Web   ·   Wiki   ·   Activities   ·   Blog   ·   Lists   ·   Chat   ·   Meeting   ·   Bugs   ·   Git   ·   Translate   ·   Archive   ·   People   ·   Donate
summaryrefslogtreecommitdiffstats
path: root/pdf
diff options
context:
space:
mode:
authorHavoc Pennington <hp@redhat.com>2004-12-23 05:12:59 (GMT)
committer Havoc Pennington <hp@src.gnome.org>2004-12-23 05:12:59 (GMT)
commit5cea5eb0365ac715463c4d6fc893ed4154821e42 (patch)
tree3999b0248e36fa2ae9065bf347912761a53f693d /pdf
parentec2aa10ab1ec9246d6c35b90f5d1df372f39c732 (diff)
make search find stuff on other pages, sort of (only returns one result on
2004-12-22 Havoc Pennington <hp@redhat.com> * pdf/xpdf/pdf-document.cc (pdf_document_find_begin): make search find stuff on other pages, sort of (only returns one result on invisible pages, to show they have results; updates full results for a page when you view it). Currently repaints the current page every time a new result is found on any page, which isn't so nice.
Diffstat (limited to 'pdf')
-rw-r--r--pdf/xpdf/pdf-document.cc293
1 files changed, 258 insertions, 35 deletions
diff --git a/pdf/xpdf/pdf-document.cc b/pdf/xpdf/pdf-document.cc
index efee462..64d592c 100644
--- a/pdf/xpdf/pdf-document.cc
+++ b/pdf/xpdf/pdf-document.cc
@@ -1,4 +1,4 @@
-/* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8; c-indent-level: 8 -*- */
+/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8; c-indent-level: 8 -*- */
/* pdfdocument.h: Implementation of EvDocument for PDF
* Copyright (C) 2004, Red Hat, Inc.
*
@@ -28,6 +28,21 @@
#include "PDFDoc.h"
#include "PSOutputDev.h"
+typedef struct
+{
+ PdfDocument *document;
+ gunichar *ucs4;
+ glong ucs4_len;
+ guint idle;
+ /* full results are only possible for the rendered current page */
+ int current_page;
+ GArray *current_page_results;
+ guchar *other_page_flags; /* length n_pages + 1, first element ignored */
+ int start_page; /* skip this one as we iterate, since we did it first */
+ int search_page; /* the page we're searching now */
+ TextOutputDev *output_dev;
+} PdfDocumentSearch;
+
typedef struct _PdfDocumentClass PdfDocumentClass;
#define PDF_DOCUMENT_CLASS(klass) (G_TYPE_CHECK_CLASS_CAST ((klass), PDF_TYPE_DOCUMENT, PdfDocumentClass))
@@ -54,11 +69,15 @@ struct _PdfDocument
PDFDoc *doc;
gboolean page_valid;
+
+ PdfDocumentSearch *search;
};
static void pdf_document_document_iface_init (EvDocumentIface *iface);
static void pdf_document_ps_exporter_iface_init (EvPSExporterIface *iface);
static void pdf_document_find_iface_init (EvDocumentFindIface *iface);
+static void pdf_document_search_free (PdfDocumentSearch *search);
+static void pdf_document_search_page_changed (PdfDocumentSearch *search);
G_DEFINE_TYPE_WITH_CODE (PdfDocument, pdf_document, G_TYPE_OBJECT,
{
@@ -80,6 +99,12 @@ document_validate_page (PdfDocument *pdf_document)
0, gTrue, gTrue);
pdf_document->page_valid = TRUE;
+
+ /* Update the search results available to the app since
+ * we only provide full results on the current page
+ */
+ if (pdf_document->search)
+ pdf_document_search_page_changed (pdf_document->search);
}
}
@@ -285,38 +310,86 @@ pdf_document_render (EvDocument *document,
}
static void
-pdf_document_find_begin (EvDocumentFind *document,
- const char *search_string,
- gboolean case_sensitive)
+pdf_document_search_emit_found (PdfDocumentSearch *search)
{
- /* FIXME make this incremental (idle handler) and multi-page */
- /* Right now it's fully synchronous plus only does the current page */
-
- PdfDocument *pdf_document = PDF_DOCUMENT (document);
- gunichar *ucs4;
- glong ucs4_len;
- int xMin, yMin, xMax, yMax;
- GArray *results;
- EvFindResult result;
+ PdfDocument *pdf_document = search->document;
+ int n_pages;
+ double pages_done;
+ GArray *tmp_results;
+ int i;
+
+ n_pages = ev_document_get_n_pages (EV_DOCUMENT (search->document));
+ if (search->search_page > search->start_page) {
+ pages_done = search->search_page - search->start_page;
+ } else {
+ pages_done = n_pages - search->start_page + search->search_page;
+ }
- /* FIXME case_sensitive (right now XPDF
- * code is always case insensitive for ASCII
- * and case sensitive for all other languaages)
+ tmp_results = g_array_new (FALSE, FALSE, sizeof (EvFindResult));
+ g_array_append_vals (tmp_results,
+ search->current_page_results->data,
+ search->current_page_results->len);
+
+ /* Now append a bogus element for each page that has a result in it,
+ * that is not the current page
*/
+ i = 1;
+ while (i <= n_pages) {
+ if (i != pdf_document->page &&
+ search->other_page_flags[i]) {
+ EvFindResult result;
+
+ result.page_num = i;
+
+ /* Use bogus coordinates, again we can't get coordinates
+ * until this is the current page because TextOutputDev
+ * isn't good enough
+ */
+ result.highlight_area.x = -1;
+ result.highlight_area.y = -1;
+ result.highlight_area.width = 1;
+ result.highlight_area.height = 1;
+
+ g_array_append_val (tmp_results, result);
+ }
+
+ ++i;
+ }
- g_assert (sizeof (gunichar) == sizeof (Unicode));
- ucs4 = g_utf8_to_ucs4_fast (search_string, -1,
- &ucs4_len);
+ ev_document_find_found (EV_DOCUMENT_FIND (pdf_document),
+ (EvFindResult*) tmp_results->data,
+ tmp_results->len,
+ pages_done / (double) n_pages);
+
+ g_array_free (tmp_results, TRUE);
+}
+
+static void
+pdf_document_search_page_changed (PdfDocumentSearch *search)
+{
+ PdfDocument *pdf_document = search->document;
+ int current_page;
+ EvFindResult result;
+ int xMin, yMin, xMax, yMax;
- results = g_array_new (FALSE,
- FALSE,
- sizeof (EvFindResult));
+ current_page = pdf_document->page;
- if (pdf_document->out->findText (ucs4, ucs4_len,
+ if (!pdf_document->page_valid) {
+ /* we can't do anything until displayPage() */
+ search->current_page = -1;
+ return;
+ }
+
+ if (search->current_page == current_page)
+ return;
+
+ /* We need to create current_page_results for the new current page */
+ g_array_set_size (search->current_page_results, 0);
+
+ if (pdf_document->out->findText (search->ucs4, search->ucs4_len,
gTrue, gTrue, // startAtTop, stopAtBottom
gFalse, gFalse, // startAtLast, stopAtLast
&xMin, &yMin, &xMax, &yMax)) {
-
result.page_num = pdf_document->page;
result.highlight_area.x = xMin;
@@ -324,11 +397,11 @@ pdf_document_find_begin (EvDocumentFind *document,
result.highlight_area.width = xMax - xMin;
result.highlight_area.height = yMax - yMin;
- g_array_append_val (results, result);
+ g_array_append_val (search->current_page_results, result);
/* Now find further results */
- while (pdf_document->out->findText (ucs4, ucs4_len,
+ while (pdf_document->out->findText (search->ucs4, search->ucs4_len,
gFalse, gTrue,
gTrue, gFalse,
&xMin, &yMin, &xMax, &yMax)) {
@@ -340,26 +413,173 @@ pdf_document_find_begin (EvDocumentFind *document,
result.highlight_area.width = xMax - xMin;
result.highlight_area.height = yMax - yMin;
- g_array_append_val (results, result);
+ g_array_append_val (search->current_page_results, result);
+ }
+ }
+
+ /* needed for the initial current page since we don't search
+ * it in the idle
+ */
+ search->other_page_flags[current_page] =
+ search->current_page_results->len > 0;
+
+ pdf_document_search_emit_found (search);
+}
+
+static gboolean
+pdf_document_search_idle_callback (void *data)
+{
+ PdfDocumentSearch *search = (PdfDocumentSearch*) data;
+ PdfDocument *pdf_document = search->document;
+ int n_pages;
+ double xMin, yMin, xMax, yMax;
+ gboolean found;
+
+ /* Note that PDF page count is 1 through n_pages INCLUSIVE
+ * like a real book. We are looking to add one result for each
+ * page with a match, because the coordinates are meaningless
+ * with TextOutputDev, so we just want to flag matching pages
+ * and then when the user switches to the current page, we
+ * will emit "found" again with the real results.
+ */
+ n_pages = ev_document_get_n_pages (EV_DOCUMENT (search->document));
+
+ if (search->search_page == search->start_page) {
+ goto end_search;
+ }
+
+ if (search->output_dev == 0) {
+ /* First time through here... */
+ search->output_dev = new TextOutputDev (NULL, gTrue, gFalse, gFalse);
+ if (!search->output_dev->isOk()) {
+ goto end_search;
}
}
+
+ pdf_document->doc->displayPage (search->output_dev,
+ search->search_page,
+ 72, 72, 0, gTrue, gFalse);
+
+ if (search->output_dev->findText (search->ucs4,
+ search->ucs4_len,
+ gTrue, gTrue, // startAtTop, stopAtBottom
+ gFalse, gFalse, // startAtLast, stopAtLast
+ &xMin, &yMin, &xMax, &yMax)) {
+ /* This page has results */
+ search->other_page_flags[search->search_page] = TRUE;
+
+ pdf_document_search_emit_found (search);
+ }
- ev_document_find_found (document,
- (EvFindResult*) results->data,
- results->len,
- 1.0);
+ search->search_page += 1;
+ if (search->search_page > n_pages) {
+ /* wrap around */
+ search->search_page = 1;
+ }
+
+ return TRUE;
- g_array_free (results, TRUE);
+ end_search:
+ /* We're done. */
+ search->idle = 0; /* will return FALSE to remove */
+ return FALSE;
}
static void
-pdf_document_find_cancel (EvDocumentFind *document)
+pdf_document_find_begin (EvDocumentFind *document,
+ const char *search_string,
+ gboolean case_sensitive)
{
PdfDocument *pdf_document = PDF_DOCUMENT (document);
+ PdfDocumentSearch *search;
+ int n_pages;
+ gunichar *ucs4;
+ glong ucs4_len;
- /* FIXME this will do something once begin_find queues
- * an incremental find
+ /* FIXME handle case_sensitive (right now XPDF
+ * code is always case insensitive for ASCII
+ * and case sensitive for all other languaages)
*/
+
+ g_assert (sizeof (gunichar) == sizeof (Unicode));
+ ucs4 = g_utf8_to_ucs4_fast (search_string, -1,
+ &ucs4_len);
+
+ if (pdf_document->search &&
+ pdf_document->search->ucs4_len == ucs4_len &&
+ memcmp (pdf_document->search->ucs4,
+ ucs4,
+ sizeof (gunichar) * ucs4_len) == 0) {
+ /* Search is unchanged */
+ g_free (ucs4);
+ return;
+ }
+
+ if (pdf_document->search) {
+ pdf_document_search_free (pdf_document->search);
+ pdf_document->search = NULL;
+ }
+
+ search = g_new0 (PdfDocumentSearch, 1);
+
+ search->ucs4 = ucs4;
+ search->ucs4_len = ucs4_len;
+
+ search->current_page_results = g_array_new (FALSE,
+ FALSE,
+ sizeof (EvFindResult));
+ n_pages = ev_document_get_n_pages (EV_DOCUMENT (document));
+
+ /* This is an array of bool; with the first value ignored
+ * so we can index by the based-at-1 page numbers
+ */
+ search->other_page_flags = g_new0 (guchar, n_pages + 1);
+
+ search->document = pdf_document;
+
+ search->idle = g_idle_add (pdf_document_search_idle_callback,
+ search);
+
+ search->output_dev = 0;
+
+ search->start_page = pdf_document->page;
+ search->search_page = search->start_page + 1;
+ if (search->search_page > n_pages)
+ search->search_page = 1;
+
+ search->current_page = -1;
+
+ pdf_document->search = search;
+
+ /* Update for the current page right away */
+ pdf_document_search_page_changed (search);
+}
+
+static void
+pdf_document_find_cancel (EvDocumentFind *document)
+{
+ PdfDocument *pdf_document = PDF_DOCUMENT (document);
+
+ if (pdf_document->search) {
+ pdf_document_search_free (pdf_document->search);
+ pdf_document->search = NULL;
+ }
+}
+
+static void
+pdf_document_search_free (PdfDocumentSearch *search)
+{
+ if (search->idle != 0)
+ g_source_remove (search->idle);
+
+ if (search->output_dev)
+ delete search->output_dev;
+
+ g_array_free (search->current_page_results, TRUE);
+ g_free (search->other_page_flags);
+
+ g_free (search->ucs4);
+ g_free (search);
}
static void
@@ -399,6 +619,9 @@ pdf_document_finalize (GObject *object)
{
PdfDocument *pdf_document = PDF_DOCUMENT (object);
+ if (pdf_document->search)
+ pdf_document_search_free (pdf_document->search);
+
if (pdf_document->target)
g_object_unref (pdf_document->target);