1170210 - Password manager does not save passwords

Status: RESOLVED WORKSFORME

(SeaMonkey :: Passwords & Permissions, defect)

Description

[Rainer Bielefeld]

Steps how to reproduce with EN-US SeaMonkey 2.38a1 build by Ruediger Lahl, Gecko/20100101, Build 20150531140409 (Default Theme) on German WIN7 64bit;

1. With new test-profile in browser open a web page with login, for 
   example this Bugzilla page, make sure that you are not logged in
2. Click 'log in'
   » input lines for user name and password appear
3. Fill input lines with your user data and click [Log in]
   » Login process will start
   Expected: Password manager brings up “Save Password?” message
   Actual: Nothing happens

Additional information:
a) I used a rather exotic user build; creator " Ruediger Lahl" told 
   about this problem at Newsgroup 
   "de.comm.software.mozilla.nightly-builds", 
    Subject "[SM] Selfmade Mozille"
b) Might have appeared before 2.38, I can't tell

Comment 1

[Rainer Bielefeld]

c) I did not find a DUP with <https://bugzilla.mozilla.org/buglist.cgi?cmdtype=runnamed&namedcmd=DUPs1170210&list_id=12294030>

Comment 2

[Rainer Bielefeld]

d) MailNews is not affected, PM works fine for me there.

Comment 3

[Rainer Bielefeld]

Worked foine with EN-US SeaMonkey 2.33.1 (German Language pack)  Gecko/20100101 Build 20150321194901 (Default Theme) on German WIN7 64bit, so REGRESSION!

Comment 4

[therube]

> SeaMonkey 2.38a1 build by Ruediger Lahl, Build 20150531140409

Using /that/ build, I can confirm there is no password save prompt.
(Note: it has calendar & who knows what else in there.)

And there is mention of dataman.js Error Console:

> Error: TypeError: invalid 'instanceof' operand Components.interfaces.nsIDOMStorageEvent
> Source File: chrome://communicator/content/dataman/dataman.js
> Line: 2570

And (from what I can make of it) there is an entire section of code that is wacked?
 
 
And ...
 
 
> SeaMonkey 2.38a1 build by Ratty, Build 20150522012631 (a few days, 9, earlier then above)

Now Ratty's build also generates that same error, but that is "just" one line that is highlighted:

> Error: TypeError: invalid 'instanceof' operand Components.interfaces.nsIDOMStorageEvent
> Source File: chrome://communicator/content/dataman/dataman.js
> Line: 2570

And even so, that build does successfully prompt & save passwords, as expected.

Now if a regression crept in between the two builds, or if is just /how/ the builds were built that is affecting things, don't know?
 
 
(And this entire post could be pointless, as we [on Windows] have nothing "official" to test with.)

Comment 5

[Rainer Bielefeld]

NOT reproducible with EN-US SeaMonkey 2.38a1 build by Philip Chee, Gecko/20100101 , Build 20150601233926 (Classic Theme) on German WIN7 64bit.

but: I read at de.comm.software.mozilla.nightly-builds, 06.05.2015 01:15, Subject "[SM Aurora] Passwort speichern geht nicht" by Susanne Jäger that she also observes the problem. 
User agent: Mozilla/5.0 (X11; Linux x86_64; rv:39.0) Gecko/20100101 Firefox/39.0 SeaMonkey/2.36a2
Build identifier: 20150504013001

Comment 6

[Hartmut Figge]

I have tested the STR with my self compiled SM-Trunk Linux x86_64 of today. I can confirm the issue. The error console says

Timestamp: 03.06.2015 09:15:04
Error: TypeError: LoginManagerContent.onFormPassword is undefined
Source File: resource://gre/components/nsSuiteGlue.js
Line: 15

Timestamp: 03.06.2015 09:15:04
Error: TypeError: Components.classes[JSD_CTRID] is undefined
Source File: jar:file:///home/hafi/.mozilla/seamonkey/3r4xm4e3.t2/extensions/%7Bf13b157f-b174-47e7-a34d-4815ddfdfeb8%7D.xpi!/components/venkman-service.js
Line: 195

Will search for the regression range.

Comment 7

[Hartmut Figge]

Regression range:

Last good: 2015-05-13 15:46:00 PDT   c-c:b557915203f8 m-c:1fab94ad196c
First bad: 2015-05-14 19:59:00 PDT   c-c:c59e01475ae2 m-c:25c5525a1000

I am hoping that bisecting will not be necessary. :)

Comment 8

[Hartmut Figge]

Of course, the second error in Comment 6 is not relevant. I should have checked before C&P. ;)

Comment 9

[Hartmut Figge]

The first bad revision is:
changeset:   243891:ea97ed03aa16
user:        Paolo Amadini <[e-mail]>
date:        Wed May 13 15:34:14 2015 +0100
summary:     Bug 1149975 - Part 1 of 2 - Handle visibility of the login fill doorhanger anchor. r=MattN

Verified by backout. In current source backout is no longer possible.

Comment 11

[Rainer Bielefeld]

(In reply to Philip Chee from comment #10)
Still REPRODUCIBLE with  en-US SeaMonkey 2.40a2 Mozilla/5.0 (Windows NT 6.1; x64; rv:43.0 from akalla download area)  Gecko/20100101  Firefox/ 43.0  Build 20151010043152, (Classic Theme) on German WIN7, so Target of Bug 1158496 does not match.
Version 2.38 here does not match with Version of Bug 1158496 "2.33"
So I reopen this one.

@Hartmut Figge:
Can you confirm 2.40?

Comment 12

[Rainer Bielefeld]

NEW due to comment 7

Comment 13

[Hartmut Figge]

(In reply to Rainer Bielefeld from comment #11)

> @Hartmut Figge:
> Can you confirm 2.40?

The bug is fixed for my SM-Trunk from today. Tested with two different new profiles.
User agent: Mozilla/5.0 (X11; Linux x86_64; rv:44.0) Gecko/2015101401 SeaMonkey/2.41a1-h

Comment 14

[Philip Chee]

(In reply to Hartmut Figge from comment #13)

> The bug is fixed for my SM-Trunk from today. Tested with two different new
> profiles.
> User agent: Mozilla/5.0 (X11; Linux x86_64; rv:44.0) Gecko/2015101401
> SeaMonkey/2.41a1-h