506064 - Thunderbird filters should ignore the accents

Status: RESOLVED DUPLICATE of bug 284856

(MailNews Core :: Search, defect)

Description

[Fernando García Gómez, stripTM]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; es-ES; rv:1.9.1.1) Gecko/20090715 Firefox/3.5.1
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; es-ES; rv:1.9.1.1) Gecko/20090715 Thunderbird/3.0b3

Thunderbird should ignore whether a word has or not an accent when performing a search, in the same way that it is not "case sensitive".

For example, if there are two messages with the following subjects: "Vacaciones en París" and "hoteles paris", a search for "París" should return both results.

In Spanish language, unfortunately, there are too many people who writes without the accents, then it is important that á é í ó ú ü are equivalent to a a e i o u u.

Reproducible: Always

Steps to Reproduce:
1. Inbox with a message that contains an accented word in the subjet or receiver, ej. 'parís'
2. In search box write 'paris'
Actual Results:  
The message is not found, because it thinks that parís != paris

Expected Results:  
The message with is founded 'paris'

Comment 1

[Kent James (:rkent)]

This is the kind of search customization that is well suited to the new custom search terms functionality from bug 495519.

Comment 2

[rsx11m]

Well, this sounds like a good idea in general not just customization, where I'm wondering though if there are clear transcription rules for all character sets.

Comment 3

[RickieES aka Ricardo Palomares]

Confirmed also in SeaMonkey 2.0b1 (Mozilla/5.0 (X11; U; Linux i686; es-ES; rv:1.9.1.1pre) Gecko/20090717 SeaMonkey/2.0b1).

Comment 4

[Ludovic Hirlimann [:Usul]]

I was unable to find a dup - marking this RFE as valid.

Comment 5

[Wayne Mery (:wsmwk)]

Fernando, you issue is about the old "Search Messages", or the new "Search all messages"

Comment 6

[Fernando García Gómez, stripTM]

Wayne, Is that a question?

My issue is about the search box on the top-right corner of Thunderbird. For example when selecting "By subject..." and then typing "guia" a message with the subject "Guía para principiantes" does not match the search.

Comment 7

[Wayne Mery (:wsmwk)]

(In reply to comment #6)
> Wayne, Is that a question?
> 
> My issue is about the search box on the top-right corner of Thunderbird. For
> example when selecting "By subject..." and then typing "guia" a message with
> the subject "Guía para principiantes" does not match the search.

I thought I saw another bug about this, that wasn't bug 284856 or bug 525537. Aureliano, do you recall seeing one?

Comment 8

[Wayne Mery (:wsmwk)]

[getting harder to finish a thought]

(In reply to comment #6)
> Wayne, Is that a question?

yes :(
 
> My issue is about the search box on the top-right corner of Thunderbird. For
> example when selecting "By subject..." and then typing "guia" a message with
> the subject "Guía para principiantes" does not match the search.

so filters

Comment 9

[steve hayes]

This should apply to  
1 filters 
2 search all messages, and 
3 search within a message

and (as an aside) search within a page in firefox

If you actually want to search for an accented character you can, and it all operates quite correctly
but normally you want to search for paris or parís however it was typed.

In French at least it's particularly relevant, because it is very common to omit the accent on upper case characters. Filters/searches are case independent by default, it would be better if they were also accent independent by default. 

Google seems to do it right:
if you click on http://www.google.fr/search?q=site:usefulshortcuts.com+e&sourceid=navclient-ff&rlz=1B3GGLL_en-GBFR382FR382&ie=UTF-8&hl
you will see that all versions of letter e (with or without accent) are bolded ie marked as hits

whereupon http://www.google.fr/search?q=site:usefulshortcuts.com+e&sourceid=navclient-ff&rlz=1B3GGLL_en-GBFR382FR382&ie=UTF-8&hl=#sclient=psy&hl=fr&rlz=1B3GGLL_en-GBFR382FR382&source=hp&q=site:usefulshortcuts.com+%C3%AB&aq=f&aqi=&aql=&oq=&gs_rfai=&pbx=1&fp=32bebe5f540a3320
only finds the accented character 

I've just tested the behaviour of IE and Opera and they both seem to do as TB/FF do, which is a shame.

Steve

Comment 10

[sunk818]

Don't forget The cédille ¸ (cedilla) is found only on the letter C. It changes a hard C sound (like K) into a soft C sound (like S), e.g., garçon. The cedilla is never placed in front of E or I, because C always sounds like an S in front of these vowels.

I also recently received an e-mail from Norway whether the O had a forward slash through it.

Comment 11

[Gene Scharmann]

I don't know if this is relevant to this discussion of accents, nor have I noticed any problems.

The Norwegian alphabet consists of 29 letters.

The 'O with Slash' "Ø" and "ø"  (transcribed as OE oe)
The 'A with Ring' "Å" and "å"  (transcribed as AA aa)
And 'Latin AE' "Æ" and "æ"  (transcribe as AE ae)

These are Latin letters and NOT accents.

The sorting order is 'ABC...XYZÆØÅ' (ie: AE OE AA after Z)

To complicate matters they is sometimes written with the Swedish characters
the Å 'A with Dieresis' "Ä" and "ä"
the Ö 'O with Dieresis' "Ö" and "ö"
again, these are NOT accents, but normal letters of the alphabet.

Comment 12

[Ludovic Hirlimann [:Usul]]

Removing myslef on all the bugs I'm cced on. Please NI me if you need something on MailNews Core bugs from me.