Cache object numbers in CPDF_Parser::ParseIndirectObject().

Cache object numbers in CPDF_Parser::ParseIndirectObject().

R=jun_fang@foxitsoftware.com, tsepez@chromium.org

Committed: https://pdfium.googlesource.com/pdfium/+/0c8e6c1f39f20985a3efb17292e46c476194183a

[Lei Zhang, 2015-11-19 01:03:16]

thestig@chromium.org changed reviewers:
+ jun_fang@foxitsoftware.com, tsepez@chromium.org

[Lei Zhang, 2015-11-19 01:03:17]

With ParseIndirectObject(), the number of bytes that passes through its
CPDF_SyntaxParser increases, as callers call it with increasing object numbers.
For some large PDFs, like https://html.spec.whatwg.org/print.pdf from
https://crbug.com/498709, PDFium spends 70% of its time (re)parsing.

This CL adds an object cache and improves the out/Release/pdfium_test print.pdf
run time from 5m 45s to 8s on my workstation.

[jun_fang, 2015-11-19 16:19:45]

https://codereview.chromium.org/1458633004/diff/1/core/src/fpdfapi/fpdf_parse...
File core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp (right):

https://codereview.chromium.org/1458633004/diff/1/core/src/fpdfapi/fpdf_parse...
core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp:1215:
syntax.RestorePos(saved_pos);
nit: It seems that there is no need to restore the position here.

[jun_fang, 2015-11-19 16:20:40]

On 2015/11/19 16:19:45, jun_fang wrote:
>
https://codereview.chromium.org/1458633004/diff/1/core/src/fpdfapi/fpdf_parse...
> File core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp (right):
> 
>
https://codereview.chromium.org/1458633004/diff/1/core/src/fpdfapi/fpdf_parse...
> core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp:1215:
> syntax.RestorePos(saved_pos);
> nit: It seems that there is no need to restore the position here.

LGTM

[Tom Sepez, 2015-11-19 16:46:25]

lgtm

https://codereview.chromium.org/1458633004/diff/1/core/include/fpdfapi/fpdf_p...
File core/include/fpdfapi/fpdf_parser.h (right):

https://codereview.chromium.org/1458633004/diff/1/core/include/fpdfapi/fpdf_p...
core/include/fpdfapi/fpdf_parser.h:547: CFX_MapPtrToPtr m_ObjectStreamMap;
nit: wish we had a comment about what this map does, and whether it can become a
std::map.

[Lei Zhang, 2015-11-20 07:11:42]

https://codereview.chromium.org/1458633004/diff/1/core/include/fpdfapi/fpdf_p...
File core/include/fpdfapi/fpdf_parser.h (right):

https://codereview.chromium.org/1458633004/diff/1/core/include/fpdfapi/fpdf_p...
core/include/fpdfapi/fpdf_parser.h:547: CFX_MapPtrToPtr m_ObjectStreamMap;
On 2015/11/19 16:46:25, Tom Sepez wrote:
> nit: wish we had a comment about what this map does, and whether it can become
a
> std::map.

Added, and likely yes.

https://codereview.chromium.org/1458633004/diff/1/core/src/fpdfapi/fpdf_parse...
File core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp (right):

https://codereview.chromium.org/1458633004/diff/1/core/src/fpdfapi/fpdf_parse...
core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp:1215:
syntax.RestorePos(saved_pos);
On 2015/11/19 16:19:45, jun_fang wrote:
> nit: It seems that there is no need to restore the position here.   

Done. I needed it in an earlier iteration, but then I arranged the code and it
was cut + pasted.

[Lei Zhang, 2015-11-20 07:12:09]

Committed patchset #2 (id:20001) manually as
0c8e6c1f39f20985a3efb17292e46c476194183a (presubmit successful).