1141034 - Autocomplete of addresses is slow

Status: RESOLVED DUPLICATE of bug 1012397

(Thunderbird :: Message Compose Window, defect)

Description

[Christian Moling]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:36.0) Gecko/20100101 Firefox/36.0
Build ID: 20150305021524

Steps to reproduce:

Since the version 30.x the autocomplete of email-addresses is very slow. I know, now the lookup works different. But now, when I type a part of an address, i must wait the results and cannot confirm with ENTER immediatly.


Actual results:

for example, I enter
Paul + ENTER
no email-address appears


Expected results:

Paul + Enter
the most used email-address appears: [e-mail]

Comment 1

[Wayne Mery (:wsmwk)]

(priority is for developers)
What is the full number of the version are you using now?

Comment 2

[Christian Moling]

Now I use the version 31.5.0.

Comment 3

[:aceman]

(In reply to Christian Moling from comment #0)
> Actual results:
> 
> for example, I enter
> Paul + ENTER
> no email-address appears

What do you mean that no address appears? Even after you wait for a long time?

Comment 4

[Christian Moling]

I'm waiting for 1 seconde. I know, it's not so much, but with olter Thunnderibrd-versions I wrote a part of the email-address, confirmend with ENTER immediatly and the email-address appears.

Comment 5

[:aceman]

How many contacts do you have in all your addressbooks in total?

Comment 6

[Christian Moling]

I have 150 contacts (local and remote, sogo connector). But now I simualted the problem with only 11 contacts, ONLY in the local adddressbook. It's the same.

Comment 7

[:aceman]

With such a small local addressbook it should not be so slow. Yes, there is an initial delay before the autocomplete even starts, but that is like half a second.
We specifically tested the autocomplete and only notice it being slow for 10000s of cards. In such a case it does take several seconds to return the results.
In your case it would be instantaneous. What is your CPU? How did you test the local addressbook case? How did you make TB ignore the remove AB?

Comment 8

[Christian Moling]

What is your CPU?
My CPU: Intel Xeon Quad-Core 2,8 GHz (the "problem" is on all PC-clients in our company...)

How did you test the local addressbook case?
How did you make TB ignore the remove AB?
I removed the remote addressbooks and wrote some contacts in the Personal Address book.

So when you say the response-time (initial delay) of a half second is normaly, we can close the bug. But in older TB-Versions there was no initial delay. The "old" selection of contacts was very fast.

Comment 9

[:aceman]

(In reply to Christian Moling from comment #8)
> So when you say the response-time (initial delay) of a half second is
> normaly, we can close the bug. But in older TB-Versions there was no initial
> delay. The "old" selection of contacts was very fast.

I checked the code and the timeout is 0.3seconds. So you shouldn't see a 1second delay. As you do, there may be something wrong.
Some initial delay must be there to prevent 1-character searches across big addressbooks (see bug 1085674), if the user types slowly.

Do you also see the slowness when just opening the addressbook window and showing all contacts in the Personal AB?

Comment 10

[Christian Moling]

No, there is no slowness when I open the addressook window and show the contacts in any AB.

Here you can see the initial delay. Is ist ok, normal?
https://storage.control.troyer.it/public.php?service=files&t=7c1079ee64d978a71a0620523d68d7f2
pwd: thunder

(Normally the addresss "[e-mail] would have to appear at first positon, most used address, but for these case I will open another/nef Bug-report).

Comment 11

[:aceman]

You have left the first 2 fields before it could autocomplete anything. That looks like bug 1012397.
In the third field you waited a split second and got a result. You can't count the time it took you to fill the 3 fields together.

Comment 12

[Christian Moling]

Yes it's the bug 1012397. I don't count the time for all attempts.
So we can close this bug and i follow the bug 1012397.
Thanx a lot for yout support!

Comment 13

[:aceman]

Thanks.