Pdf search doesn't find strings running over lines

Binary package hint: evince

E.g. in the following pdf:

http://www.ico.gov.uk/upload/documents/library/data_protection/detailed_specialist_guides/review_of_eu_dp_directive.pdf

On page 33, paragraph 6:

The string "13 years" runs over the first and onto the second line of this paragraph.

Searching for "13 years" finds no results.

Searching for "13" will find the incidence of this string on this page.

o/s:- Jaunty
App:- Evince 2.26.1 Using poppler 0.10.5 (cairo)

works fine with acrobat it seems it doesn't search for words which are compose with another one on a second line.

it was indeed a poppler issue and known upstream: http://bugs.freedesktop.org/show_bug.cgi?id=11381

This should be high priority.

As it stands, the Ubuntu Document viewer cannot be trusted. The search function does that work.

How many professions can you think of where it is crucial to know whether a string occurs in a document or not?

Law, journalism, health, security, law enforcement.

Need I go on?

Can anyone trust to use the search in Ubuntu's document viewer? Of course they can't they shall have to manually read a document before they can establish whether the text they are looking for is there or not.

Now consider an instance where one of these professionals has to establish that a string DOES NOT appear in a document. These documents can stretch to hundreds of pages. Are you going to sit down and search through manually and trust your judgment when you need to establish the absence of something as a matter of life and death, of national security, of legal importance? Of course you're not. You are going to get another piece of software to do it, aren't you?

This fix should be high priority.

It can be hard to prove a negative.
The same goes for documents.
I.e. Professional people who use .pdf readers sometimes have a need to prove something does not exist in a document.
They can do this by searching for the thing they need to disprove.

But if the search does not work properly because it can't spot strings that fall over lines, and they don't know that this fault exists, they will end up making decisions based on faulty information given them by the software.

This error is still effecting me. It has been this way for four years.

I have a pdf open in Okular. If I search for a term with two words, Okular will not find it if the term runs over a line. So you have to remember to search for either of the words in the term, and lug past all those matches where the single word was found but the term wasn't there. Linux pdf search can't handle two words. It's a joke.

Evince uses poppler for the searching, and the bug agains poppler is reported here https://bugs.freedesktop.org/show_bug.cgi?id=61104