1571724 - Thunderbird 69.0 b1 trucates recepients (5 first characters) using MAPI's nsMapiHook::PopulateCompFieldsW()

Status: RESOLVED FIXED

(MailNews Core :: Simple MAPI, defect)

Description

[Jérôme Baldo - Cogidis]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:68.0) Gecko/20100101 Firefox/68.0

Steps to reproduce:

Send an email from another application (for instance from Microsoft Access with "DoCmd.SendObject") specifying the recepient (ex : [e-mail])

Actual results:

The recipient's address is truncated : the first 5 characters disappear :

[e-mail]

in the "For :" field

Some beta versions since the 60 branch sometimes displayed nothing, some other times also truncate the 5 first characters.

Expected results:

As in the non beta versions that work OK with the same program, it should display
[e-mail].

Comment 1

[Jorg K (CEST = GMT+2)]

Hmm, that's bad, I'll take a look. Not many applications using MAPI specify the recipient.

Five rings a bell, it must be this code, but that looks right:
https://searchfox.org/comm-central/rev/2d079399f2abf8119b05fbd6bb5782b28412d881/mailnews/mapi/mapihook/src/msgMapiHook.cpp#605

But this doesn't :-(
https://searchfox.org/comm-central/rev/2d079399f2abf8119b05fbd6bb5782b28412d881/mailnews/mapi/mapihook/src/msgMapiHook.cpp#713

nsMapiHook::PopulateCompFieldsW() is new.

Comment 2

[Jorg K (CEST = GMT+2)]

Attachment: 1571724-mapi.patch

I don't know how this slipped through my review :-(

First reviewer wins.

Comment 3

[Magnus Melin [:mkmelin]]

Comment on attachment 9083441 [details] [diff] [review]
1571724-mapi.patch

Review of attachment 9083441 [details] [diff] [review]:
-----------------------------------------------------------------

Looks ok to me, r=mkmelin

Comment 4

[Jorg K (CEST = GMT+2)]

Attachment: 1571724-mapi.patch (v2)

I've chatted a bit with Eric Rahm on IRC to see whether we need nsDependentSubstring here. In the end it's cheaper to use native Windows functions, in this case _wcsnicmp, wide string N ignore case compare.

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/comm-central/rev/c423618c2ac6
Fix logic error in string comparison in nsMapiHook::PopulateCompFieldsW(). r=mkmelin

Comment 6

[Jorg K (CEST = GMT+2)]

TB 68.0 ESR:
https://hg.mozilla.org/releases/comm-esr68/rev/35d9da68475a55b796a9aecbe9944b9c9c5eb234

Comment 7

[Jorg K (CEST = GMT+2)]

Jérôme, please try the TB 68.0 ESR pre-release:
32bit: https://queue.taskcluster.net/v1/task/chOrjxdaRHG2WS0uIE9lWw/runs/0/artifacts/public/build/install/sea/target.installer.exe
64 bit: https://queue.taskcluster.net/v1/task/RfkCLtIjQJChRi8St_iDzA/runs/0/artifacts/public/build/install/sea/target.installer.exe

I'm not sure whether the 32bit version will execute the code path we fixed.

Comment 8

[Jérôme Baldo - Cogidis]

I tested the 32 bit TB 68.0 ESR version on another PC, it worked like a charm with the full address. You can consider the bug as fixed and don't worry, if I see it appear again I will let you know, I use MAPI very frequently. Thanks again.

Comment 9

[Jorg K (CEST = GMT+2)]

TB 69 beta 3:
https://hg.mozilla.org/releases/comm-beta/rev/dadc53787af48f91dfecd8f82973acc7e4908395

Comment 10

[Jérôme Baldo - Cogidis]

I confirm that the problem is fixed with TB 69 beta 3.