Re: [Tracker] searching for words within MSWord docs

[Ivan Frade <ivan frade nokia com>]

Hi Jin,

El jue, 21-08-2008 a las 11:09 -0700, ext Jin Suh escribiÃ:
> 1. I can't find words within MSWord or PDF documents. It doesn't find any but it finds words within HTML.
>  I have installed the wv, libgsf, poppler, poppler-utils. The configure script found the libgsf.

1) Check that you have "wvText" and "pdftotext" installed your system. 
2) Call the filter in /usr/lib/tracker/filters/applications/pdf_filter
manually to check the result

 If both things are ok, then the problem could be that tracker is not
calling the extractor. In that case, please fill a bug.

> 2. I have a word list in metadata of a xls file. The tracker-search-tool finds only the first word out of 
> the list not for the 2nd line or other keywords. For example, I have 
> cat
> dog
> cow
> Then, I can only search for cat. When I enter the word "dog", it says, did you mean "ndog"? The metadata 
> has "cat\r\ndog\r\n". And the rest words are the same way.

 Check the output of
tracker-extract my-file.xls application/vnd.ms-excel

 You should see something like
MetadataX=cat dog cow;

 We must check:
1) The output of the extractor is correct (could be a problem in the
"msoffice" module of the extract)
2) That metadata type support multiple values. Check the definition of
that metadata type in the /usr/local/share/tracker/services/*.metadata
files.

 Please fill a bug with this information and attach a small (few Kb)
excel file we can use for test.

> Please let me know what I am missing.

 Thank you very much for the feedback!

Ivan