% is a special character that is not indexed AFAIK

Sorry Matt, but that cant be the explanation as the failure I am seeing is not confined to % sign.

I have got the same promblem with "-" and "#"

My non-geek-qualified experiments indicate that it just IGNORES the subjecct line in the search

o(

Ted

What does "AFAIK" mean?

tedv said

What does "AFAIK" mean?

Google is your friend https://www.google.com.au/?gfe_rd=cr&ei=Gab4V5n7GMfN8gflmpCYDw#newwindow=1&q=afaik+meaning

GIYF.

AFAIK = As far as I know.

Special characters appear to be those set aside for specific rôles in URIs and email addresses. I think you'll find that @ and + are also ignored. I must admit I'm surprised to see - in your list.

I use regular expression based searches on the very rare occasions when I do need to search on these characters.

Ta Both,

Methinx you are answering my question as if i am asking about Special Characters.

I am not asking about special characters.

I *am* asking how to get the software to search *ALL* the fields in a single search.

My K9 mail client in Android does that straight off. I'm confused as to how to get Thunderbird to behave the same way.

As far as I can detect by trial&error, Thunderbirds global search box at the top right ONLY searched the BODIES of messages?

Ted

My experience is contrary to yours. I find it searches subjects and correspondents and is less good at bodies. In the case of IMAP-connected accounts, message body text is not necessarily downloaded and so the search needs to be transferred to the server.

Have you tried a simple subject search without any of these alleged special characters? If that fails, I'd suggest you reset your "gloda" - the global database for this particular search mode.

It is not a precise search. It will find the search term inside longer words and will deliver near misses too. If I need a precise search I use the older search mode accessed via ctrl+shift+f and if I need it to span multiple folders I make it into a Saved Search.

Usually if I do use Global Search, I leave it quite vague and generalised and immediately switch to the traditional view via "show email as list" where I can use the more precise search tools associated with folders.

I don't know where the Global Search came from. Its very different look-and-feel has me thinking that it is something like the results of a sprint or Google Summer of Code project.

tedv said

Sorry Matt, but that cant be the explanation as the failure I am seeing is not confined to % sign. I have got the same promblem with "-" and "#" My non-geek-qualified experiments indicate that it just IGNORES the subjecct line in the search

o(

Ted

I tell you one special character so you off and list more special characters. and from that draw a conclusion.

I just typed caret position wrong into global search on my machine. the very first result only had those words in the subject line and nowhere else. I was on the third match before the word wrong appeared in the body of the email.

Perhaps your index is dead and needs to be deleted. See https://support.mozilla.org/en-US/kb/rebuilding-global-database

Or you are only searching using wildcards. see http://www.w3schools.com/sql/sql_wildcards.asp

There is some fairly heavy reading here https://developer.mozilla.org/en-US/docs/Mozilla/Thunderbird/gloda

Discussing tokenizers, whittlers etc. all designed to index only useful stuff. Things like search terms of 2 letters or less will return nothing as words that short are not indexed.

Thank you both for trying, but, your dictionary is plainly a different one to mine as I barely understood half of your recent posts.

Sounds like we should stop now, and for me to hope for a response to my simple question that I might understand.

Ted

Can you please point out a single difficult technical word either of us has used?

Please follow the link provided to reset your global database.

" traditional view"

"tokenizers, whittlers etc."

"Perhaps your index is dead and needs to be deleted"

"caret position wrong "

" only searching using wildcards"

I am wasting your effort here.

Its Saturday 1036 here, and I have to go shopping now.

Thanks for trying, but I am not geek enough to follow your words

Ted

I dare not "reset my global database" without knowing what will be the consequences of that...

I have had terrible experiences of following such instructions in the past only to find I have lost data

Ted

What I learn from the article on the global search is that it simplifies the text and throws out punctuation. So that probably explains why it doesn't see a - character. If I understand it correctly, it converts uppercase characters to their lower case forms and converts characters with accents to their non-accented cousins. So it might treat ä as a.

It ignores "words" shorter than three letters. So whether or not % is treated as punctuation (and therefore ignored) in the sense that it may have a special meaning in links and addresses, it won't support a search for that single character, particularly if it is surrounded by spaces making it into a one-letter word. The % symbol is used a lot in encoding things. You can't have a space in a web address for instance, but you can represent a space by the code "%20".

I thought there was a simpler way to reset the global database. I'd like to see if there is a button for this in the settings, but I don't have Thunderbird here right now. The advice given to delete a file in the profile is indeed rather too geeky.

The article on the global database tells us that it is just an index. If you delete it it will be rebuilt. No data is lost.

Thank you Zenos, for the info about not losing data.

I'm very annoyed with whomever created the search function... from what you say they made it do something other than what the plebean user is led to believe it does...

• and then failed to tell that pleb that they had forced that change*.

Thus causing confusion, irritation and time wasting. like we have done here.

Again, thank you for trying; If you know who those people are, please direct heir attention to my words here.

If the article Matt links to is to be believed, it appears to be very carefully thought through in terms of computer science and its internal structure and methodology. They don't seem to have done a very good job in thinking through what the average user wants from it. They do recognise the the two-letter word size limit is a bit silly, and that it has no concept of common "stop" words ("and", "the", etc) which could usefully be ignored (but that's hard work when you consider providing support for multiple languages).

I'd like to see support for explicit exact matches so it doesn't throw up near misses.

It feels like a work in progress, but it's way down the list of priorities given the short supply of people to address these and many other issues. Don't hold your breath waiting for improvements.