532753 - Gloda search has problems finding words in subject line containing non-ASCII characters

Status: VERIFIED DUPLICATE of bug 529824

(Thunderbird :: Search, defect)

Description

[Karel Koubek]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2b4) Gecko/20091124 Firefox/3.6b4
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.5) Gecko/20091130 Thunderbird/3.0

I've played with the new Gloda search bar and found that:

1) searching for "bakalářská práce" (bachelor thesis in Czech language) yields zero results, even though this exact phrase is included in the subject of one of my e-mails! All of my e-mail seemed to be successfully indexed before search performed. Content-Transfer-Encoding: base64 of this e-mail.

2) searching for "bakalarska prace" (without diacritics) seems to find all emails containing this phrase, thus working as expected.

There are other Czech words with diacritics in the subject field, that Gloda seems to have problem with.

Reproducible: Always

Steps to Reproduce:
1. Perform search for words included in the subject line of your e-mail that contains non-ASCII characters (i.e. ěščřžýáíé)

Actual Results:  
"No messages match your search"

Expected Results:  
Gloda search should be able to find e-mails with non-ASCII words in the subject line, otherwise it's quite useless for non-English users.

I would really like to find what's going on here. I find it glaring that such a bug would get into a release candidate! I will gladly provide any other piece of information needed to settle this out, thanks!

Comment 1

[Andrew Sutherland [:asuth] (he/him)]

Please attach an example message that gloda does not find and the exact search string that you are using to find the message.  You can use "File... Save As... File" to export the message to a form suitable for attachment.

Feel free to alter portions of the message that are not critical to the test to avoid spammers picking up on it or the revelation of information you do not want the world to see.

Comment 2

[[:Aureliano Buendía]]

Could be a dupe of bug #529824

Comment 3

[Karel Koubek]

Attachment: Sample e-mail demonstrating the problem

Comment 4

[Karel Koubek]

Sorry for cross posting, my mistake when submitting the attachment. For comments see the above bug #529824. I'll just add that performing standard search (Ctrl+Shift+F) in subject field works as expected. Even when I select the folder in which the e-mail resides and select 'Subject filter' in the Gloda bar and perform the search, all instances are found and shown.

Maybe it has something to do with the encoding as Christos R. pointed out. If I look into the source of the message, ISO-8859-2 is used in the subject.

How can make sure, that all of my messages had been successfully indexed? Thanks.

Comment 5

[Andrew Sutherland [:asuth] (he/him)]

Duping to the other bug; the description of the problem is consistent with the problem found and addressed on that bug.