2
1
mirror of https://github.com/qpdf/qpdf.git synced 2024-12-22 10:58:58 +00:00

Externalize inline images now includes form XObjects

This commit is contained in:
Jay Berkenbilt 2021-01-02 09:47:27 -05:00
parent bedf35d6a5
commit 98da4fd835
9 changed files with 473 additions and 27 deletions

View File

@ -1,3 +1,18 @@
2021-01-02 Jay Berkenbilt <ejb@ql.org>
* QPDFPageObjectHelper::externalizeInlineImages can be called with
form XObjects as well as pages.
* Bug fix: QPDFPageObjectHelper::externalizeInlineImages was not
descending into form XObjects on a page. It now does this by
default. In the extremely unlikely event that anyone was actually
depending on the old behavior, it is available by passing
shallow=true to the externalizeInlineImages call.
* Bug fix: QPDFObjectHandle::filterPageContents was broken for
pages with an array of content streams. This caused
externalize-inline-images to also be broken for this case.
2021-01-01 Jay Berkenbilt <ejb@ql.org> 2021-01-01 Jay Berkenbilt <ejb@ql.org>
* Add methods to QPDFPageObjectHelper: forEachXObject, * Add methods to QPDFPageObjectHelper: forEachXObject,

12
TODO
View File

@ -21,18 +21,6 @@ Candidates for upcoming release
* big page even with --remove-unreferenced-resources=yes, even with --empty * big page even with --remove-unreferenced-resources=yes, even with --empty
* optimize image failure because of colorspace * optimize image failure because of colorspace
* Externalize inline images doesn't walk into form XObjects. In
general:
* Check QPDFPageObjectHelper and see what can be applied to form
XObjects. Maybe think about generalizing it to work with form
XObjects.
* There is an increasing amount of logic in qpdf.cc that should
probably move into the library. This includes externalizing inline
images and page splitting as those operations become more
elaborate, particularly with handling of form XObjects.
* See if the tokenizer is a performance bottleneck and, if so, * See if the tokenizer is a performance bottleneck and, if so,
optimize it. We might end up with a high-performance tokenizer that optimize it. We might end up with a high-performance tokenizer that
has a different interface but still ultimately creates the same has a different interface but still ultimately creates the same

View File

@ -123,8 +123,15 @@ class QPDFPageObjectHelper: public QPDFObjectHelper
QPDF_DLL QPDF_DLL
std::map<std::string, QPDFObjectHandle> getFormXObjects(); std::map<std::string, QPDFObjectHandle> getFormXObjects();
// Convert each inline image to an external (normal) image if the // Converts each inline image to an external (normal) image if the
// size is at least the specified number of bytes. // size is at least the specified number of bytes. This method
// works with pages or form XObjects. By default, it recursively
// processes nested form XObjects. Pass true as shallow to avoid
// this behavior. Prior to qpdf 10.1, form XObjects were ignored,
// but this was considered a bug.
QPDF_DLL
void externalizeInlineImages(size_t min_size, bool shallow);
// ABI: make shallow optional (default false) and merge
QPDF_DLL QPDF_DLL
void externalizeInlineImages(size_t min_size = 0); void externalizeInlineImages(size_t min_size = 0);

View File

@ -486,20 +486,50 @@ QPDFPageObjectHelper::getFormXObjects()
void void
QPDFPageObjectHelper::externalizeInlineImages(size_t min_size) QPDFPageObjectHelper::externalizeInlineImages(size_t min_size)
{ {
QPDFObjectHandle resources = getAttribute("/Resources", true); externalizeInlineImages(min_size, false);
// Calling mergeResources also ensures that /XObject becomes }
// direct and is not shared with other pages.
resources.mergeResources( void
QPDFObjectHandle::parse("<< /XObject << >> >>")); QPDFPageObjectHelper::externalizeInlineImages(size_t min_size, bool shallow)
InlineImageTracker iit(this->oh.getOwningQPDF(), min_size, resources); {
Pl_Buffer b("new page content"); if (shallow)
filterContents(&iit, &b);
if (iit.any_images)
{ {
getObjectHandle().replaceKey( QPDFObjectHandle resources = getAttribute("/Resources", true);
"/Contents", // Calling mergeResources also ensures that /XObject becomes
QPDFObjectHandle::newStream( // direct and is not shared with other pages.
this->oh.getOwningQPDF(), b.getBuffer())); resources.mergeResources(
QPDFObjectHandle::parse("<< /XObject << >> >>"));
InlineImageTracker iit(this->oh.getOwningQPDF(), min_size, resources);
Pl_Buffer b("new page content");
filterContents(&iit, &b);
if (iit.any_images)
{
if (this->oh.isFormXObject())
{
this->oh.replaceStreamData(
b.getBuffer(),
QPDFObjectHandle::newNull(),
QPDFObjectHandle::newNull());
}
else
{
this->oh.replaceKey(
"/Contents",
QPDFObjectHandle::newStream(
this->oh.getOwningQPDF(), b.getBuffer()));
}
}
}
else
{
externalizeInlineImages(min_size, true);
forEachFormXObject(
true,
[min_size](QPDFObjectHandle& obj,
QPDFObjectHandle&, std::string const&) {
QPDFPageObjectHelper(obj).externalizeInlineImages(
min_size, true);
});
} }
} }

View File

@ -4991,6 +4991,26 @@ print "\n";
</listitem> </listitem>
</itemizedlist> </itemizedlist>
</listitem> </listitem>
<listitem>
<para>
Bug Fixes
</para>
<itemizedlist>
<listitem>
<para>
<function>QPDFPageObjectHelper::externalizeInlineImages</function>
was not externalizing images referenced from form XObjects
that appeared on the page.
</para>
</listitem>
<listitem>
<para>
<function>QPDFObjectHandle::filterPageContents</function>
was broken for pages with multiple content streams.
</para>
</listitem>
</itemizedlist>
</listitem>
</itemizedlist> </itemizedlist>
</listitem> </listitem>
</varlistentry> </varlistentry>

View File

@ -905,6 +905,7 @@ $td->runtest("check output",
my @eii_tests = ( my @eii_tests = (
['inline-images', 80], ['inline-images', 80],
['large-inline-image', 1024], ['large-inline-image', 1024],
['nested-form-xobjects-inline-images', 20],
); );
$n_tests += 4 * scalar(@eii_tests); $n_tests += 4 * scalar(@eii_tests);
$n_compare_pdfs += 2 * scalar(@eii_tests); $n_compare_pdfs += 2 * scalar(@eii_tests);

View File

@ -0,0 +1,385 @@
%PDF-1.3
%¿÷¢þ
%QDF-1.0
%% Original object ID: 1 0
1 0 obj
<<
/Pages 2 0 R
/Type /Catalog
>>
endobj
%% Original object ID: 2 0
2 0 obj
<<
/Count 1
/Kids [
3 0 R
]
/Type /Pages
>>
endobj
%% Page 1
%% Original object ID: 3 0
3 0 obj
<<
/Contents 4 0 R
/MediaBox [
0
0
612
792
]
/Parent 2 0 R
/Resources <<
/Font <<
/F1 6 0 R
>>
/ProcSet 7 0 R
/XObject <<
/Fx1 8 0 R
/IIm1 10 0 R
/IIm2 12 0 R
>>
>>
/Type /Page
>>
endobj
%% Contents for page 1
%% Original object ID: 20 0
4 0 obj
<<
/Length 5 0 R
>>
stream
q
BT
/F1 24 Tf
72 720 Td
(Page) Tj
ET
q
100 0 0 100 72 600 cm
/IIm1 Do
Q
q
100 0 0 100 192 600 cm
/IIm2 Do
Q
Q
q
1.00000 0.00000 0.00000 1.00000 72.00000 200.00000 cm
/Fx1 Do
Q
endstream
endobj
5 0 obj
186
endobj
%% Original object ID: 10 0
6 0 obj
<<
/BaseFont /Helvetica
/Encoding /WinAnsiEncoding
/Name /F1
/Subtype /Type1
/Type /Font
>>
endobj
%% Original object ID: 11 0
7 0 obj
[
/PDF
/Text
/ImageC
]
endobj
%% Original object ID: 12 0
8 0 obj
<<
/BBox [
0
0
300
500
]
/Resources <<
/Font <<
/F1 14 0 R
>>
/ProcSet 15 0 R
/XObject <<
/Fx1 16 0 R
/IIm1 18 0 R
/IIm2 20 0 R
>>
>>
/Subtype /Form
/Type /XObject
/Length 9 0 R
>>
stream
BT
/F1 24 Tf
0 320 Td
(FX1) Tj
ET
q
100 0 0 100 000 200 cm
/IIm1 Do
Q
q
100 0 0 100 120 200 cm
/IIm2 Do
Q
q
1.00000 0.00000 0.00000 1.00000 0.00000 0.00000 cm
/Fx1 Do
Q
endstream
endobj
9 0 obj
177
endobj
%% Original object ID: 18 0
10 0 obj
<<
/BitsPerComponent 8
/ColorSpace /DeviceGray
/Height 15
/Subtype /Image
/Type /XObject
/Width 15
/Length 11 0 R
>>
stream
`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
endstream
endobj
%QDF: ignore_newline
11 0 obj
225
endobj
%% Original object ID: 19 0
12 0 obj
<<
/BitsPerComponent 8
/ColorSpace /DeviceGray
/Height 15
/Subtype /Image
/Type /XObject
/Width 15
/Length 13 0 R
>>
stream
@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
endstream
endobj
%QDF: ignore_newline
13 0 obj
225
endobj
%% Original object ID: 14 0
14 0 obj
<<
/BaseFont /Helvetica
/Encoding /WinAnsiEncoding
/Name /F1
/Subtype /Type1
/Type /Font
>>
endobj
%% Original object ID: 15 0
15 0 obj
[
/PDF
/Text
/ImageC
]
endobj
%% Original object ID: 16 0
16 0 obj
<<
/BBox [
0
0
300
200
]
/Resources <<
/Font <<
/F1 14 0 R
>>
/ProcSet 15 0 R
/XObject <<
/IIm1 22 0 R
/IIm2 24 0 R
>>
>>
/Subtype /Form
/Type /XObject
/Length 17 0 R
>>
stream
BT
/F1 24 Tf
0 120 Td
(FX2) Tj
ET
q
100 0 0 100 0 0 cm
/IIm1 Do
Q
q
100 0 0 100 120 0 cm
/IIm2 Do
Q
endstream
endobj
17 0 obj
108
endobj
%% Original object ID: 21 0
18 0 obj
<<
/BitsPerComponent 8
/ColorSpace /DeviceGray
/Height 15
/Subtype /Image
/Type /XObject
/Width 15
/Length 19 0 R
>>
stream
@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
endstream
endobj
%QDF: ignore_newline
19 0 obj
225
endobj
%% Original object ID: 22 0
20 0 obj
<<
/BitsPerComponent 8
/ColorSpace /DeviceGray
/Height 15
/Subtype /Image
/Type /XObject
/Width 15
/Length 21 0 R
>>
stream
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
endstream
endobj
%QDF: ignore_newline
21 0 obj
225
endobj
%% Original object ID: 23 0
22 0 obj
<<
/BitsPerComponent 8
/ColorSpace /DeviceGray
/Height 15
/Subtype /Image
/Type /XObject
/Width 15
/Length 23 0 R
>>
stream
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
endstream
endobj
%QDF: ignore_newline
23 0 obj
225
endobj
%% Original object ID: 24 0
24 0 obj
<<
/BitsPerComponent 8
/ColorSpace /DeviceGray
/Height 15
/Subtype /Image
/Type /XObject
/Width 15
/Length 25 0 R
>>
stream
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@`````@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
endstream
endobj
%QDF: ignore_newline
25 0 obj
225
endobj
xref
0 26
0000000000 65535 f
0000000052 00000 n
0000000133 00000 n
0000000242 00000 n
0000000563 00000 n
0000000804 00000 n
0000000852 00000 n
0000000998 00000 n
0000001071 00000 n
0000001536 00000 n
0000001584 00000 n
0000001996 00000 n
0000002045 00000 n
0000002457 00000 n
0000002506 00000 n
0000002653 00000 n
0000002727 00000 n
0000003107 00000 n
0000003156 00000 n
0000003568 00000 n
0000003617 00000 n
0000004029 00000 n
0000004078 00000 n
0000004490 00000 n
0000004539 00000 n
0000004951 00000 n
trailer <<
/Root 1 0 R
/Size 26
/ID [<55269d37282af9edc76855e4cb859987><31415926535897932384626433832795>]
>>
startxref
4972
%%EOF

Binary file not shown.