tris/recoll
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

rclpdf: also escape text inside meta content attributes

Browse Source
This commit is contained in:
Jean-Francois Dockes 2014-08-25 14:16:45 +02:00
parent 26c7e1c690
commit 552eb0965b
1 changed files with 19 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
src/filters/rclpdf
View File

@ -1,22 +1,18 @@
#!/bin/sh #!/bin/sh
# @(#$Id: rclpdf,v 1.10 2007-07-12 17:13:38 dockes Exp $ (C) 2004 J.F.Dockes
# This is copied almost verbatim from Estraier:
#================================================================ #================================================================
# Some parts are Copyright Estraier (GPL v2).
# Estraier: a personal full-text search system # Estraier: a personal full-text search system
# Copyright (C) 2003-2004 Mikio Hirabayashi # Copyright (C) 2003-2004 Mikio Hirabayashi
# Copyright (C) 2014 J.F. Dockes
# This file is licensed under the GPL v2
#================================================================ #================================================================
#================================================================ #================================================================
# Convert a pdf file to HTML. # Convert a pdf file to HTML.
# #
# We use pdftotxt from the xpdf package. This does not perfect results as # We use pdftotext from the xpdf/poppler-utils package.
# whitespace is sometimes either arbitrarily inserted or stripped from the
# text. This seems to depend on the usage of option -raw, and,
# unfortunately also of the document itself, so that there does not seem to
# be an universally good solution
# #
# Also, the filter sometimes seems to output problematic utf-8. I did not # pdftotext sometimes outputs unescaped text inside HTML text sections.
# check if it was actually incorrect or just mis-understood by qtextedit # We try to correct.
# (tobedone)
# Uncomment the following if you get better results without. The # Uncomment the following if you get better results without. The
# pdftotext manual says that the option is no longer recommended The # pdftotext manual says that the option is no longer recommended The
@ -133,6 +129,19 @@ awk 'BEGIN'\
mid = "<title>" mid "</title>" mid = "<title>" mid "</title>"
$0 = part1 mid part2 $0 = part1 mid part2
} }
if(doescape == 0 && $0 ~ /content=".*"\/>/){
match($0, /content=".*"\/>/)
part1 = substr($0, 0, RSTART-1)
mid = substr($0, RSTART, RLENGTH)
part2 = substr($0, RSTART + RLENGTH, length($0))
gsub(/content="/, "", mid)
gsub(/"\/>/, "", mid)
gsub(/&/, "\\&amp;", mid)
gsub(/</, "\\&lt;", mid)
gsub(/>/, "\\&gt;", mid)
mid = "content=\"" mid "\"/>"
$0 = part1 mid part2
}
# Recoll treats "Subject" as a "title" element (based on emails). The PDF # Recoll treats "Subject" as a "title" element (based on emails). The PDF
# "Subject" metadata field is more like an HTML "description" # "Subject" metadata field is more like an HTML "description"