dany/pdfio
1
0
Fork 0
mirror of https://github.com/michaelrsweet/pdfio.git synced 2025-08-29 15:22:06 +02:00
Code Issues Packages Projects Releases Wiki Activity

Increase the maximum number of object streams in a file (Issue #58) - most files

Browse Source 
only contain 1 or 2...

Change the implementation of add/find object to use a custom binary insertion
sort algorithm rather than doing a qsort after every addition.  This results in
a significant improvement in open speed - from 2371 seconds (about 39.5 minutes)
to 3.1 seconds for one large test file (an ESRI standard).
This commit is contained in:
Michael R Sweet
2023-12-13 12:21:59 -05:00
parent 2b92044504
commit 2a85baaf81
3 changed files with 107 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
CHANGES.md
View File

@@ -5,16 +5,20 @@ Changes in PDFio
v1.2.0 (Month DD, YYYY)
-----------------------
- Now use autoconf to configure the PDFio sources (Issue #54)
- Added `pdfioFileCreateNumberObj` and `pdfioFileCreateStringObj` functions
(Issue #14)
- Added `pdfioContentTextMeasure` function (Issue #17)
- Added `pdfioContentTextNewLineShow` and `pdfioContentTextNewLineShowf`
functions (Issue #24)
- Renamed `pdfioContentTextNextLine` to `pdfioContentTextNewLine`.
- Now use autoconf to configure the PDFio sources (Issue #54)
- Updated the maximum number of object streams in a single file from 4096 to
8192 (Issue #58)
- Updated the token reading code to protect against some obvious abuses of the
PDF format.
- Updated the xref reading code to protect against loops.
- Updated the object handling code to use a binary insertion algorithm -
provides a significant (~800x) improvement in open times.
- Fixed handling of encrypted PDFs with per-object file IDs (Issue #42)
- Fixed handling of of trailer dictionaries that started immediately after the
"trailer" keyword (Issue #58)