opts.do_expand = 0;
 	opts.do_ascii = 0;