[Digikam-users] Words match in digiKam search

[Jean-François Rabasse]

Hello,

On Sun, 28 Aug 2011, Gilles Caulier wrote:

> Date: Sun, 28 Aug 2011 00:01:40 +0200
> From: Gilles Caulier <caulier.gilles at gmail.com>
> Reply-To: digiKam - Home Manage your photographs as a professional with the
>     power of open source <digikam-users at kde.org>
> To: digiKam - Home Manage your photographs as a professional with the power of
>      open source <digikam-users at kde.org>
> Subject: Re: [Digikam-users] Words match in digiKam search
> 
> 2011/8/28 Jean-François Rabasse <jean-francois.rabasse at wanadoo.fr>:
>>
>> Hello,
>>
>> I have some questions about the way digiKam performs strings
>> comparisons when searching for keywords.
>>
>> 1. The digiKam handbook says "searches are case insensitive".
>> Well, doesn't seem to be the case in my environment (digiKam 1.2.0,
>> Linux OpenSuSE 11.3 with KDE 4) with non US ASCII characters.
>>
>> I have several pictures with captions containing the word "église"
>> (for non French readers, église is a church). E.g. picture 1 is
>> titled "Petite église rouge", and picture 2 is titled "Église à
>> vendre". (With an uppercase É because all my titles are
>> capitalised.)
>>
>> Now, if I type (in the words search input field) the word "église",
>> I get picture 1 but not picture 2. And if I type "Église", I get
>> picture 2 but not picture 1. Clearly a case sensitive match,
>> concerning only the ISO Latin characters (typing "éGLISE" finds
>> picture 1 too, and "ÉGLISE" finds picture 2).
>>
>> I suspect it could be a problem with non ASCII characters encoding.
>> My desktop environment is set to French, ISO-8859-1 charset.
>> How digiKam internally encodes its text strings ? UTF-8 maybe ?
>> This could possibly explain a mismatch.
>>
>> Has anyone already seen such an issue, and what could be the best
>> way to walk around ? Is there a way to tell digiKam about user input
>> encoding, us-ascii, iso-8859-x, utf-8 ? Or is it shadowed by the
>> X11/Qt/KDE layers ?
>>
>> 2. Another question related to strings match, just because I'm
>> curious :-)
>> In the digiKam configuration window, folder "Miscellaneous",
>> there's a configuration parameter named "String comparison type",
>> with a selection menu proposing two options, Natural or Normal.
>>
>> I couldn't find anything about that in the handbook. What is a
>> "Normal" comparison type, and a "Natural" comparison type ?
>>
>
> SHIFT+F1 over the drop-down menu doesn't help ?
>
> Gilles Caulier

Well, doesn't help a lot :-(
Shift-F1 says : "Not defined. There no "What's This" assigned to this
widget".
And the menu comparison options menu, Natural vs. Normal, has a tooltip
but not very readable. Tooltip text starts with "Sets the way in which
strings are compared inside digiKam, etc.", but seems to be a long text
and I can't have it displayed correctly. I can see that text embeds some
XHTML formating, e.g. "<br/>" to trig a line break. And this markups are
juste displayed as raw text, not processed, so the whole tooltip text
appears on one long line, truncated at the right edge of my screen.

But this was just a matter of curiousness.
My major problem is the case insensitive search when words contains
non US-ASCII characters.
Could be a problem with my graphic interface setup, or not.
Could non english writers digiKam users confirm - or not - that case
insensitive search works well with their country charset ?
If yes, I'll investigate in my X11/Qt/KDE configuration.
If not, I'll just do my searches twice, typing the lowercase version
then uppercase version of words containing Latin1 characters :-)

Thanks in advance,
Jean-François