leafmold writes

I used Recoll to index a bunch of latin1 text files from Project Gutenberg, and found that character accents aren’t handled right. Example from 22403-8.txt:

  • shéer plód makes plough down sillion (from text editor)

  • sh er pl d makes plough down sillion (from Recoll abstract)

  • sh?er pl?d makes plough down sillion (from Recoll preview)

A search for either "sheer plod" or "shéer plód" (with quot. marks) returns nothing. Searching for "sh er pl d" does work, though.

Specs: Recoll 1.16.2, Xapian 1.2.7, Linux Mint 12 32-bit

medoc writes

What is your machine’s locale ? (Type "locale" on the command line)

If it’s not using iso-8859-1 as default encoding, then recoll has no way to know that your text/plain files are iso-8859-1, there is no encoding specification inside them.

In this case, you need to specify a local default encoding for the directory where the files are stored, you can do this in Preferences- >Indexing configuration- >Local parameters inside the GUI, or add something like the following to ~/.recoll/recoll.conf: {{{

defaultcharset = iso-8859-1 }}}

leafmold writes

>What is your machine's locale ? (Type "locale" on the command line)

en_CA.UTF-8

>or add something like the following to /.recoll/recoll.conf

That fixed it. Thank you!